THE JOURNAL OF COMPARATIVE NEUROLOGY 312:l-18 (1991)

Serotoninergic Innervation of the Thalamus in the Primate: An Immunohistochemical Study BRIGITTE LAVOIE AND ANDRE PARENT Centre de Recherche en Neurobiologie, Universite Laval et HBpital de 1'Enfant-Jesus, Quebec, QC, Canada

ABSTRACT Little is known of the serotoninergic innervation of the thalamus in primates; therefore, we undertook a detailed study of the distribution of 5-hydroxytryptamine (5-HT)-immunoreactive neuronal profiles in the thalamus of the squirrel monkey (Saimiri sciureus) with a specific antibody directly raised against 5-HT. All thalamic nuclei in the squirrel monkey displayed 5-HT-immunoreactive fibers, but none contained immunopositive cell bodies. The 5-HT innervation of the thalamus derived from extrinsic fibers arising mostly from the midbrain raphe nuclei and forming the transtegmental system. Most of the fibers destined to the thalamus collected into a major bundle that swept dorsoventrally within the midbrain tegmentum and coursed beneath the thalamus along its entire caudorostral extent. Several fiber fascicles broke off from this main bundle at different levels and ascended dorsally to innervate the various thalamic nuclei. Overall, the 5-HT innervation of the thalamus in the squirrel monkey was more massive than would have been expected from earlier studies in nonprimate species. Marked differences in the regional density of innervation were noted both between the various nuclei and within single nuclei. The most densely innervated nuclei were those delineating the principal subdivisions of the thalamic mass, that is, the midline, rostral intralaminar, limitans, and reticular nuclei, where very dense fields of isolated axonal varicosities occurred. In contrast to the rostral intralaminar nuclei, which were rather uniformly innervated, the centre median/parafascicular complex contained immunoreactive fibers and isolated varicosities distributed according to a mediolateral gradient. The habenula and the ventral anterior nucleus were among the most weakly innervated nuclei. In the latter nucleus, as well as in more densely innervated nuclei, thin varicose fibers formed numerous pericellular contacts on cell bodies and proximal dendrites of thalamic neurons. The 5-HT innervation of the lateral nuclear group as well as that of the medial and lateral geniculate nuclei ranged from very weak to dense. The mediodorsal nucleus displayed a highly heterogeneous 5-HT innervation that varied from weak in its central portion to moderate or dense in its medial and lateral borders. A moderate 5-HT innervation was observed in the anterior nuclear group. The surprisingly dense and heterogeneous 5-HT innervation of the thalamus noted in the present study suggests that serotonin may be involved in several specific functions of the thalamus in primates. Key words: 5-hydroxytryptamine,diencephalon, thalamic afferents, monoaminergic systems, raphe nuclei ascending projections, squirrel monkey, Suirniri sciureus

Among the most extensively studied chemospecific brainstem afferents to the thalamus are those utilizing either acetylcholine or monoamines as neurotransmitters. The brainstem cholinergic innervation of the thalamus arises from neurons located in the pedunculopontine and laterodorsal tegmental nuclei (Woolf and Butcher, '86; Hallanger et al., '87; Pare et al., '88; Steriade et al., '88; Steriade and Biesold, 'go), whereas the monoaminergic innervation arkh a t e s from noradrenergic neurons located in the locus

o 1991 WILEY-LISS, INC.

coeruleus (Lindvall et al., '74; Moore and Card, '84) and serotoninergic neurons lying in the midbrain raphe nuclei (Azmitia and Segal, '78; Moore et al., '78; Moore, '81; Tork, '85). In addition, a rather discrete dopaminergic innervaAccepted June 18, 1991. Address reprint requests to Brigitte Lavoie, Centre de recherche en neurobiologie, HBpitg de YEnfant-Jesus, 1401 18' rue, Quebec CQC), Canada, G i j 124.

B. LAVOIE AND A. PARENT

2 tion of the habenular nuclei originating from neurons in the ventral tegmental area has been described (Phillipson and Griffith, '80; Bjorklund and Lindvall, '84; Skagerberg et al., '84). However, by comparison with the wealth of knowledge on the organization of the cholinergic and catecholaminergic inputs to the thalamus, little is known of the serotoninergic innervation of this structure. Different approaches have been used to study the cellular localization of serotonin (5-hydroxytryptamine, 5-HT) in the central nervous system (CNS). First, the formaldehydeinduced fluorescence (FIF) procedure, developed by Falck and Hillarp ('62), was extensively employed by Dahlstrom and Fuxe in their pioneering study of the distribution of 5-HT-containing cell bodies and axon terminals in the rat CNS (Dahlstrom and Fuxe, '64; Fuxe, '65). In addition, intraventricular injections of tritiated 5-HT were also utilized to visualize neurons endowed with a powerful reuptake mechanism for 5-HT in the rat CNS (Chan-Palay, '77; Parent et al., '81). The latter two approaches have demonstrated the presence of fibers and axonal varicosities exhibiting fluorescence of the 5-HT type or accumulating tritiated 5-HT in the rodent thalamus. However, because the FIF method is poorly sensitive for 5-HT and tritiated 5-HT cannot easily reach structures located deep from the ventricular surface, information on the 5-HT innervation of the thalamus gathered with these two approaches has remained parcellary. A much more detailed knowledge of the 5-HT innervation of the CNS was obtained following the development of highly specific antibodies directed against the neurotransmitter itself (Ranadive and Sehon, '67; Steinbusch et al., '78; Wallace et al., '82). The immunohistochemical approach has been used in numerous studies of the distribution of 5-HT fibers, particularly in the cerebral cortex and basal ganglia (Tork, '85; Azmitia and Gannon, '86; Fallon and Loughlin, '87; O'Hearn et al., '88; Wilson et al., '89; Lavoie and Parent, '90). By comparison, little is known of the 5-HT innervation of the thalamus. Except for the detailed immunohistochemical study of the thalamus in the rat by Cropper et al. ('84), the information of the 5-HT thalamic innervation is limited to findings scattered within investigations describing the overall organization of the 5-HT innervation in the rodent CNS (see Steinbusch, '81). Our knowledge of the 5-HT innervation of the thalamus in primates is even more fragmentary.

Morrison and Foote ('86) have described the noradrenergic and serotoninergic innervation of some of the visual thalamic nuclei and of the reticular thalamic nucleus in macaques and squirrel monkeys. These authors have shown, among other things, the existence of some species variations in the distribution of these two monoaminergic fiber systems in the lateral geniculate nucleus. The present study, which is part of an ongoing research program on the monoaminergic innervation of the forebrain in primates (Lavoie et al., '89; Lavoie and Parent, '901, provides the first detailed description of the 5-HT innervation of the thalamus in a primate, namely, the squirrel monkey.

MATERIALS AND METHODS Preparation of tissue Three female, adult (approximately 3 years old), squirrel monkeys (Saimiri sciureus), 550 to 800 g in body weight, were used in the present study. The animals were perfused transcardially under deep pentobarbital anesthesia. The vascular system was first rinsed with a heparinized saline solution (500 ml of 0.9% NaCl), followed by 1 liter of a fixative solution containing 4% paraformaldehyde and 0.2% glutaraldehyde. The brains were cut in the transverse plane into 5 to 10-mm-thick slabs according to the stereotaxic coordinates of the atlas of Emmers and Akert ('63). These blocks were postfixed for 1hour in a 4% paraformaldehyde solution, kept overnight at 4°C in a phosphate buffer solution (PBS, 0.1 M, pH 7.4), and then cut at 50 km with a Vibratome (Oxford Instruments). Sections from the brainstem raphe nuclei to the rostral thalamus were serially collected in a PBS solution prior to immunohistochemical processing.

5-HT immunohistochemistry The sections were processed for the demonstration of 5-HT immunoreactivity according to the peroxidase antiperoxidase (PAP) procedure of Sternberger ('86). All the immunohistochemical reagents were diluted in a PBS solution containing 1%normal goat serum (NGS) and 0.1% Triton X-100. Following a preincubation for 30 minutes at room temperature in a solution containing 10% NGS, 0.4% Triton X-100, and PBS, the sections were incubated over-

Abbreviations

AD AM AV C CeL CeM CM

cs

dLG EML H H1 Hm HP LD Li LP MD MG MT Pc

anterodorsal nucleus anteromedial nucleus anteroventral nucleus central nucleus centrolateral nucleus centromedial nucleus centre median nucleus central superior lateral nucleus dorsal part of lateral geniculate nucleus external medullary lamina H-fields of Fore1 lateral habenular nucleus medial habenular nucleus habenulo-interpedunculartract lateral dorsal nucleus nucleus limitans lateral posterior nucleus mediodorsal nucleus medial geniculate nucleus mammillothalamic tract paracentral nucleus

PC Pd Pf Po Pi P1 Pm PT

R Re SG SM

VA VL vLG VPi VPl VPm VPmc ZI

posterior commissure peripeduncular nucleus parafascicular nucleus oral part of pulvinar inferior part of pulvinar lateral part of pulvinar medial part of pulvinar pretectal area reticular nucleus nucleus reuniens suprageniculate nucleus stria medullaris ventral anterior nucleus ventral lateral nucleus ventral part of lateral geniculate nucleus inferior part of ventral posterior nucleus lateral part ofventral posterior nucleus medial part of ventral posterior nucleus pars compacta of the medial part of ventral posterior nucleus zona incerta

SEROTONINERGIC INNERVATION OF PRIMATE THALAMUS

3

TABLE 1. Variations in the Density of the 5-HT Innervation of Various Thalamic Nuclei Along Their Caudorostral Extent

