Serotonin axon terminals in the ventral tegmental area of the rat: fine structure and synaptic input to dopaminergic neurons

Brain Research. 435 ( 1987171-83 Elsevier

71

BRE 13082

Serotonin axon terminals in the ventral tegmental area of the rat: fine structure and synaptic input to dopaminergic neurons Denis Hervd ~, Virginia M. Pickel 2, Tong H. Joh 2 and Alain Beaudet 1 "~Mo~,trealNeurological Institute. McGill Ulfirersio', Montreal Quebec tCanada) and -'Department qfNeuroh~gy. ¢.'ornell Unirersity Medical School. New York. ,VY 10021 ~U. S.,4. ) (Accepted 12 May 1987) Key words: Tyrosine hydroxylase; Midbrain; Electron microscopy; Radioautography; lmmunocytochemistry

The serotoninergic (5-hydroxytryptaminc, 5-HT) inner, at|on of the rat ventral tcgmental area (VTAJ wa,,~examined by light and electron microscopic radioautography I~llowing intravcntricular infusion of [3HI5-HT. The [3HI5-HT labeled processes were characterized with respect to their regional distribution, ultrastructure and relationships with all neurons, including dopaminergic :~eurons, identified in the same sections using immunoeytochcmistry for the localization of the catecholamine-synthesizing enzyme, t.vrosinc hvdroxylase (TH). By light microscopy. [3H]5-HT labeled axons and axoaal varicosities were detected throughout the intcrfascicuh~r nucleus and ventral portion of the VTA. By electron microscopy, [3H]5-llT-labeled axons were lk~und to be mainly small and unmyclinatcd, although a few showed several lamcllae of myelin. The labeled varicosities measured 11.6!tin in mean diameter and contained many small, round or flaltcncd agranular vesicles and a few large granular vesicles. More than 18c~ showed synaptic specializations in single thin sections. Most of these synapses were asymmetric and estabZshed on dendritic shafts. Based on the probability of seeing such synoptic specializations in single thin sections, it was estimated that as m:my as 511% of the labeled 5-HT terminals fornlcd synoptic contacts in the VTA. In dually labeled light microscopic sections, [~Hl5-HT-accumulating processes often appeared adjacent to TH-immunore;tctive pcrikarya and proximal dendrites. Electron microscopy demonstrated that terminuls with radioautographic lubeling fl~r 5-1-1"I"lk)rmed conventional synapses both with TH-labclcd and unlabeled dendrites in the VTA. Many additional 5-H'F terminals lacking recognizable synaptic dens|lies were directly apposcd to TH-labeled dendrites and were isohltcd from the rest of the neuropil by thin glial leaflets. These results suggest that 5-HT neunms innervate both dopaminergic and non-dopaminergic ncnrollS ill lilt VTA and 111113'illflt|ence mesocortical and mcsolimbic efferent systems tlu'talgh synoptic as well as non-synaptic mechanisms.

INTRODUCTION

The ventromedial midbrain tegmentum is comprised of the ventral tegmental area of Tsai (VTA) and a number of associated nuclei, among which are the interfascicular and rostral linear nuclei23"42"53. This region contains the largest collection of dopamine (DA)-containing neurons in the brain (Group A10, according to the nomenclature of Dahlstrom and Fuxe 2°) and projects mainly to limbic structures, namely the nucleus accumbens, the olfactory tubercle, the lateral septum, the amygdala, the pre- and supragenuai fields of the frontal cortex and the entorhinal cortex8'9"24"25"37"5°'53'55. It is also a major route of passage for both descending and ascending path-

ways linking forebrain to hindbrain. Prominent among these is the so-called transtegmental serotonin (5-HT) fiber pathway, which originates from the dorsal and median mesencephalic raphe nuclei and arches ventralward across the tegmentum to join the medial forcbrain bondle 2"n'l'r26'4t. Several lines of evidence suggest that ascending 5-HT axons not only pass through but also terminate within the VTA. These include: (1) the biochemical demonstration of high levels of endogenous 5-H'l~s, tryptophan hydrox~,lase activity~ and high affinity uptake of 5-HT 3 and 12i the presence of a dense network of 5-HT axonal varicosities, identified both by radioautography41 and immunohistochemistry5~ within the VTA. There is still no ultrastructural evidence, however, that 5-

Correspondence: A. Beaudet. Neuroanatomy Laboratory, Montreal Neurological institute, 38111 University Street, Montreal. Ouc. H3A 2B4, Canada. 0006-8993/87/$03.511 © 1987 Elsevier Science Publishers B.V. (Biomedical Division)

72 HT axons directly contact neurons of the VTA. I,~ the present study, the 5-HT innervation of rat ventromedial tegmentum was visualized by light and electron microscopic radioautography following intraventricular infusion of [3H]5-HT, and characterized in terms of its distribution, fine structural features and synaptic relationships. A second phase of the study was directed towards the identification of structural relationships between [3H]5-HT-labeled processes and DA-containing neurons, identified in the same sections by immunocytochemical localization of the biosynthetic enzyme, tyrosine hydroxylase (TH). This double labeling study was prompted by (1) the fact that DA neurons account for almost ;wo-thirds of neurons in the VTA 53 and (2) biochemical27"2s and pharmacological3 evidence to the effect that these neurons might be regulated by 5-HT. MATERIALS AND METHODS Adult male Sprague-Dawley rats (180-200 g; n = 6) were pretreated with a monoamine oxidase inhibitc, r (pargyline, 75 mg/kg i.p.. 1 I1 pre.~operative), anesthetized with chloral hydrate (300 mg/kg i.p.) and recounted in a Kopf stereotaxic instrument. Tritiated st'rotonin ([3tt]5-HT, spee. act. 11-27 Ci/mmol, Am~rsham) was infused into the left lateral ventricle through a gauge-10 cannula coupled to a Sage microperfusion pump. The tracer was evaporated from stock solution under a gentle stream of nitrogen and reconstituted at a concentration of 10-4 M with isotonic saline containing 0.1% ascorbic acid and 10 -3 M non-radioactive norepinephfine (1-noradrenaline-HCl, Sigma). The tracer was delivered at an initial rate of 4.5 ~tl/min during 15 min and at a rate of 0.8 ~tl/min thereafter for a total volume of 200 gtl over a 3-h period. At the end of the infusion, the brains were fixed by intra-aortic arch perfusion of a mixture of 0.5-1% glutaraldehyde and 4% paraformaldehyde in 0.12 M phosphate buffer (pH 7.4). The entire midbrain tegmentum was then blocked for vibratome sectioning and postfixed for 1 h at room temperature in the same fixative. The blocks were serially sectioned at 20/~m thickness on an Oxford vibratome. The sections were collected in phosphate buffer, rinsed twice in 0.5 M Tris-buffered saline (TBS; pH 7.6) and immunocytochemically stained for TH according to a modifica-

tion ~5 of the peroxidase-antiperoxidase (PAP) method of Sternberger 52. In brief, the sections were preincubated in normal goat serum diluted 1:30 with TBS and sequentially incubated at room temperature with: (a) a 1:750 dilution of TH antiserum 33 overnight; (b) a 1:50 dilution of goat anti-rabbit lgG's (Miles Laboratory) for 30 min; and (c) a 1:50 dilution of rabbit (PAP) complex, also for 30 min. All dilutions were made with 0.5 M TBS containing 1% normal goat serum. The peroxidase was reacted for 6 min in a 0.05% 3-3'-diaminobenzidine solution containing 0.01% H202. The sections were briefly rinsed in distilled water and processed for either light or electron microscopic radioautography. For light microscopic radioautography, the sections were mounted onto gelatin-coated slides, dehydrated in graded ethanols, defatted in xylene and coated by dipping in Kodak NTB-2 emulsion diluted 1:1. The radioautographs were exposed for 12-15 days and developed for 11/2 min at 18 °C in Dektoi (Kodak). For electron microscopic radioautography, the sections were postfixed for 1 h in 2% OSO4, dehydrated in ethanols and fiat-embedded in Epon between two plastic coverslips. After polymerization, they were re-embedded in Beem capsules and trimmed for thin sectioning according to light microscopic observations. Ultrathin sections were collected from the surface of each block, deposited onto parlodion-coated slides, stained with uranyl acetate and lead citrate and sprayed with a thin layer of carbon. The slides were then dipped in Ilford emulsion L-4 diluted 1:5 and developed in MicrodoI-X (1.5 min at 18 °C) or paraphenylenediamine (1 min at 20 °C) after 2 and 6 months of exposure, respectively. The sections were collected by floating on 300mesh copper grids and examined with a JEOL 100 CX microscope after thinning the parlodion film in amyl acetate. A first electron microscopic analysis was carried out in 3 rats, on sections that showed good overlap between radioautographic and immunocytochemical labeling at the light microscopic level. The sections were systematically scanned and every [3H]5-HTlabeled profile (defined as such when overlaid by more than 3 silver grains) photographed at an initial magnification of 10-15,000. This first analysis yielded a sampling of 285 labeled elements. A second analy-