3.0

4.5

AD AM AV C CeL

dLG

++

+-I+

+

HI Hm LD Li LP MD MG Pc Pf Po Pi PI Pm R Re SG VA VL vLG

VPi VPl VPm VPmc ZI

6.5

7.0

7.5

8.0

8.5

9.0

9.5

10.5

t

+

t

+

t

t

++I+++

CeM CM

cs

5.5

+-I+ -I+-

+-I+

+I+++ tt

tt

ttt

-It

-I+

t-

+-

-

t+

+ -

+-It +-it

-

+t t+t+

+-I+ +-I++

+-It t-/++

+

+

+t tt

+I++

+I+++

++ +-

+I++

+I++

+I++

tt

++

tt

tt

+

+

t

t-

+-

+-

+-

++ +-It

+-It

+

++

+

t t i

+ + ++

+

t

t-

+It+

+I++

+I++

++

+I++

+It+ +t

+It+ t+

+-I+

t-/+

+-it

+

+

++

+-It

t++

+

+ +

tt

t t+

+I++ t

++

+-I+

+-I+

-

-

-

t-

t-

t-

t-

+-I+

+-I+

+-I+

+I++

+

+ +

++

-

+

-

-

+

+

+I++

++

tt

+t

t+

-1t-

-

+I++

t

The different nuclei are identified by their abbreviations on the left, whereas the different anteroposterior coordinates are indicated at the top. The scale of innervation density used ranges from very weak (-1, weak (+-), moderate (+I,dense (+ t),to very dense (++ tj.

night at 4°C with a polyclonal rabbit anti-5-HT (1:500, Immunonuclear Gorp. Stillwater, MN). The sections were then rinsed in PBS and incubated for 1 hour at room temperature in a goat antirabbit IgG (1:30, Sigma), rinsed again in PBS, and finally incubated for another hour at room temperature with rabbit PAP complex (1:100, Sigma). After this last incubation, the sections were rinsed three times in PBS and the bound peroxidase was revealed by a last incubation in a solution of 3-3' diaminobenzidine tetrahydrochloride (DAB, 0.05%, Sigma) and hydrogen peroxide (H,O,, 0.008%). The sections were mounted on gelatine-coated slides with Permount and observed with a Leitz microscope under both bright- and darkfield illuminations. Furthermore, a complete series of contiguous brain sections were stained with cresyl violet to facilitate the identification of the various thalamic nuclei. The nomenclature used in the present study was adapted from that found in the atlas of the squirrel monkey by Emmers and Akert ('63) and in the more detailed cytoarchitectonic study of the thalamus of the rhesus monkey (Macaca mutattu) by Olszewski ('52). However, we have followed the current trend in thalamic terminology by avoiding as much as possible the Latin terminology (see Jones, '85).

Control experiments The specificity of the 5-HT immunostaining obtained with the Immunonuclear 5-HT antibody in brain tissue from the squirrel monkey has been tested previously (Lavoie and Parent '90). I n brief, a series of contiguous sections were processed as above except that the anti-5-HT serum was replaced by nonimmunized rabbit serum. Other sections were incubated in rabbit anti-5-HT serum preabsorbed for 16 hours at 4°C with serotonin-creatine sulfate at concentrations ranging from lo-' to M. Sections

incubated in normal rabbit serum remained virtually free of immunostaining, whereas a marked decrease of immunostaining was observed in sections incubated with anti-5-HT serum preabsorbed with 5-HT. In these experiments the intensity of the immunostaining was inversely proportional to the concentration of the serotonin used for the immunoabsorption (see Steinbusch, '81).

RESULTS General considerations In the squirrel monkey, all thalamic nuclei displayed 5-HT-immunoreactive (5-HT-IR) fibers, but none contained 5-HT-positive cell bodies. The density of the 5-HT innervation varied from one nucleus to the other, and a heterogeneous distribution of the 5-HT fibers and axonal varicosities was frequently observed within individual nuclei. The density of the immunoreactive neuronal profiles encountered within the different thalamic nuclei ranged from very weak to very dense. We have arbitrarily chosen to classify the relative density of the 5-HT innervation into five categories: very weak (-), weak (+-I, moderate (+I, dense (+ +), and very dense (+++). This scale was used in Table 1, which provides an overview of the density of the 5-HT innervation of the major thalamic nuclei such as disclosed along the various caudorostral levels. In addition, the range of densities of the 5-HT innervation for a particular nucleus is indicated by the use of the slash mark (e.g., weak to moderate: +-/+I. Based on the morphological criteria used by Steinbusch ('81) in his immunohistochemical study of 5-HT distribution in the rat CNS, the nonperikaryal5-HT-IR neuronal profiles observed in the present investigation were classified into: (1) smooth, nonvaricose fibers, (2) fibers endowed with varicosities, and (3) varicosities without clear intervar-

B. LAVOIE AND A. PARENT

4 icose segments termed isolated varicosities. In the thalamus of the squirrel monkey, the 5-HT nonvaricose fibers often followed well-defined myelinated fiber tracts, such as the internal and external medullary laminae, and formed the major 5-HT fiber fascicles innervating the thalamus. They most likely correspond to the more proximal portion of the 5-HT axons. The varicose fibers were often seen to break off from the nonvaricose fibers. These fibers were generally sinuous, without specific orientation, and most probably represent the more terminal portion of 5-HT axons. They frequently displayed typical pericellular arrangements around nonimmunoreactive thalamic neurons. Varicosities without intervaricose segments were also seen in close appositions to unlabeled cell bodies and proximal dendrites of individual neurons in various thalamic nuclei (see below). These isolated varicosities, reminiscent of terminal boutons, are considered here as part of the terminal arborization of 5-HT axons.

Fiber pathways The 5-HT innervation of the thalamus in the squirrel monkey derived largely from the transtegmental pathway, which is composed of numerous fiber fascicles originating from both the dorsal and median (central superior) raphe nuclei. As they arched ventrally, these fiber fascicles formed a very dense and intricate network covering most of the central portion of the midbrain tegmentum (Parent et al., '81; Lavoie and Parent, '90). More rostrally these fascicles formed a massive bundle, which ascended within the dorsal portion of the lateral hypothalamic area. This bundle first swept ventrolaterally through the core of the midbrain tegmentum (Fig. 1A-C), then pierced the H-fields of Forel (Fig. 2A,B),accumulated at the basis of the internal capsule and zona incerta (Fig. 31, and finally fanned out in the ansa peduncularis and the inferior thalamic peduncle (Fig. 3D,E). As it ascended through the diencephalon, the 5-HT fiber bundle became progressively thinner because fiber fascicles detached themselves from it at different caudorostral levels and swept dorsally to innervate the various thalamic nuclei. Other fibers curved laterally to reach several components of the basal ganglia (see Lavoie and Parent, '90). At caudal diencephaliclevels, small fiber fasciclesleft the main bundle and traversed the pretectal area to arborize within the most caudal components of the lateral nuclear group, the nucleus limitans and the caudal portion of the habenula (Fig. lA,B). More rostrally the main bundle coursed along the ventral surface of the thalamus and fanned out into a medial and a lateral radiation. The medial radiation ascended partly within the habenulo-interpedunculartract to innervate the rostral portion of the habenula, whereas other fibers coursed slightly more laterally to arborize in the thalamic nuclei located medial to the internal medullary lamina as well as within the lamina itself. The lateral radiation ran along the dorsal surface of the substantia nigra and invaded the reticular nucleus and the lateral and medial geniculate nuclei, as well as other laterally located thalamic nuclei (Fig. 1C). More rostrally the bulk of 5-HT fibers was located within the H-fields of Forel. The major bundle broke up into several smaller fascicles traversing the H field, whereas more elongated fiber fascicles coursed laterally along the HI and H, fields en route to the reticular nucleus and the rostral pole of the lateral geniculate nucleus (Fig. 2). Other fibers left these long fascicles and ascended toward various thalamic nuclei located more dorsomedially.

In the rostral portion of the thalamus, 5-HT fibers abounded at the basis of the internal capsule and the zona incerta. Here again these fibers formed two divergent fascicles, one reaching the medially located nuclei and the other the laterally located ones (Fig. 3C). The most rostral portion of the thalamus received its 5-HT innervation from fibers coursing within the inferior thalamic peduncle medially and from other fibers ascending along the lateral border of the thalamus (Fig. 3D,E). In addition t o the 5-HT innervation of the thalamus deriving from the main ascending 5-HT bundle described above, a contribution from 5-HT fibers belonging to the periventricular system (Parent et al., '81) has also been noted. These fibers arborized principally within the midline nuclei.

Nucleus limitans and habenula The nucleus limitans traces the limit between the dorsal thalamus and the pretectal area (Fig. 1A). Several fiber fascicles coursed along its dorsal and ventral surfaces. The nucleus itself received a dense 5-HT innervation consisting mostly of isolated and uniformly distributed varicosities. The habenula is commonly divided into a medial and a lateral nucleus. The density of 5-HT innervation was very weak in both of these nuclei (Fig. 1B). In fact, the habenula represented one of the most poorly innervated area of the thalamus. Only a few scattered varicose fibers occurred in the medial and lateral nuclei, except in the ventralmost portion of the lateral nucleus where fibers were slightly more numerous. These fibers reached the habenula ventrally by coursing along the habenulo-interpedunculartract and arborized mostly in the form of an elongated band that lay along the medial border of the lateral habenular nucleus (see Fig. 5A).

Lateral nuclear group This group includes the pulvinar, lateral posterior, and lateral dorsal nuclei. Both isolated varicosities and varicose fibers occurred in varying number in each of these nuclei. The pulvinar is commonly divided in medial, lateral, inferior, and oral parts. The 5-HT innervation of this nucleus ranged from weak to moderate. In the medial part of the pulvinar, 5-HT varicose fibers were more abundant medially along the internal medullary lamina than laterally (Fig. 1A). In the lateral part of the pulvinar, few isolated varicosities were present in its medial half compared to a larger number of isolated varicosities and varicose fibers in its lateral half. Most of these fibers lay parallel to the external medullary lamina and displayed a typical bandlike pattern. These bands were interrupted by fascicles of myelinated nonimmunoreactive fibers coursing perpendicularly to the lateral border of the thalamus and among which some 5-HT nonvaricose fibers were found (Fig. 1A). The innervation of the inferior part of the pulvinar was weak; a small number of isolated varicosities occurred dorsally within this portion of the nucleus. However, the major 5-HT component consisted of nonvaricose fibers that tra-

Figs. 1-3. Camera lucida drawings of transverse sections through the caudal (11, middle (21,and rostral(3) portions of the thalamus in the squirrel monkey illustrating the distribution of 5-HT-immunoreactive fibers (lines) and isolatedvaricosities (dots). In each figure the drawings are set out in a caudorostral order. The numbers in the parentheses indicate the approximate level of each section according to the atlas of Emmers and Akert ('63).