Figs. 1-3. Light microscopic immuno-radioautographs from the ventral tegmental-interfascicular area after intraventricular infusion of [3H]5-HT. Tyrosine hydroxylase (TH)-immui.oreactive neurons are filled with an orange-brown reaction product. [3H]5-HTlabeled axons and/or axon terminals are detected in the form of small and dense aggregates of silver grains. Strands of silver grains are also apparent over transtegmental axon bundles (arrows). Fig. 1. Overview of the medioventral tegmentum illustrating the overlap between radioautographic and immunocytochemical labciings. [3H]5-HT-labeled varicosities form a particularly dense network within the interfascicular nucleus (IF). Note that the intensity of the radioautographic labeling decreases from the surface of the brain inward. VTA, ventral tegmental a~ea; III, oculomotor nerve. Scale bar: 50#m. Fig. 2. Higher magnification of the transitional zone between the interfascicular nucleus (IF) and the adjoining ventral tegmental area (VTA). Radiolabeled 5-HT varicosities are interspersed among, and often lie adjacent to TH-immunoreactive perikarya and/or dendrites. Scale bar: 30#m. Fig. 3. In the ventral tegmental area proper, radioautographieally labeled beaded 5-HT axons are seen coursing alongside TH-immunoreactive profiles. Scale bar: 30/~m.

74 sis was carried out in 8 blocks from the same animals, documenting only those [3H]5-HT-iabeled profiles that were apposed to TH-immunoreactive elements. A total of 64 radiolabeled structures were thus recorded, counting labeled profiles identified in more than one thin section only once. Data from all 3 animals were pooled for each analysis. RESULTS

[~H]5-HT-labeled processes in the VTA Light microscopy. Two distinct [3H]5-HT labeling patterns were distinguished by light microscopic radioautography: (1) axonal varicosity labeling, in the form of small and dense aggregates of silver grains and (2) fiber labeling, detected as intervaricose strands of silver grains (Figs. 1-3). Labeled varicosities were apparent throughout the ventromedial midbrain tegmentum, but predominated in the interfascicular nucleus (Figs. 1,2) and ventral third of the VTA (Fig. 3). Labeled fibers were most prominent dorsolateral to the interfascicular nucleus, where they appeared as thick, non-varicose, obliquely cut bundles (Fig, 1). Labeled irrocesses of both types were consistently more numerous and more intensely reactive in sections fixed with 1% glutaraldehyde than in sections fixed with 0.5% glutaraldehyde. Electron microscopy. Of the 285 [3H]5-HT-labeled profiles detected in our systematically scanned material, 58 (20%) were identified as cross-sectioned axons and 227 (80%) as axonal varicosities (Figs. 4-9). Most of the labeled axons (45/58) were small (diameter = 0.1-0.3~m; mean = 0.2~m) and unmyelinated. They contained a moderately electrondense axoplasm, numerous microtubules and occasional vesicles. Other labeled axons (13/58) were slightly larger (mean diameter 0.5 #m) and showed several lamellae of myelin (Figs. 4,9). These myelinated axons usually were seen in tightly arranged bu,~dles including many unmyelinated axons that were isolated from the rest of the neuropii by glial ieatlct~ (Fig. 4). Labeled varicosities were usually ovoid in shape and measured between 0.4 and 1.0 #m in diameter (mean = 0.6t~m). A small proportion was somewhat larger, with diameters ranging between 0.8 and 1.6 pm. Labeled varicosities exhibited a moderately dense axoplasm and contained variable numbers of

Fig. 4, Myelinated5-HT axon. This radiolabeledaxon runs in a bundle with other myelinatcd as well as unmyelinatedaxons. Note the glial enshc~ltllnlcnt that insulates the axons from the stn'r~tmdingneuropil (arrowh¢i~ds), Scale bar: tl.41tm. Fig. 5. A [3H]5-HT-labeled varicosity, filled with small clear synaptic vesicles of approximately the same size and shape of those seen in nearby unlabeled terminals, is directlyapposed to an immunonegativedendrite. No synapticdifferentiationis visible at the site of apposition in this plane of section, but a thin glial leaflet almost entirely isolates the apposed elements from the rest of the neuropil (clear arrowheads). Note that the recipient dendrite is synapticallycontacted by an unlabeled axon terminal (filled arrowhead). It also establishes a synaptie contact on a small dendritic spine through a slender, vesicle-containing appendage (arrow). Scalebar: 0.4pro.

synaptic vesicles, often completely filling the labeled profile (Figs. 5,6). Most frequently, these synaptic vesicles were small (15-25 nm in diameter), agranular, and round or flattened in shape. They were also often associated with tiny tubular organelles of approximately the same diameter. Some of the varicosities contained larger (30-50 nm in diameter), round electron-lucent vesicles of a more common type (Figs. 6,9). Finally, most labeled varicosities exhi-

75 bited a small n u m b e r (2 p e r varicosit~ on a v e r a g e ) of r o u n d or e l o n g a t e d large d e n s e core vesicles (Figs. 6,8,9). O f t e n , labeled varicosities were s e e n to

e m e r g e f r o m thin u n m y e l i n a t e d axons (Fig. 7). In such cases, intra-axonal m i c r o t u b u l e s were interrupted a n d r e p l a c e d by synaptic vesicles (Fig. 7).