SEROTONINERGIC INNERVATION OF PRIMATE THALAMUS

Figure 1

AN3XVd 'V aNV 3IOAVT '8

9

SEROTONINERGIC INNERVATION OF PRIMATE THALAMUS

7

2 mm

Figure 3

B. LAVOIE AND A. PARENT

8

versed this part and coursed laterally among myelinated nonimmunopositive fibers to finally reach the reticular and the lateral geniculate nuclei (Fig. 5C). Numerous isolated varicosities and few varicose fibers were observed along the lateral border of the oral part of the pulvinar and their number decreased medially (Fig. 1B). In the lateral posterior nucleus, the 5-HT innervation varied from weak to dense. Isolated varicosities and varicose fibers were more abundant in the dorsolateral part of this nucleus and progressively decreased ventromedially (Figs. 1B,C, ZA,B; see also Fig. 5C). The fibers in the lateral third of the lateral posterior nucleus were more homogeneously distributed than those in the lateral part of the pulvinar. In the lateral dorsal nucleus, the innervation was weak and consisted of varicose fibers and few isolated varicosities (Figs. lC, 2, 3A,B).

Reticular nucleus The reticular nucleus lies lateral to the external medullary lamina along the internal capsule. Ventrally, this nucleus is continuous with the zona incerta. The 5-HT innervation of the reticular nucleus was heterogeneous and ranged from moderate to very dense. Caudally, the innervation of the reticular nucleus was very dense and isolated varicosities were organized in clusters, which were in register with aggregates of reticular neurons (Figs. 1, 2A; see also Fig. 5B). Thin varicose fibers oriented ventrodorsally linked these clusters. A few 5-HT nonvaricose fibers were also intertwined with the varicose fibers present inbetween the clusters. More rostrally, the number of immunoreactive fibers in the dorsal part of the nucleus was larger than in its ventral part (Figs. ZB,C, 3A-D, 4B). In the rostral pole of the reticular nucleus, the fibers were homogeneously distributed along the dorsoventral axis (Fig. 3E). Some 5-HT nonvaricose fibers coursed along the ventral and dorsal borders of the zona incerta, turned dorsally and continued their route within and along the medial and lateral parts of the reticular nucleus. A moderate number of isolated varicosities occurred in the zona incerta itself (Figs. ZB,C, 3A-D).

Medial nuclear group The mediodorsal thalamic nucleus is the essential component of this group. On the basis of cytoarchitectural characteristics, this nucleus can be divided into: (1) a magnocellular part, medially, (2) a parvocellular part, centrally, and (3) a paralamellar part (or pars multiformis), laterally. The 5-HT innervation of the mediodorsal nucleus was highly heterogeneous and ranged from weak to dense. In its caudalmost portion, two moderately dense clusters of isolated varicosities occurred dorsally and ventrally along the medial aspect of the nucleus (Figs. lA,B, 5D). The remaining portion of the nucleus was weakly innervated and contained mostly nonvaricose fibers at this level. In the largest rostrocaudal portion of the nucleus, the 5-HT innervation displayed three distinct patterns (Figs. lC, 2 , 3A). Medially, adjacent to the midline nuclei, the innervation varied from moderate to dense and consisted mostly of isolated varicosities that were more abundant ventrally. In contrast, the large central portion of the nucleus was weakly innervated and contained only varicose fibers. Laterally, near the internal medullary lamina, a moderate number of both varicose fibers and isolated varicosities was found. In the rostralmost portion of the mediodorsal nucleus, the innervation was more homogeneous and consisted of a moderate number of varicose fibers among which a few isolated varicosities were scattered (Fig. 3B,C).

Intralaminar nuclear group

The intralaminar nuclei are generally divided into a rostral and a caudal group. The nuclei included in the rostral group are the paracentral, centrolateral, and centromedial nuclei, whereas the caudal group contains essentially the centre median and parafascicular nuclei. The 5-HT innervation was weak to moderate in the caudal group and generally dense in the rostral group. Varicosities without clear intervaricose segments were the major component of the 5-HT innervation in the intralaminar nuclei. Thin varicose fibers were more often observed in the caudal group, particularly in regions where the innervation was weak. The innervation of the centre medianlparafascicular complex was heterogeneous. Caudally, fibers were more Lateral and medial geniculate nuclei numerous in the centre median nucleus, particularly in its In the dorsal lateral geniculate nucleus, the 5-HT inner- lateral part, than in the parafascicular nucleus (Fig. 1B). vation ranged from very weak ventromedially to moderate This mediolateral gradient was progressively inverted along dorsolaterally. Hence, the superficial layers of the lateral the caudorostral extend of this complex. Indeed, rostrally, geniculate nucleus were more densely innervated than the the parafascicular nucleus received a moderate innervation, deeper layers (Fig. 4A).In some regions the 5-HT varicose whereas only few 5-HT varicose fibers were observed in the fibers were heterogeneously distributed and closely sur- centre median nucleus (Figs. 2A, 6A,B). At the level of the rounded myelinated nonimmunoreactive fibers (Figs. 1, rostral intralaminar nuclei, the 5-HT innervation was ZA). In the ventral lateral geniculate nucleus, 5-HT fibers dense and generally homogeneous, except in the caudal half were numerous and homogeneously distributed (Figs. lC, of the paracentral nucleus where 5-HT isolated varicosities 2A). Fibers arborizing in both the dorsal and ventral were more abundant dorsally (Figs. lC, 2, 3A,B). In the components of the lateral geniculate nucleus entered dorso- rostralmost portion of the paracentral nucleus, isolated laterally from the reticular nucleus (Figs. 1, 5F) and varicosities closely surrounded nonimmunoreactive myelimedially from the inferior part of the pulvinar. They nated fiber fascicles (Fig.3D). formed a capsule surrounding the entire lateral geniculate Midline nuclear group nucleus complex. The medial geniculate nucleus received a This group is composed of the paraventricular, central, very weak innervation consisting of varicose fibers and isolated varicosities that were more abundant in its ventro- and reuniens nuclei localized along the midline, and the lateral part than in its dorsomedial part (Figs. 1A,B, 4A). parataenial and central superior lateral nuclei, which are The suprageniculate nucleus received a moderate innerva- more dorsolaterally located. The densest 5-HT innervation tion mostly composed of uniformly distributed isolated of the thalamus was observed in the nuclei located directly varicosities (Fig. lA,B). along the midline. These nuclei received a moderate to very

SEROTONINERGIC INNERVATION OF PRIMATE THALAMUS

Fig. 4. Low power, darkfield photomontages showing the distribution of 5-HT fibers and isolated varicosities in lateral and medial geniculate nuclei (A),and in different rostra1 thalamic nuclei, including the reticular, ventral anterior, anteromedial and central nuclei (B). Scale bar: 500 pm.

9

10

B. LAVOIE AND A. PARENT

Figure 5

SEROTONINERGIC INNERVATION OF PRIMATE THALAMUS dense innervation consisting of 5-HT isolated varicosities that were heterogeneously distributed. A few nonvaricose fibers oriented ventrodorsally were also observed along the lateral border of these nuclei. Caudally, the 5-HT innervation of the central nucleus was heterogeneous and ranged from dense to very dense. The superior portion of the central nucleus together with the overlyingparaventricular nucleus contained the highest number of 5-HT isolated varicosities. Because there was no clear delineation between the superior part of the central nucleus and the paraventricular nucleus in our material, both nuclei are illustrated as a single entity (Figs. lC, 2, 3). The innervation of the intermediate part of the central nucleus was less dense than that of the inferior and superior parts. Furthermore, the number of isolated varicosities was significantly greater in the caudal and rostral poles than in the middle portion of the intermediate part of the central nucleus (Figs. lC, 2, 3). The nucleus reuniens received a moderate to dense innervation. In its caudal part the innervation was lighter than that of the adjacent inferior part of the central nucleus, whereas more rostrally both nuclei received a similarly dense innervation (Figs. lC, 2,3A-D). The central superior lateral nucleus contained a moderate to high number of varicose fibers as well as a few isolated varicosities. In this nucleus the 5-HT innervation displayed slight mediolateral and caudorostral decreasing gradients (Figs. lC, 2, 3A-C). The region of the parataenial nucleus just beneath the stria medullaris received a weak innervation similar to that of the underlying mediodorsal nucleus. This structure does not form a distinct entity in the squirrel monkey (see Emmers and Akert, '63) and is not illustrated as such in the present study.

Ventral nuclear group This group comprises the ventral posterior, ventral lateral, and ventral anterior thalamic nuclei. At this level the 5-HT innervation varied from very weak to dense and arose from fibers coursing along the ventral border of the thalamus. Some of these long nonvaricose fibers ascended laterally along the external medullary lamina, whereas others penetrated more directly within the ventral anterior nucleus. In the ventral nuclear group, the 5-HT varicose fibers were generally more numerous along the internal and external medullary laminae and their distribution was highly heterogeneous. The ventral posterior nucleus is divided into lateral, medial, and inferior parts. The medial part is further separated into a lateral diffusely organized portion and a more compact medial region termed the pars compacta. The inferior part of the ventral posterior nucleus received a very

Fig. 5. Darkfield photomicrographs depicting some patterns of the 5-HT innervation in different thalamic nuclei. The relatively weak innervation of the habenula is illustrated in A, whereas the distribution of 5-HT fibers and isolated varicosities along the external medullary lamina in the reticular nucleus and in the lateral posterior nucleus are shown in B and C, respectively. Some clusters of isolated varicosities in the reticular nucleus can be seen in B. The two dense clusters of 5-HT fibers and isolated varicosities occurring in the caudalmost portion of the mediodorsal nucleus are illustrated in D, whereas the patterns of 5-HT innervation in the ventral anterior and anteromedial nuclei are shown in E and G, respectively. Some of the fibers coursing from the reticular nucleus dorsally to the lateral geniculate nucleus ventrally are illustrated in F. Scale bars: 200 +m in A,D, and F and 50 km in B,C,E, and G.

11

weak innervation composed mostly of varicose fibers, which were rather uniformly distributed (Figs. lC, 2, 3A). In the medial part, the 5-HT innervation was very weak laterally and moderate to dense medially, within the pars compacta (Fig. lC, 2). Varicose fibers and isolated varicosities in the lateral part of the ventral posterior nucleus were more numerous dorsolaterally and their number decreased ventromedially (Figs. lB, C, 2,3A). The ventral lateral nucleus received a weak to dense 5-HT innervation (Figs. 2, 3A-C). The varicose fibers were more abundant in the lateral part of this nucleus, where isolated varicosities were also observed, and their number decreased medially and caudally. The 5-HT innervation of the ventral anterior nucleus was homogeneous, very weak, and consisted mostly of thin varicose fibers (Fig. 5E). Isolated varicosities were noted exclusively in the rostral part of this nucleus, along the external medullary lamina (Fig. 3C-E). Although occurring in most thalamic nuclei, it is in the weakly innervated portions of the ventral nuclear group that the typical pericellular arrangements displayed by the varicose 5-HT fibers around nonimmunoreactive thalamic neurons were more easily distinguishable (Fig. 6C,D).

Anterior nuclear group This group includes the anterodorsal, anteroventral, and anteromedial thalamic nuclei. These nuclei received a moderate and homogeneous 5-HT innervation (Fig. 5G). Fibers entered mostly by the ventral tip of the anteromedial nucleus. Some 5-HT fibers coursing in the capsule bordering ventrolaterally the anterior nuclei were also seen to invade this group (Fig. 4B). The thin and varicose 5-HT fibers arborized throughout the anterior nuclear group but were more abundant in the anterodorsal nucleus than in the anteroventral nucleus. A significant number of isolated varicosities were also scattered among the varicose fibers, particularly within the anterodorsal nucleus (Fig. 3).

DISCUSSION The present study has provided the first overall description of the 5-HT innervation of the thalamus in a primate. In the squirrel monkey the 5-HT innervation of the thalamus was more massive than would have been expected from earlier studies in nonprimate species. This finding is particularly significant since most nonprimate investigations were undertaken in animals pretreated with monoamine oxydase inhibitors plus tryptophan to increase endogenous 5-HT levels (see Steinbusch, '81), whereas the monkeys used in the present study were not pharmacologically manipulated. The density of the thalamic innervation in Saimiri was comparable to, and in certain nuclei even greater than, that observed at the level of the striatum (Lavoie and Parent, '90). Overall, the so-called nonspecific thalamic nuclei, comprising the intralaminar and the midline nuclei, received the heaviest innervation. The reticular nucleus also contained a significant amount of 5-HT varicose fibers and axon terminals. By comparison, the more specific relay nuclei and the association nuclei were less densely innervated. Our results are here compared to findings gathered in different species with various techniques, with a special attention to the origin and course of the 5-HT thalamic afferents and to their patterns of arborization in the different thalamic nuclei.

B. LAVOIE AND A. PARENT

12

Fig. 6 . Lightfield photomicrographs comparing the 5-HT innervation of the centre median (A)and parafascicular (B)nuclei at the middle rostrocaudal level of this complex. Some 5-HT-immunoreactive fibers forming pericellular contacts with nonimmunoreactive thalamic neu-

rons in the ventral anterior and ventral lateral nuclei are shown in C and D, respectively. Scale bars: 60 pm in A and B, and 30 pm in C and

D.

SEROTONINERGIC INNERVATION OF PRIMATE THALAMUS

Cellular origin and trajectory of 5-HT thalamic afferents Earlier anterograde tracing studies in rats (Conrad et al., '74; Moore et al., '78; Moore, '81; Peschanski and Besson, '84; Tork, '85) and cats (Bobillier et al., '75, '76) have indicated that both the dorsal and median raphe nuclei are the major sources of the thalamic innervation. As suggested by Brodal et al. ('60), the most caudally located raphe nuclei, principally nuclei raphe pallidus and raphe magnus, also provide inputs to the thalamus (Bobillier et al., '76; Taber-Pierce et al., '76; Azmitia and Segal, '78; Peschanski and Besson, '84). These caudal raphe nuclei appear to innervate more especially the rostral intralaminar nuclei (Bobillier et al., '76). It is important to note, however, that the 5-HT nature of these thalamic d e r e n t s cannot be established beyond doubt because these tract-tracing studies were undertaken with nonspecific anterograde or retrograde tracers. In addition, the raphe nuclei are not chemically homogeneous entities; they contain a significant proportion of nonserotoninergic neurons, including dopaminergic cells that have been detected principally in the dorsal raphe nucleus (Steinbusch et al., '80; Descarries et al., '86). This further complicates the interpretation of the results of axonal transport studies of raphe nuclei efferents that were undertaken without combination with 5-HT immunohistochemistry . In the squirrel monkey the 5-HT-immunoreactive neurons of the dorsal and median (central superior) raphe nuclei were seen to contribute heavily to the formation of the major fiber bundle ascending toward the forebrain. However, because of the close intermingling of the ascending 5-HT-positive fibers at midbrain level, it was not possible to determine the specific cellular origin of the 5-HT fibers that innervate the thalamus, This could only be achieved by combining the use of retrograde tracers with 5-HT immunohistochemistry. The trajectories of the 5-HT fibers innervating the thalamus of the squirrel monkey appear very similar to the ascending fiber pathways traced after tritiated amino acid injections into the midbrain raphe nuclei of the rat (Conrad et al., '74; Azmitia and Segal, '78; Moore et al., '78; Moore, '81; Tork, '85) and the cat (Bobillier et al., '76; Taber-Pierce et al.,'76). In both of these species, labeled fibers originating in the dorsal and median raphe nuclei formed a prominent bundle that ascended through the ventral portion of the midbrain tegmentum. From there some fibers coursed along the habenulo-interpeduncular tract to reach the habenula. A larger component of fibers swept laterally within the internal medullary lamina to innervate several thalamic nuclei (Bobillier et al., '76; Moore et al., '78; Moore, '81). In the cat, labeled fibers were traced more rostrally within the H-fields of Fore1 where, as in the squirrel monkey, they spread into a medial and a lateral component innervating the forebrain structures located medially and laterally, respectively (Bobillier et al., '76). Labeled fibers were also encountered more rostrally in the ansa lenticularis and ansa peduncularis through which they reached the basal ganglia and the septal region. Histofluorescence and immunohistochemical studies in monkeys (Schofieldand Everritt, '81; Schofield and Dixon, '82; Felten and Sladek, '83; Azmitia and Gannon, '86; Lavoie and Parent, '90) revealed that the overall organization of the ascending 5-HT fiber systems in primates is similar to that in rats and cats. These studies have not

13

provided detailed information regarding the 5-HT innervation of the thalamus.

Patterns of 5-HT innervation of thalamic nuclear groups Nucleus limitans and habenula The nucleus limitans has been well characterized in monkeys and cats but has not been identified as such in rats. In the cat, no labeled axon terminals were noted in this nucleus after tritiated amino acid injections in the midbrain raphe nuclei, despite that a significant number of positive terminals occurred in the adjoining pretectal area (Bobillier et al., '76). Immunohistochemical studies in rodents have shown that the entire pretectal area was weakly innervated (Steinbusch, '81; Cropper et al., '84),except for a small region overlying the medial pretectal area, which displayed an intense 5-HT immunoreactivity (Steinbusch, '81) and may correspond, at least in part, to the nucleus limitans. In the squirrel monkey the nucleus limitans displayed a dense 5-HT innervation as shown in the present investigation. In other studies involving primates, the 5-HT innervation of this caudalmost region of the thalamus has not been analyzed. In the habenula of the rat, moderate concentrations of 5-HT were biochemically detected in the medial nucleus compared to low concentrations in the lateral nucleus (Saavedra, '77). These data confirmed the earlier histofluorescence findings of Fuxe ('65). Likewise, immunohistochemical study in the rat showed a moderate number of 5-HT fibers in the medial habenular nucleus, whereas the lateral habenular nucleus appeared virtually devoid of such fibers (Steinbusch, '81). In the squirrel monkey, the 5-HT innervation of the habenula was overall very weak, except for a relatively dense band of isolated varicosities aligned along the medial border of the lateral habenular nucleus and some varicose fibers present in the ventralmost portion of the lateral nucleus. Anterograde tracing studies in rats (Azmitia and Segal, '78) and cats (Bobillier et al., '76) revealed that the lateral habenular nucleus is innervated by both the dorsal and median raphe nuclei, whereas the medial habenular nucleus is afferented mainly by the median raphe nucleus. Lateral nuclear group. The histofluorescenceinvestigation of Fuxe ('65) showed that the nuclei of the lateral thalamic group in the rat contained only a few scattered 5-HT varicosities and this was confirmed by the immunohistochemical study of Steinbusch ('81). In a more detailed immunohistochemical study of the 5-HT innervation of the rodent thalamus, the lateral dorsal nucleus was reported t o be well innervated, except in its posterior part, which received only a moderate 5-HT input (Cropper et al., '84). Interestingly, the ventral part of the lateral dorsal nucleus appeared more densely innervated than its dorsal part, as it is the case in the squirrel monkey. In the same study the so-called posterior complex, which appears to correspond, at least in part, to the pulvinar in primates, was shown to be moderately well innervated. An immunohistochemical study of the monoaminergic innervation of thalamic visual nuclei in cynomolgus and squirrel monkeys showed that the pulvinar complex exhibited a fairly uniform and moderately dense 5-HT innervation (Morrison and Foote, '86). In the present study the 5-HT innervation of the nuclei of the lateral group in the squirrel monkey was found to be much more heterogeneous and ranged from weak to dense. In the pulvinar and the