Fig. 6. A 5-HT varicosity, filled with clear, round synaptic vesicles and exhibiting one large granular vesicle rests on a stout nmmunonegative dendritic shaft. No membrane specialization is appar¢~t at the sLte of contact in this plane of section, Note that the labeled varicosity is ensheathed over the rest of its surface by a thin astroglial leaflet (arrowheads). Scale bar: 0.2~m. Fig. 7. A thin, unmyelinated [3H]5-HTdabeled axon courses between two immunonegative dendritic shafts. The labeling overlies a small varicose dilation, exhibiting clustered vesicular organelles, A density, which may be presynaptic, is apparent within the 5-HT varicosity, along part of its contact with one of the two dendrites (arrow). No clear postsynaptic differentiation is visible in this plane of section• Scale bar: 0.2 pro.

76 Most [3H]5-HT-labeled varicosities (=80%) were directly apposed to medium-size dendritic profiles without intervening glia (Figs. 7-9). However, only 42 out of the 227 varicosities examined exhibited a well-developed junctional complex in the plane of section (Figs. 7-9), Irrespective of whether or not they showed a synapse, these varicosities were often encapsulated over the rest of their surface by a thin glial leaflet (Figs. 5,6). When present, synaptic complexes were usually of the asymmetrical variety (Fig. 8). Symmetrical synapses were only rarely encountered (Fig. 9). Active zones measured on average 0.3 l*m in length. Although often juxtaposed to more than one dendrite, labeled varicosities were never found in synaptic contact with more than one den-

drite in a given plane of section (Fig. 7). Several [3H]5-HT-labeled axonal varicosities were adjacent to large proximal dendrites, but none was seen in synaptic junction with, or even directly apposed to, a neuronal perikaryon.

Relationships between [3H]5-HT-labeled varicosities and TH-irnrnunoreactive elements Light microscopy. In the light microscope, TH-immunoreactive nerve cell bodies and dendrites were apparent throughout the V T A as well as within interfascicular and rostral linear nuclei and substantia nigra (Figs. 1-3). The distribution and cytological features of these immunoreactive elements were similar

Fig. 8. An intensely labeled 5-HT varicosity, filled with clear pleomorphic synaptic vesicles, establishes an asymmetrical synapse on a dendritic shaft, Subsynaptic dense bodies-~sare clearly visible beneath the active zone (arrowhead). Note that the recipient dendrite is also contacted by two unlabeled axon terminals. Scale bar: 0.21~m. Fig. 9. A [3H]5-HT-labeled axonal varicosity, containing numerous clear pleomorphic vesicles as well as several large granular vesicles, forms a symmetrical synaptic junction with a dendritic branch. The same dendrite is also contacted by an unlabeled terminal similarly endowed with small clear and large granular vesicles. A labeled myelinated 5-HT axon is visible in the same field (arrow). Scale bar: 0.2p~m.

77 to that of DA neurons previously described within the same region ~°'3°. Both the number of labeled cells and the intensity of immunohistochemical reaction were higher in sections fixed with 0.5% glutaraldehyde than in sections fixed with 1% glutaraldehyde. In addition, marked differences in immunostaining were noted between individual perikarya under each of these two fixation conditions. Cells in the interfascicular nucleus, in particular, consiste,itly exhibited more intense immunoreactivity than those in the dorsolateral VTA (Fig. 1). In the ventral part of the VTA, as well as throughout the interfascicular nucleus, TH-immunoreactive perikarya and dendrites were interspersed with [3H]5-HT-labeled axonal varicosities (Figs. 1-3). In many instances, the labeled varicosities were adjacent to TH-immunoreactive dendrites or perikarya (Figs. 2,3). Accumulations of radioactivity were never found superimposed over TH-immunoreactive perikarya or dendrites,. Electron microscopy. By electron microscopy, 7% of systematically sampled [3H]5-HT-labeled axonal varicosities (16/227) were seen in direct apposition to dendrites showing peroxidase labeling for TH (Figs. 10-15). All 64 appositions subsequently visualized between [3H]5-HT-labeled terminals and TH-immunoreactive elements also involved immunoreactive dendrites. Most of the sections exhibiting dual labeling were from the outermost surface of tissue fixed with 0.5% glutaraidehyde, where the penetration of immunoreagents was optimal. [3H]5-HT-labeled terminals apposed to TH-immunoreactive dendrites were morphologically similar to those abutting unlabeled profiles. The TH-labeled dendrites were usually filled with dense peroxidase reaction product which was concentrated on the inner surface of the cytoplasmic membrane and around microtubules and mitochondria (Figs. 10-15). Only 9 out of the 80 [3H]5-HT-labeled axonal varicosities found adjacent to TH-immunorcactive dendrites (total sampling) showed a well-defined junctional complex at the site of contact (Figs. 14,15). In the rest of the cases, synaptic specializations were either absent or obscured by subplasmalenmal accumulations of immunoreaction product (Figs. 11-13). Most synaptic junctions were asymmetrical and established on medium size dendritic shafts (Figs. 14,15).