14 lateral posterior nucleus, the 5-HT innervation was denser in the laterodorsal region of these structures. Very little is known of the cellular origin of the 5-HT innervation of the lateral nuclear group. In the rat the lateral dorsal nucleus was reported to be innervated by the dorsal raphe nucleus by Azmitia and Segal ('78). In a more recent study, both the lateral dorsal and lateral posterior nuclei were said to receive inputs from the caudal part of the dorsal raphe nucleus (Peschanski and Besson, '84). As far as we know, no information is available on the origin of the 5-HT innervation to this region of the thalamus in. primates. Reticular nucleus. In the rat only a few scattered fluorescent 5-HT axon terminals were visualized in the reticular nucleus (Fuxe, '65). This finding is in agreement with more recent immunohistochemical studies, which described this nucleus as being poorly innervated by 5-HT axons (Steinbusch, '81; Cropper et al., '84). However, the rostra1 pole of the reticular nucleus has been shown to contain a moderate number of 5-HT fibers and isolated varicosities in rodents (Cropper et al., '84). In contrast, the present findings in Saimiri as well as those of Morrison and Foote ('86) in both macaques and squirrel monkeys revealed dense patches of 5-HT-immunoreactive terminals throughout the rostrocaudal extent of the reticular nucleus. These patches were in register with reticular neuron clusters and appeared composed of numerous axon terminals in primates. Very little information is available on the cellular origin of the 5-HT innervation of the reticular thalamic nucleus. In the cat, the nucleus raphe pontis has been suggested as a possible source of innervation (Bobillier et al., '76), whereas in the rat, the dorsal raphe nucleus has been shown to contribute to the innervation of the reticular nucleus (Hallanger et al., '87). Lateral and medial geniculate nuclei. In the rat, 5-HT concentrations were found to be higher in the lateral than in the medial geniculate nucleus (Saavedra, '77). Histofluorescence data in rodents (Fuxe, '65) indicated that the number of 5-HT axon terminals varied from low to moderate in the dorsal lateral geniculate nucleus and was uniformly low in the ventral geniculate nucleus. By comparison, only scattered 5-HT fluorescent varicosities were encountered in the medial geniculate nucleus. In the immunohistochemical study of Steinbusch ('811, the density of the 5-HT innervation was reported to be low in the medial geniculate nucleus, medium to high in the dorsal lateral geniculate nucleus, and medium to very high in the suprageniculate nucleus. Cropper et al. ('84) showed that the 5-HT immunoreactivity was dense in the ventral lateral geniculate nucleus, moderate in the dorsal lateral geniculate nucleus, and very weak in the medial geniculate nucleus. In the cat, 5-HT immunoreactive fibers were more abundant in the parvicellular or superficial layers than in the magnocellular or deep layers of the lateral geniculate nucleus (Fitzpatrick et al., '891, whereas in cynomolgus and squirrel monkeys the 5-HT immunoreactive fibers appeared rather uniformly distributed with only a slight dorsoventral increasing gradient (Morrison and Foote '86). However, the density of the 5-HT innervation of the lateral geniculate nucleus was reportedly more uniform in the squirrel monkey than in the cynomolgus monkey, where the laminar distribution of the 5-HT varicosities was more

B. LAVOIE AND A. PARENT differentiated with a slightly higher density in the magnocellular than in the parvicellular layers. In the present study the medial geniculate nucleus appeared less densely innervated than the lateral geniculate nucleus. At variance with the findings of Morrison and Foote ('861, the deep layers of the lateral geniculate nucleus in Saimiri were found to contain less 5-HT varicosities than the superficial layers. In the same species the 5-HT innervation of the suprageniculate nucleus was less dense than that of the corresponding structure in the rat (Steinbusch, '81).However, because of marked species differences in the relative size and topographical position of the various components of the lateral geniculate complex, the interspecific variations in the 5-HT innervation of this complex must be interpreted with caution. Anterograde tracing studies in the rat (Azmitia and Segal, '78; Pasquier and Villar, '82) and the cat (Bobillier et al., '76) pointed to the dorsal raphe nucleus as the major source of afferents to the lateral geniculatenucleus. Furthermore, retrogradely labeled cells were encountered in the dorsal raphe nucleus of the rat after WGA-HRP injections in lateral and/or medial geniculate nuclei (Ahlsen and Lo, '82; Mackay-Sim et al., '83; Ahlsen, '84; Hallanger et al., '87; Luth and Seidel, '87). Studies combining retrograde fluorescent tracers with 5-HT immunofluorescence in the cat have confirmed the projection from the dorsal raphe nucleus to the lateral geniculate nucleus and further indicated that the median (central superior) raphe nucleus also contributes to this innervation. (De Lima and Singer, '87a). Interestingly, a similar double-labelingstudy in the rat has revealed the existence of dorsal raphe neurons with axons sending collaterals to both the lateral geniculate nucleus and the superior colliculus (Villar et al., '88). Since both of these structures are well-known, relay nuclei in visual pathways, this finding indicates that 5-HT is strongly implicated in the modulation of visual information (see below). Medial nuclear group. Immunohistochemical studies reported that the rat mediodorsal nucleus contained only a few 5-HT-positive fibers (Steinbusch, '81; Cropper et al., '84). By comparison, the mediodorsal nucleus in the squirrel monkey displayed a pattern of 5-HT immunoreactivity that was dense and very heterogeneous. The central portion of the nucleus was weakly innervated, whereas its medial portion received a dense innervation. The lateral portion of the nucleus contained a moderate number of 5-HT fibers and Varicosities. Although the 5-HT innervation of the mediodorsal nucleus was organized according to three distinct patterns mediolaterally, the exact correspondence between the distribution of the 5-HT-IR profiles and the three cytoarchitectonic divisions of the mediodorsal nucleus remained to be established. The origin of the 5-HT innervation of the mediodorsal nucleus has been studied with anterograde tracing methods in both rats and cats (Bobillieret al., '76; Azmitia and Segal, '78; Peschanski and Besson, '84). These studies suggested that the median raphe nucleus is a major source of innervation of the mediodorsal nucleus. Retrograde cell labeling study in the rat confirmed these results (Hallanger et al., '87). In macaque monkeys, however, retrogradely labeled cells were encountered in the dorsal raphe nucleus after WGA-HRP injections into the magnocellular division of the mediodorsal nucleus (Russchen et al., '87). Similarly, injections of fluorescence tracers in the mediodorsal nucleus of the rat resulted in retrograde labeling of neurons in the

SEROTONINERGIC INNERVATION OF PRIMATE THALAMUS lateral wing of the dorsal raphe nucleus and surrounding central gray (De Olmos and Heimer, '80). Interestingly, some of these neurons appeared to send axon collateral to different non-thalamic structures. Intralaminar nuclear group. In the rat, the intralaminar nuclei, including the centrolateral, centromedial, paracentral, and parafascicular nuclei, were reported to contain very few 5-HT-immunoreactive profiles, except the caudal part of the parafascicular nucleus, which appeared moderately innervated (Steinbusch, '81). In contrast, in the study of Crooper et al., ('84), the anterior intralaminar nuclei in rodents appeared well labeled, although less densely innervated than the midline nuclei. In the same study the parafascicular nucleus was shown to contain numerous 5-HT-immunoreactive fibers that were less homogeneously distributed than in the adjoining midline nuclei. In the squirrel monkey the 5-HT innervation was dense in the anterior intralaminar nuclei, but only weak to moderate in the posterior group. In the centre medianlparafascicular complex, the 5-HT fibers and varicosities were distributed according to a mediolateral increasing gradient caudally, whereas this gradient was progressively reversed rostrally. Here again the interpretation of interspecific variations in the 5-HT thalamic innervation is hindered by marked species differences in the relative development of thalamic nuclei. For instance, a centre median is not recognized as such in rodents where the entire posterior intralaminar complex is believed to be solely composed of the parafascicular nucleus. In contrast, the centre median in primates is thought to be greatly enlarged by comparison to the parafascicular nucleus. Anterograde labeling studies in the rat (Peschanski and Besson, '84) and the cat (Bobillier et al., '76) have revealed that the intralaminar nuclei as a whole received afferents from virtually all raphe nuclei, including the most caudally located ones, such as the raphe pallidus and raphe magnus nuclei. The dorsal raphe nucleus appeared to project more heavily to the parafascicular nucleus than to other intralaminar nuclei. In the cat, retrogradely labeled cells were visualized in the dorsal raphe nucleus after WGA-HRP injections into various portions of the intralaminar nuclei (Kaufman and Rosenquist, '85). Midline nuclear group. The midline nuclei in the rat received a relatively dense and uniform 5-HT innervation as determined by immunohistochemistry (Steinbusch, '81; Cropper et al., '84). Individual fibers were difficult to visualize in these nuclei, which contained a multitude of isolated varicosities. In one immunohistochemical study (Steinbusch, '811, the nucleus reuniens was described as displaying a rostrocaudal decreasing gradient and the paraventricular nucleus appeared much more intensely innervated than the other midline nuclei. These findings largely confirmed previous observations made with the histofluorescence method (Fuxe, '65; Aghajanian et al., '73). In the present study, the midline nuclei in Saimiri were the most densely innervated nuclei of the thalamus. This innervation consisted mostly of dense terminal fields composed of innumerable isolated varicosities among which some thin varicose fibers were scattered. The densest of these terminal fields were encountered in the dorsal portion of the midline nuclei. Results from anterograde and retrograde labeling studies revealed that the midline nuclei of rats and cats received a dense innervation deriving from both the dorsal and median raphe nuclei (Bobillier et al., '76; Moore et al., '78;