DISCUSSION

Methodology The present study describes the fine structural teatures of [~H]5-HT-labeled axon terminals and their relationship to TH-immunoreactive as well as to nonimmunoreactive neurons within the ventromediai midbrain tegmentum of the rat. The specificity of the axonal labeling detected by radioautography after in vivo administration of [3H]5-HT under experimental conditions comparable to the present ones has been firmly established (for a review, see ref. 6). That this labeling selectively involved 5-HT-containing elements was further supported here by the absence of tracer accumulation within TH-immunoreactive neurons, previously shown to be the most likely to take up [3H]5-HT non-specifically21-36-~9. The immunolabeling detected using the present anti-TFl antibody has been shown to be selective for catecholaminergic neurons (see ref. 43 for control studies). Catecholamine nerve cell bodies within the VTA are all believed to be dopaminergic, based on early pharmacohistofluorescence data (for a review, see ref. 10) and on the absence of detectable immunoreactivity for the noradrenaline synthesizing enzyme dopamine-fl-hydroxylase (ref. 54 and unpublished observations). Fixation conuition,~ were found to markedly influence both the radioartographic detection of [~H]5HT and the immunocytochemical localization of TH. High concentrations of glutaraldehyde favored the detection of [3H]5-HT-labelcd axonal varicosities, particularly in regions far from the ventricles and the base of the brah~ (which were less permeated by the tracer), thereby confirming the importance of glutaraldehyde for cross-linking ['~H]5-HT molecules within their uptake and/or ~torage sites -'t'3-'. However, high concentrations of glutaraldehyde also led to decreased immunocytochemical detection of TH. as with most antigens, presumably through both loss of ~,ntigenicity and reduced penetration of antisera across well fixed membranes 4~. It is therefore likely that [~H]5-HT-labeled and TH-immunoreactive profiles detected under either one of the fixation conditions used in the present study represent only a fraction of elements conlaining each marker, and that the true incidence of contacts between them is actually higher than that estimated.

78

Figs. 10-15. ?!ectron microscopic radioautographs of [~H]5-HT-labeled varicositiesadjacent to TH-immunoreactive elements. Fig, I0. A largeTH-immunoreactive dendrite iscontacted by I labeled and 3 unlabeled axon terminals.The labeled varicosityappears somewhat lesselectron dense than itsunlabeled congeners, probably owing to itslower content in synaptic vesicles.It is also the only one that shows no synaptic specializationin thisplane of section.Scale bar: 0.4 ~m. Fig. 11. An intenselylaoeled [3H]5,HT.labeled varicosity,filledwith clear synaptic vesicles,liesimmediately adjacent to a large THimmunoreactive denuritic shaft. Scale bar: 0.4 :tm. Fig. 12. This labeled 5-HT varicosity is also apposed to a TH-immunoreactive dendritic shaft, but again shows no junctional specialization in the pl~qe of section. In fact, within the varicosity, synaptic vesicles appear to be streaming in the opposite direction, towards a small immun~t~egative dendritic spine (arrow), Scale bar: 0.4#m, Fig. • Other example of a [3HI5-HT-labcled axonal varicosity in :,~,ntact with a TH-immunoreactive c~endrite. The labeling appears to arise from a ciustcr of large granular vesicles within the varicosity. The rest of the bouton is filled with small clear svnaptic vesicles. No synapfic :en,.ily is apparent in tbis pl~,ne of section. The immu~oteactive dendrite is synaptically contacted by an unlabeled axon. Inset: a ,aali [3Hl5-HT-labeled varicosity, filled with plcomorphic ve~icular organelles, leans alongside a TH-irnmunoreactive dendrite. "~dense osmium deposit is s isiblc on the dendritic shaft, beneath the junctional zone (arrow). This densi~.y appears to correspond to a submembrane PAP deposit rather than to a post-synaptic specialization. Scale bar: 0,4t, m.

-mr/17"0 :J~q al~,zS "ol!apuap ~At.lzeoaountum! oql u!ql.tza solnqnloa~!m pzuo!~,~as -ssoao ql!m s~,~.sodapd V d jo uo.zle.t3osst~oql oloN uojnoq IH-S oql op!su! OlO!SZA~elnuea~ oSae I e tuoaj old!pea ol sae~dde I~qel 2.tqdea$ -omeo!pea oq£ 'ljeqs 3!l!Jptlop OA!~eoaounmtu!. H& e uo asd~.u~s i~!.ij~mtu~s e ~Uplem 2{1.tso3~.te^,LH'S p~loq~l e jo zldtuvxT/"gl "$!:1 "troll,'0 :J~.q ~1~3S "uolnoq pal~qe, lun uP, £q pal~,luo30Sl~, s! 31!Jpuop lua -!d!3~l ~ q l "l~3!.ll~t'ltll,(s~,s! uo!l~'!lu~.l~jj[p 3!ldeu,{s ~q,L "q~uP.aq 311]apuap ~A!13P.~JountutU!-H&P, ql!~ l~~P,l,llO3 3!ldP,u/~s u! u~os sl "s~uo aP,inunafi o~ae I jo uo!!3OllO3 I~ sP, ilZat se s0p!s~a 3!ldl~uKs a~Ol3 IletUs snoaotunu ~u.tu!eluo3 '~{I!SO3]JUApoloqel-lH-~[Hf] s!q,L 'l'1 'fi!rl

..~

6L

. ¢ .~ ,i ~ . •

• ..

~, . ~..~.

~'.~r,,~ ~ .~ , .~t. • ~

~'~

,%.~ ~ . ~

~*~'~ . ~

J

8(1

Serotonin innervation of the ventral tegmental area Radiolabeled elements detected in the VTA following intraventricular infusion of [3H]5-HT included both axons and axon terminals. By light microscopy, labeled axons were most apparent dorsolateral to the interfascicu!ar nucleus, in the position of the ascending transtegmental 5-HT pathways-~.19'26.~. [3H]5-HT-labeled axonal varicosities were mainly detected within the interfascicular nucleus and the ventral portion of the VTA. This topographic d,stribution conformed with that observe~ by immunohistochemistry using anti-5-HT antibodiesst and is therefore not merely due to greater tracer accessibility. By electron microscopy, [3H]5-HT-labeied axons appeared primarily unmyelinated, although some showed several sheaths of myelin, in keeping with the previous description of a contingent of myelinated 5HT axons xithin the medial forebrain bundle of rat and monkey I. Myelinated 5-HT axons have also been described in the lower brainstem, both in the paratrigeminal nucleus ~s and substantia gelatinosa .af the spinal trigeminal nucleus 47. The occurrence of these different populations of axons (myelinated and unmyelinated) suggests that ascending as well as descending 5-HT axons may have variable conduction velocities, as noted previously for catccholaminergic neurons t5,3'~, Most [3H]5-HT-labeled axonai varicosities exhibited a den~; axoplasmic matrix filled with minute, pleomorphic synaptic vesicles. Comparable vesicular contents were reported in 5-HT varicosities from the brainstem ~Tas, the locus coeruleus35, the cerebellum 7A7. the inferior olive 56 and the suprachiasmatic nucleus of the hypothalamus l'~. While more than 80% of the 5-HT-labeled varicosities abutted dendritic profiles, only 18% showed a synaptic specialization (density) in random thin sections. From this figure, and based on the assumption that 5-HT varicosities correspond to spheres 0.6 t~m in diameter and active zones to caps 0.31tm in diameter, at least 50% of 5-HT varicosities may be estimated to form synaptic contacts in the VTA and interfascicular nucleus7, This incidence is markedly higher thak~ that reported for 5-HT axon terminals in other areas of thebrain (for reviews, see refs. 4 and 6) and might imply that the midbrain tegmentum is innervated by a particular subset of 'junctional' 5-HT

neurons. Such a subset, if it indeed exists, would not correspond to any morphologically defined sub-class of 5-HT axons since 5-HT varicosities comparable in size, shape and internal constituents to the ones detected here have been shown to establish frequent synapses in certain brain areas (e.g. in the sup~'achiasmatic nucleus 12) and not in others (e.g. in cerebellar cortex7'17). An alternate interpretation is that epigenic factors, such as !he territory of innervation and/or chemical nature of target neurons, account for the higher junctional frequency of 5-HT axons. This interpretation would be in keeping with the fact that the synaptic incidence of central 5-HT boutons may be modified by inducing changes in their target zone 7. Whatever the case may be, the occurrence of an important contingent of synaptic 5-HT anon terminals within the VTA confirms that the varicose 5HT axons previously identified in this area with the light microscope provide a true local innervation and do not merely represent transtegmental axons of passage.