15

Moore, '81). The more caudally located raphe nuclei also appeared to contribute to the innervation of the midline nuclei in both rats and cats (Bobillier et al., '76; Peschanski and Besson, '84). Ventral nuclear group. Immunohistochemical studies in the rat showed that, except for the ventromedial nucleus, which is moderately innervated, the ventral nuclear group contained only a few labeled fibers (Steinbusch, '81; Cropper et al., '84). Tract-tracing studies revealed that the dorsal and median raphe nuclei and, to a lesser extent, the nucleus raphe pontis, contributed to the innervation of the ventral nuclear group in rats and cats (Bobillier et al., '76; Moore et al., '78; Moore, '81; Peschanski and Besson, '84; Hallanger et al., '87). These findings were confirmed with more specific investigations in the rat combining the use of retrograde tracers with 5-HT immunohistochemistry (Consolazione et al., '84). In the squirrel monkey, the 5-HT innervation of the ventral nuclear group is weak but denser in its dorsolateral portion. In the cynomolgus monkey, the lateral part of the ventral posterior nucleus received a sparse 5-HT innervation (Westlund et al., '90). Thin immunoreactive fibers were seen closely associated with cell clusters characteristic of this nucleus. I n the same study, retrogradely labeled 5-HT cells were found in the dorsal, median (central superior), and pontis raphe nuclei after WGA-HRP injection in the lateral part of the ventral posterior nucleus (Westlund et al., '90). Anterior nuclear group. A small number of 5-HT fluorescent varicosities were encountered in the anteroventral nucleus of rodents, whereas the anterodorsal nucleus contained only a few scattered varicosities (Fuxe, '65). A large number of 5-HT-immunoreactive profiles were found in the anteroventral nucleus compared to a small number in both the anterodorsal and anteromedial nuclei (Steinbusch, '81; Cropper et al., '84). Studies of the efferent projections of the midbrain raphe nuclei in the rat have demonstrated that both dorsal and median raphe nuclei project to the anterior nuclear group, with a slightly higher contribution of the median nucleus to the innervation of the anterodorsal nucleus (Azmitia and Segal, '78; Moore et al., '78; Moore, '81; Peschanski and Besson, '84; Hallanger et al., '87). In the squirrel monkey the 5-HT innervation of the anterior nuclear group was moderate and rather homogeneous, although a larger number of 5-HT fibers and isolated varicosities occurred in the anterodorsal nucleus.

Functional considerations Numerous behavioral and physiological roles have been attributed to 5-HT in the CNS. Indeed, changes in 5-HT levels may alter thermoregulation, sleep, pain, locomotor activity, sensory processing, and diurnal rhythms (see Azmitia and Gannon, '86; Glennon, '90). This indolamine has also been implicated in various human disorders, such as Parkinson and Huntington's diseases, depression, and schizophrenia (Dray, '81; Azmitia and Gannon, '86; Fallon and Loughlin, '87; Glennon, '90). Despite extensive knowledge of the overall anatomical and functional organization of the 5-HT system in the CNS, very little is known about the functions of this biogenic amine at thalamic level. The dense and highly heterogeneous 5-HT innervation of the thalamus observed in the squirrel monkey suggests a role of 5-HT in various specific thalamic functions. Furthermore, numerous 5-HT fibers were seen in close apposition with cell bodies located in different thalamic nuclei, which

16 indicates close pericellular interactions between thalamic neurons and 5-HT afferents. Indeed, electron microscopic studies have revealed typical synaptic contacts between 5-HT fibers and nonimmunoreactive neurons in the lateral geniculate nucleus of the rat (Papadopoulosand Parnavelas '90). In cats and monkeys, 5-HT-IR fibers were seen to display both junctional and nonjunctional appositions on lateral geniculate neurons (De Lima and Singer, '87b; Pasik et al., '88; Wilson and Hendrickson, '88). The presence of numerous 5-HT receptors heterogeneously distributed in the thalamus of rats and primates strongly supports the hypothesis of a specificrole of 5-HT in thalamic functions. The 5-HT receptors in the CNS have been divided into four subtypes: 5-HT,, 5-HT,, 5-HT3,and 5-HT4 (Glennon, '90). In most species the 5-HT receptors disclosed at thalamic level were mainly of the 5-HT1 subtype. Different subclasses of 5-HT1receptors have been described and termed 5-HT1,, 5-HT,,, 5-HT1,, 5-HT1,, and 5-HTl, (Glennon, '90). The distribution of the 5-HT,, receptor subclass at thalamic level has been described in detail in several species, and the presence of 5-HTl, and 5-HT1, receptor subclasses has also been reported in the thalamus (Pazos and Palacios, '85; Herrick-Davis and Titeler, '88; Molineaux et al., '89; Mengod et al., '90). In nonhuman primates, high levels of 5-HT1 binding sites were observed in most of the midline nuclei as well as in the parafascicular nucleus, whereas moderate levels are noted in the central and anteroventral nuclei and in the ventral part of the lateral geniculate nucleus. The remaining thalamic nuclei displayed low binding levels (Stuart et al., '86). In the human a similar distribution of 5-HT1 receptors, mainly 5-HT,, subclass, was observed (Pazos et al., '87a). Only a small number of 5-HT, receptors occurred in the thalamus, except in the parafascicular nucleus where an intermediate concentration of this receptor subtype was noted in rats and humans (Pazos et al., '85, '87b). The levels of 5-HT, receptors were reportedly low in the entire thalamus (Kilpatrick et al., '87; Waeber et al., '88). The involvement of 5-HT in pain modulation is one of the most well documented roles of this indolamine at thalamic level. Several studies in the rat have demonstrated that stimulation of the dorsal raphe nucleus attenuates the increase in the firing rate of the parafascicular neurons induced by noxious stimuli. In 5-HT-depletedrats, both the stimulation of the dorsal raphe nucleus and noxious stimuli elicited an increase in the firing rate of the parafascicular neurons. The fact that this effect was not additive indicates that the serotoninergic projection to the parafascicular nucleus has a tonic inhibitory influence on the response to noxious stimuli (Andersen and Dafny, '83). Likewise, iontophoretic applicationof 5-HT upon neurons of the parafascicular nucleus reduced their response t o noxious stimuli (Andersen and Dafny, '82). Furthermore, two classes of neurons have been identified in the parafascicular nucleus: those whose spontaneous activity is either increased or decreased by noxious stimuli and respectivelytermed "nociceptive-on" and "nociceptive-off' cells by Dafny et al. ('90). The same authors showed that the concomitant application of noxious stimuli and electrical or chemical stimulation of the dorsal raphe neurons suppressed the excitation induced by noxious stimuli on nociceptive-on cells. The role of 5-HT in the processing of visual information in the thalamus has also been an area of active research, but the exact action of 5-HT at this level is still unclear. For instance, several electrophysiological studies in rats and

B. LAVOIE AND A. PARENT cats, involving either iontophoretic 5-HT application or electrical stimulation of the dorsal raphe nucleus, revealed that 5-HT decreases both spontaneous and evoked neuronal activities in the dorsal lateral geniculate nucleus (Rogawaski and Aghajanian, '80; Kemp et al., '82; Yoshida et al., '84; Marks et al.,'87; Kayama et al., '89). However, facilitatory effects have also been observed after electrical stimulation of the dorsal raphe nucleus in the cat (Foote et al., '74). These data are difficult to interpret because the lateral geniculate nucleus contains GABAergic inhibitory interneurons (Fitzpatrick et al., '84; Montero and Zempel, '85; Mugnaini and Oertel, '85). Electron microscopic investigations have shown that 5-HT terminals make synapses on both relay cells and interneurons in the lateral geniculate nucleus (Papadopoulos and Parnavelas, '90). Therefore, the action of 5-HT at this level could be the result of a direct effect on relay cells and/or indirect action on GABAergic inhibitory interneurons. The action of 5-HT at thalamic level has also been studied in the ventrobasal nucleus, which is devoid of GABAergic interneurons (Mugnaini and Oertel, '85). In 5-HT-depleted cats, slow fluctuations of neuronal activities were extracelMarly recorded from neurons of the ventrobasal nucleus by Kodama et al. ('89). In such cases, the administration of 5-HT agonists was shown t o suppress these slow fluctuations. However, it must noted that the modulatory effect of 5-HT in this nucleus appeared to be dose-dependent. For instance, low doses of 5-HT produced a marked facilitation of responses to excitatory amino acids, whereas higher doses reversed the facilitation or inhibited the excitatory amino acid responses and synaptic transmission (Eaton and Salt, '89). Interesting information has been obtained on the possible role of 5-HT at thalamic level by in vitro experiments. For instance, it has been shown that 5-HT reduces the ability of neurons to generate burst firing in lateral geniculate nucleus (Pape and McCormick, '89). A similar effect has been observed in the reticular and suprageniculate nuclei (McCormickand Wang, '91). In the latter study, local 5-HT application resulted in a pronounced and prolonged neuronal excitation associated with the occurrence of single spike activity. This excitation was specifically blocked by 5-HT, and 5-HT1, antagonists and appeared to be mediated through a decrease in potassium current. The same study has shown that noradrenaline produces a similar excitation and that both noradrenaline and 5-HT converge onto the same potassium conductance. In contrast, application of acetylcholineon the same neurons resulted in a brief period of inhibition (McCormickand Wang, '91). Taken together, the latter results indicate that serotonin exerts a specific excitation on thalamic neurons and that this neurotransmitter is also involved in complex interactions with other neurotransmitters, particularly noradrenaline and acetylcholine,that are used by well-characterized brainstem ascending systems.

ACKNOWLEDGMENTS We thank Carole Harvey and Lisette Bertrand for technical assistance, and Suzanne Bilodeau for typing the manuscript. This investigation was supported by grant MT-5781 from the Medical Research Council of Canada (MRC) to A.P. The financial support of the FRSQ and FCAR is also acknowledged.B. L. was the recipient of a Studentship from the MRC.