Serotonin innervation of dopaminergic neurons The present study provides the first evidence for direct synaptic connections between 5-HT axon ter~ minals and DA neurons in the ventral midbrain tegmentum. The existence of such contacts suggests that ascending ~-HT axons may directly influence the activity of DA cells. However, it is unclear whether differentiated synaptic contacts provide the sole morphological substrate for interactions between 5-HT and DA-eontaining elements within the VTA, or if a transfer of information is also possible between nonjunctional!y apposed elements. Less than 12% of [3H]5-HT-labeled boutons here found abutting THimmunoreactive dendrites showed a junctional complex at the site of contact. This low incidence suggests that only a flaction of 5-HT varicosities apposed to TH-immunoreactive processes is in synaptic contact with them. If they all were, the occurrence of synaptic specializations between 5-HT and TH-containing elements would have been higher, especially since planes of section in which the two are seen in contact are also plane~ in which synaptic specializations between them are most likely to be found. Although 5HT boutons abutting TH-immunoreactive dendrites could obviously synapse on neighboring non-DA neurons, others might be deprived of junctional, spe-

81 cialization and release 5 - H T extrasynaptically as proposed for 5-HT varicosities in other brain areas 22 (for reviews, see refs. 4 - 6 ) . If this is indeed the case, the thin glial leaflets which were often found to isolate apposed, but non-synoptically linked structures from the rest of the neuropil might provide a shield against diffusion of 5-HT in the extracellular space and therefore make the adjacent D A dendrite a prime target for the released biogenic amine. Junctional as well as non-junctional relationships between 5-HT and catecholamine-containing elements have previously been reported in various regions of the rat CNS. In the medial zona incerta and ~.rcuate nucleus of the hypothalamus, 5-HT axo~l terminals were shown to contact, and synapse upon, D A dendrites v~'34. In the median eminence, 5 - H T containing terminals were found abutting D A nerve endings in the vicinity of portal perivascular spaces 13. Finally, in the locus coeruleus 35"46and within the medial and commissural portion of the nucleus tractus solitarius 44, 5-HT axons were shown to contact both the perikarya and dendrites o~ noradrenaline-containing neurons. The present study provides another example of what appears to be a multi-level network of proximal and distal links between serotonin and catecholamine systems in the mammalian neuraxis. Pharmacological studies have shown that 5-HT may act both proximally 3 and distally 16 upon mesotelencephalic D A neurons. In keeping with these observations is the fact that electrolytic lesions of nucleus raphe dorsalis induce an increase in D A turnover in nucleus accumbe,as with no apparent effect on

REFERENCES 1 Azmitia, E.C. and Gannon, P., The ultrastructural localization of serotonin immunoreactivity in myelinated and unmyelinated axons within the medial forebrain bundle of rat and monkey, J. Neurosci., 3 (1983) 2083- 2090. 2 Azmitia, E.C. and Segal, M., Autoradiographic analysis of the differential ascending projections of the dorsal and median raphe nuclei in the rat, J. Comp. Neurol., 179 (1978) 641-668. 3 Beart, P.M. and McDonald, D., 5-Hydroxytryptamine and 5-hydroxytryptaminergic-dopaminergic interactions in the ventral tegmental area of rat brain, J. Pharm. Pharmacot.. 34 (1982) 591-593. 4 Beaudet, A. and Desearries, L., The fine structure of central serotonin neurons, J. Physiol. (Paris), 77 (1981.) 193-203. 5 Beaudet, A. and De, arries, L., Fine structure of monoamine axon termina:s in cerebral cortex. In L. Descarries,

D A utilization in the prefrontal cortex, whereas lesions of nucleus raphe medianus, which also result in an increase in D A turnover in nucleus accumbens, reduce dihydroxyphenylacetic acid/DA ratios in prefrontal cortical areas 27.28. The cellular relationships here demonstrated between 5-HT axons and D A dendrites provide a morphological substrate for direct proximal interactions between 5-HT and D A neurons within the VTA. The present results also suggest, however, that :hese interactions are not exclusive. Indeed, an important proportion of [3H]5HT-labeled axon terminals in the V T A was either directly apposed to, or formed synaptic junctions with dendrites that were devoid of TH-immunoreacqvity. While this finding may partially be accounted for by methodological factors (see above), it seems likely that n o n - D A neurons, which account for approximately one third of the cells in the VTA -~3, also receive a 5-HT innervation, W h e t h e r some of these n o n - D A recipient cells correspond to the G A D -4°, neurotensin -2~ or cholecystokinin-immunoreactive neurons 3~ previously detected by immunohistochemistry within the VTA, remains to be established. ACKNOWLEDGEMENTS We thank Miss K. Leonard for expert technical assistance, Mr. C. Hodge for photographic work and Miss B. Lindsay to typing the manuscript. This work was supported by MRC Grant MT-7366, N I M H Grant MH40342 and an Edith and Richard Strauss Foundation Fellowship to Dr. D. Herv6,