SEROTONINERGIC INNERVATION OF PRIMATE THALAMUS

LITERATURE CITED Aghajanian, G.K., M.J. Kuhar, and R.H. Roth (1973) Serotonin-containing neuronal perikarya and terminals: Differential effects of p-chlorophenylalanine. Brain Res. 54.85-101. Ahlsen, G. (1984) Brain stem neurones with differential projections to functional subregions of the dorsal lateral geniculate complex in the cat. Neuroscience 12.81 7-838. Ahlsen, G., and F.S. Lo (1982) Projections of the brain stem neurons to the perigeniculate nucleus and the lateral nucleus in the cat. Brain Res. 238:433438. Andersen, E., and N. Dafny (1982) Microiontophoretically applied 5-HT reduces responses to noxious stimulation in the thalamus. Brain Res. 241: 176-178. Andersen, E., and N. Dafny (1983) An ascending serotonergic pain modulation pathway from the dorsal raphe nucleus to the parafascicularis nucleus of the thalamus. Brain Res. 269:57-67. Azmitia, E.C., and P.J. Gannon (1986) The primate serotoninergic system: A review of human and animal studies and a report on Mucuca fusciculuris. I n S. Fahn (ed): Advances in Neurolom, Vol. 43. New York: Raven Press, pp. 407468. Azmitia, E.C., and M. Segal (1978) An autoradiographic analysis of the differential ascending projections of the dorsal and median raphe nuclei in the rat. J. Comp. Neurol. 179:641-659. Bjorklund, A., and 0.Lindvall(1984) Dopamine-containing cells in the CNS. In A. Bjorklund and T. Hokfelt (eds):Handbook of Chemical Neuroanatomy. Vol. 2, Part 1. Amsterdam: Elsevier, pp. 55-122. Bobillier, P., F. Petitjean, D. Salvert, M. Ligier, and S. Seguin (1975) Differential projections of the nucleus raphe dorsalis and nucleus raphe centralis as revealed by autoradiography. Brain Res. 85205-210. Bobillier, P., S. Seguin, F. Petitjean, D. Salvert, M. Touret, and M. Jouvet (1976) The raphe nuclei of the cat brain stem: A topographical atlas of their efferent projections as revealed by autoradiography. Brain Res. 113r449486. Brodal, A,, E. Taber, and F. Walberg (1960) The raphe nuclei of the brain stem in the cat. 11. Efferent connections. J. Comp. Neurol. 114:239-259. Chan-Palay, V. (1977) Indoleamine neurons and their processes in the normal rat brain and in chronic diet-inducedthiamine deficiency demonstrated by uptake of 3H-serotonin.J. Comp. Neurol. 176t467-494. Conrad, L.C.A., C.M. Leonard, and D.W. PfaE (1974) Connections of the median and dorsal raphe nuclei in the r a t An autoradiographic and degeneration study. J. Comp. Neurol. 156:179-206. Consolazione, A,, J.V. Priestley, and A.C. Cuello (1984) Serotonincontaining projections to the thalamus in the rat revealed by a horseradish peroxidase and peroxidase antiperoxidase double-staining technique. Brain Res. 3.22:233-243. Cropper, E.C., J.S. Eisenman, and E.C. Azmitia (1984) An immunocytochemical study of the serotoninergic innervation of the thalamus of the rat. J. Comp. Neurol. 224r38-50. Dafny, N., C. Reyes-Vazquez, and J.T. Qiao (1990) Modification o f nociceptively identified neurons in thalamic parafascicularis by chemical stimulation of the dorsal raphe with glutamate, morphine, serotonin and focal dorsal raphe electrical stimulation. Brain Res. Bull. 24:717-723. Dahlstrom, A., and K. Fuxe (1964) Evidence for the existence of monoaminecontaining neurons in the central nervous system. I. Demonstration of monoamines in the cell bodies of the brain stem neurons. Acta Physiol. Scand. 62(Suppl. 232):l-55. De Lima, A.D., and W. Singer (1987a) The brainstem projection to the lateral geniculate nucleus in the cat: Identification of cholinergic and monoaminergic elements. J. Comp. Neurol. 259:92-121. De Lima, A.D., and W. Singer (198713)The serotoninergic fibers in the dorsal lateral geniculate nucleus of the cat: Distribution and synaptic connections demonstrated with immunocytochemistry. J. Comp. Neurol. 258; 339-35 1. De Olmos, J., and L. Heimer (1980) Double and triple labeling of neurons with fluorescent substances, the study of collateral pathways in the ascending raphe system. Neurosci. Lett. 19:7-12. Descarries, L., F. Berthelot, S. Garcia, and A. Beaudet (1986) Dopaminergic projection from nucleus raphe dorsalis to neostriatum in the rat. J. Comp. Neurol. 249.511-520. Dray, A. (1981) Serotonin in the basal ganglia: Functions and interactions with other neuron pathways. J. Physiol. (Paris) 77r393403. Eaton, S.A., and T.E. Salt (1989) Modulatory effects of serotonin on excitatoiy amino acid responses and sensory synaptic transmission in the ventrobasal thalamus. Neuroscience 33:285-292.

17

Emmers, R., and K. Akert (1963) A stereotaxic atlas of the brain of the squirrel monkey (Suimiri sciureus). Madison: University of Wisconsin Press. Falck, B., N.A. Hillarp, G. Thieme, and A. Torp (1962) Fluorescence of catecholamines and related compounds condensed with formaldehyde. J. Histochem. Cytochem. lOr348-354. Fallon, J.H., and S.E. Loughlin (1987) Monoamine innervation of cerebral cortex and a theory of the role of monoamines in cerebral cortex and basal ganglia. In E.G. Jones and A. Peters (eds): Cerebral Cortex, Vol. 6. New York: Plenum Press, pp. 41-126. Felten, D.L., and J.R. Sladek J r (1983) Monoamine distribution in primate brain. V. Monoaminergic nuclei: Anatomy, pathways and local organization. Brain Res. Bull. 10:171-284. Fitzpatrick, D., I.T. Diamond, and D. Raczkowski (1989) Cholinergic and monoaminergic innervation of the cat's thalamus: Comparison of the lateral geniculate nucleus with other principal sensory nuclei. J. Comp. Neurol. 288:647-675. Fitzpatrick, D., G.R. Penny, and D.E. Schmechel (1984) Glutamic acid decarboxylase-immunoreactive neurons and terminals in the lateral geniculate nucleus in the cat. J. Neurosci. 4t1809-1829. Foote, W.E., R.J. Maciewicz, and J.P. Mordes (1974) Effects of midbrain raphe and lateral mesencephalic stimulation on spontaneous and evoked activity in the lateral geniculate of the cat. Exp. Brain Res. 19:124-130. Fuxe, K. (1965) The distribution of monoamines terminals in the central nervous system. Acta Physiol. Scand. 64(Suppl. 247):37-85. Glennon, R.A. (1990) Serotonin receptors: Clinical implications. Neurosci. Biobehav. Rev. 14.35-47. Hallanger, A.E., A.I. Levey, H.J. Lee, D.B. Rye, and B.H. Wainer (1987) The origins of cholinergic and other subcortical afferents to the thalamus in the rat. J. Comp. Neurol. 262105-124. Herrick-Davis, K., and M. Titeler (1988) Detection and characterization of the serotonin 5-HT,, receptor in rat and human brain. J. Neurochem. 50t1624-1631. Jones, E.G. (1985) The Thalamus. New York: Plenum Press. Kaufman, E.F.S., and A.C. Rosenquist (1985) Afferent connections of the thalamic intralaminar nuclei in the cat. Brain Res. 335.281-296. Kayama, Y., S. Shimada, Y. Hishikawa, and T. Ogawa (1989) Effects of stimulating the dorsal raphe nucleus of the rat on neuronal activity in the dorsal lateral geniculate nucleus. Brain Res. 489:l-11. Kemp, J.A., H.C. Roberts, and A.M. Sillito (1982) Further studies on the action of 5-hydroxytryptaminein the dorsal lateral geniculate nucleus of the cat. Brain Res. 246:334-337. Kilpatrick, G.J., B.J. Jones, and M.B. Tyers (1987) Identification and distribution of 5-HT3 receptors in rat brain using radioligand binding. Brain Res. Bull. 13:l-31. Kodama, T., H. Mushiake, K. Shima, T. Hayashi, and M. Yamamoto (1989) Slow fluctuations of single unit activities of hippocampal and thalamic neurons in the cats. 11. Role of serotonin on the stability of neuronal activities. Brain Res. 487:3544. Lavoie, B., and A. Parent (1990) Immunohistochemical study of the serotoninergic innervation of the basal ganglia in the squirrel monkey. J. Comp. Neurol. 299.1-16. Lavoie, B., Y. Smith, and A. Parent (1989) Dopaminergic innervation o f the basal ganglia in the squirrel monkey as revealed by tyrosine hydroxylase immunohistochemistry.J. Comp. Neurol. 289r3652. Lindvall, O., A. Bjorklund, A. Nobin, and U. Steveni (1974) The adrenergic innervation of the rat thalamus as revealed by glyoxylic acid fluorescence method. J. Comp. Neurol. 154r317-348. Liith, H.J., and I. Seidel (1987) Immunohistochemische Characterisierung serotoninerger Afferenzen im visuellen System der Ratte. J. Hirnforsch. 28:591-600. McCormick, D.A., and Z. Wang (1991) Serotonin and noradrenaline excite GABAergic neurones of guinea pig and cat nucleus reticularis thalami. J. Physiol. (London)442.235-255. Mackay-Sim, A,, A.J. Sefton, and P.R. Martin (1983) Subcortical projections to the lateral geniculate and thalamic reticular nuclei in the hooded rat. J. Comp. Neurol. 213.24-35. Marks, G.A., S.G. Speciale, K. Cobbey, and H.P. Roffwarg (1987) Serotonergic inhibition of the dorsal lateral geniculate nucleus. Brain Res. 418:76-84. Mengod, G., H. Nguyen, H. Le, C. Waeber, H. Lubbert, and J.M. Palacios (1990) The distribution and cellular localization of the serotonin 1 C receptor mRNA in the rodent brain examined by in situ hybridization histochemistry. Comparison with receptor binding distribution. Neuroscience 35:577-591.

18 Molineaux, S.M., T.M. Jessel, R. Axel, and D. Julius (1989) 5-HTlc receptor is a prominent serotonin receptor subtype in the central nervous system. Proc. Natl. Acad. Sci. (USA)86t6793-6797. Montero, J.M., and J. Zempel (1985) Evidence for two types of GABAcontaining interneurons in the A-laminae of the cat lateral geniculate nucleus: A double-labeled HRP and GABA-immunocytochemical study. Exp. Brain Res. 60:603-609. Moore, R.Y. (1981) The anatomy of the central serotonin neuron systems in the rat brain. In: B.L. Jacobs and A. Gelperin (eds):Serotonin Neurotransmission and Behavior. Cambridge:MIT Press, pp 35-71. Moore, R.Y., and J.P. Card (1984) Noradrenaline-containing neuron systems. In A. Bjorklund and T. Hokfelt (eds): Handbook of Chemical Neuroanatomy, Vol. 2, Part 1.Amsterdam: Elsevier, pp. 123-156. Moore, R.Y., A.E. Halaris, and B.E. Jones (1978) Serotonin neurons of the midbrain raphe: Ascending projections. J. Comp. Neurol. 180:417-438. Morrison, J.H., and S.L. Foote (1986) Noradrenergic and serotoninergic innervation of cortical, thalamic, and tectal visual structures in old and new world monkeys. J. Comp. Neurol. 243.117-138, Mugnaini, E., and W.H. Oertel (1985) An atlas of the distribution of GABAergic neurons and terminals in the rat CNS as revealed by GAD immunohistochemistry. In A. Bjorklund and T. Hokfelt ( 4 s ) : Handbook of Chemical Neuroanatomy, Vol. 4, Part 1. Amsterdam: Elsevier, pp. 436-595. O’Hearn, E., G. Battaglia, E.B. De Souza, M.J. Kuhar, and M.E. Molliver (1988) Methylenedioxyamphetamine(MDA)and methylenedioxymethamphetamine (MDMA) cause selective ablation of serotoninergic axon terminals in forebrain: Immunocytochemicalevidence for neurotoxicity. J. Neurosci. 82788-2803. Olszewski, J. (1952) The Thalamus of Macaca mzdattu. New York: S. Karger. Papadopoulos, G.C., and J.G. Parnavelas (1990) Distribution and synaptic organization of serotoninergic and noradrenergic axons in the lateral geniculate nucleus of the rat. J. Comp. Neurol. 294:345-355. Pape, H.-C., and D.A. McCormick (1989) Noradrenaline and serotonin selectively modulate thalamic burst firing by enhancing a hyperpolarization-activated cation current. Nature 34Ot715-718. Park, D., Y. Smith, A. Parent, and M. Steriade (1988) Projections of brainstem core cholinergic and non-cholinergicneurons of cat to intralaminar and reticular thalamic nuclei. Neuroscience 25:69-86. Parent, A., L. Descarries, and A. Beaudet (1981) Organization of ascending serotonin systems in the adult rat brain. A radioautographic study after intraventricular administration of L3H1 5-hydrolcyt~yptamine.Neuroscience 6:115-138. Pasik, P., T. Pasik, and G.R. Holstein (1988) Serotonin-immunoreactivity in the monkey lateral geniculate nucleus. Exp. Brain Res. 69.662-666. Pasquier, D.A., and M.J. Villar (1982) Specific serotonergic projections to the lateral geniculate body from the lateral cell groups of the dorsal raphe. Brain Res. 249:142-146. Pazos, A,, and J.M. Palacios (1985) Quantitative autoradiographic mapping of serotonin receptors in the rat brain. I. Serotonin-1 receptors. Brain Res. 346:205-230. Pazos, A., R. Cartes, and J.M. Palacios (1985) Quantitative autoradiographic mapping of serotonin receptors in the rat brain. 11. Serotonin-2 receptors. Brain Res. 346.231-249, Pazos, A., A. Probst, and J.M. Palacios (1987a) Serotonin receptors in the human brain. 111. Autoradiographic mapping of serotonin-1 receptors. Neuroscience 21:97-122. Pazos, A,, A. Probst, and J.M. Palacios (1987b) Serotonin receptors in the human brain. IV. Autoradiographic mapping of serotonin-2 receptors. Neuroscience 21:123-139. Peschanski, M., and J.M. Besson (1984) Diencephalic connections of the raphe nuclei of the rat brainstem: An anatomical study with reference to the somatosensory system. J. Comp. Neurol. 224:509-534. Phillipson, O.T., and A.C. Griffith (1980) Dopamine neurons in the ventral tegmentum project to both medial and lateral habenula. Exp. Brain Res. 45:89-94. Ranadive, N.S., and A.H. Sehon (1967) Antibodies to serotonin. Can. J. Biochem. 45: 1701-1 710. Rogawski, M.A., and G.K. Aghajanian (1980) Norepinephrine and serotonin: Opposite effects on the activity of lateral geniculate neurons evoked by optic pathway stimulation. Exp. Neurol. 69:678-694.

B. LAVOIE AND A. PARENT Russehen, F.T., D.G. Amaral, and J.L. Prince 11987) The afferent input to the magnocellular division of the mediodorsal thalamic nucleus in the monkey, Macaca fascicularis. J. Comp. Neurol. 256:175-210. Saavedra, J.M. (1977) Distribution of serotonin and synthesizing enzymes in discrete areas of the brain. Fed. Proc. 362134-2141, Schofield, S.P., and B.J. Everitt (1981) The organization of indolamine neurons in the brain of the rhesus monkey (Macaca rnulatta).J. Comp. Neurol. 197:369-383. Schofield, S.P., and A.F. Dixon (1982) Distribution of catecholamine and indolamine neurons in the brain of the common marmoset (Callithrix jacchus).J. Anat. 134.315-338. Skagerberg, G., 0.Lindvall, and A. Bjorklund (1984) Origin, course and termination of the mesohabenular dopamine pathway in the rat. Brain Res. 307r99-108. Steinbusch, H.W.M. (1981) Distribution of serotonin-immunoreactivity in the central nervous system of the rat-ell bodies and terminals. Neuroscience 6:557-618. Steinbusch, H.W.M., D. Van der Kooy, A.A.J. Verhofstad, and A. Pellegrino (1980) Serotoninergic and non-serotoninergic projections from the nucleus raphe dorsalis to the caudate-putamen complex in the rat studied by a combined immunofluorescence and fluorescent retrograde axonal labeling technique. Neurosci. Lett. 19:137-142. Steinbusch, H.W.M., A.A.J. Verhofstad, and H.W.M. Joosten (1978)Localization of serotonin in the central nervous system by immunohistochemistry: description of a specific and sensitive technique and some applications. Neuroscience 32311-819. Steriade, M., and D. Biesold (1990) Brain cholinergic systems. Oxford Oxford University Press. Steriade, M., D. Pare, A. Parent, and Y. Smith (1988) Projections of cholinergic and non-cholinergic neurons of the brainstem core to relay and associational thalamic nuclei in the cat and macaque monkey. Neuroscience%5:47-67. Sternberger, L.A. (1986) Immunocytochemistry. New York: John Wiley & Sons. Stuart, A.M., I.J. Mitchell, P. Slater, H.L.P. Unwin, and A.R. Crossman (1986) A semi-quantitative atlas of 5-hydrolcytryptamine-1receptors in the primate brain. Neuroscience 18:619-639. Taber Pierce, E., W.E. Foote, and J.A. Hobson (1976) The efferent connections of the nucleus raphe dorsalis. Brain Res. 107:137-144. Tork, I. (1985) Raphe nuclei and serotonin containing system. In G. Paxinos (ed): The Rat Nervous System, Vol. 2. Sydney: Academic Press, pp. 43-78. Villar, M.J., M.L. Vitale, T. Hokfelt, and A.A.J. Verhofstad (1988) Dorsal raphe serotoninergic branching neurons projecting both to the lateral geniculate nucleus and superior colliculus: A combined retrograde tracing-immunohistochemicalstudy in the rat. J. Comp. Neurol. 277t126 140. Waeber, C., K. Dixon, D. Hoyer, and J.M. Palacios (1988) Localization by autoradiography of neuronal 5-HT3receptors in the mouse CNS. Eur. J. Pharmacol. 151r351-352. Wallace, J.A., P. Petrusz, and J.M. Lauder (1982) Serotonin immunocytochemistry in the adult and developing rat brain: Methodological and pharmacologicalconsiderations. Brain Res. Bull. 9:117-129. Westlund, K.N., L.S. Sorkin, D.G. Ferrington, S.M. Carlton, H.H. Willcockson, and W.D. Willis (1990) Serotoninergic and noradrenergic projections to the ventral posterolateral nucleus of the monkey thalamus. J. Comp. Neurol. 295:197-207. Wilson, J.R., and A.E. Hendrickson (1988) Serotoninergic axons in the monkey’s lateral geniculate nucleus. Vis. Neuroscience 1.125-133. Wilson, M.A., G.A. Ricaurte, and M.E. Molliver (1989) Distinct morphologic classes of serotoninergic axons in primates exhibit differential vulnerability to the psychotropic drug 3-4-methylenedioxymethamphetamine. Neuroscience28: 121-137. Woolf, N.J., and L.L. Butcher (1986) Cholinergic systems in the rat brain: 111. Projections from the pontomesencephalic tegmentum to the thalamus, tectum, basal ganglia, and basal forebrain. Brain Res. Bull. 16:603-637. Yoshida, M., M. Sasa, and S. Takaori (1984) Serotonin-mediated inhibition from dorsal raphe nucleus of neurons in dorsal lateral geniculate and thalamic reticular nuclei. Brain Res. 290:95-105.