T.A. Reader and H.H. Jasper (Eds.), Monoamine Innervation of Cerebral Cortex, Alan R. Liss, New York, 1984, pp. 77-93. 6 Beaudet, A. and Descarrics. L.. UItrastructural identification of serotonin neurons. In H.W.M. Steinbusch (Ed.), Monoaminergic Neurons at the Light Microscopical and Ultraslructural Let','l. IBRO Hamtbouk Series: Methods in the Nenrosciences. Wiley, Chiccster, 1987, pp. 265-313.

7 Beaudet, A. and Se',elo, C., Synaptic remodeling of serotonin axon terminals in rat agranular cerebellum, Brain Research, 206 (1981) 305-309. 8 Beckstead, R.M.. Domesiek, V.B. and Nauta, W.J.H., Efferent connections of the substantia nigra and ventral tegmental area in the rat, Brain Research, 175 (1979) ~91-217. 9 Berger, B., Thierry, A.M., Tassin, J.P. and Moyne, M.A., Dopaminergic innervation of the rat prefrontal cortex: a fluorescence histochemical study, Brain Research. 106 (1976) 133-145. 10 Bjfrklund, A. and Lindvall, O., Dopamine-containing sys-

82 terns in the CNS. In A. BjOrklund and T. HOkfelt (Eds.), Handbook of Chemical Neuroanatomy, Vol. 2: Classical Transmitters in the CNS, Part L Elsevier, Amsterdam, 1984, pp. 55- '.22. 11 Bobillier, P., $eguin, S., Degueurce, A., Lewis, B.D. and Pujol. J.F., The efferent connections of the nucleus raphe ceutralis superior in the rat as revealed by radioautography, Brain Research, 166 (1979) 1-8. 12 Bosler, O. and Beaudet, A., VIP neurons as prime synaptic targets for serotonin afferents in rat suprachiasmatic nucleus: a combined radioautographic and immunocytochemical study, J. Neurocytol., 14 (1985) 749-763. 13 Bosler, O. and Beaudet, A., Relations uitrastructurales entre syst~mes monoaminergiques et peptidergiques dans I'hypothalamus. Approche radioautographique et immunocytochimique coapl6e dans le noyau arqu6 et le noyau suprachiasmatique du rat, Ann. Endocrinol., 46 (1985) 19-26. 14 Bosler, O., Joh, T.H. and Beaudet, A., Ultrastructural relationships between serotonin and dopamine neurons in the rat areuate nucleus and medial zona incerta: a combined radioautographic and immunocytoehemical study, Neurosci. Le:i., 48 (1984) 279-285. 15 Brown, D.L. and Guyenet, P.G., Cardiovascular neurons of the nucleus paragigantocellularis lateralis with pr~ iections to the spinal cord, Am. J. Physiol., 247 (1984) RI009-R1016. 16 Carter, C.J. and Pycock, C.J., A study of the sites of interaction between dopamine and 5-hydroxytryptamine for the production of fluphenazine-induced catalepsy, NaunynSchmiedeberg'sArch. Pharmacol., 304 (1978) 135-139. 17 Chan-Palay, V., Fine structure of labelled axons in the cerebellar cortex and nuclei of rodents and primates after intraventricular infusio,~ of tritiated serotonin, Anat. Em. bryol., 148 (1975) 235-265. 18 Chan-Palay, V., The paratrigeminal nucleus. II. Identification and inter-relations of catecholamine axons, indolea. mine axons and substance P immunoreactive cells in the neuropil, J. Neurocytol., 7 (1978)419-442. 19 Conrad, L.C.A., Leonard, C.M. and Pfaff, D.W., Connections of the median and dorsal raph~ nuclei in the rat: an autoradiographic and degeneration study, J. Comp Neurol., 156 (1974) 179-206. 20 Dahlstrom, A. and Fuxe, K., Evidence for the existence of monoamine containing neurons in the central nervous system. I. Demonstration of monoamines in the cell bodies of brain stem neurons, Acta Physiol, Scand., 232 (1964) 1-55. 21 Descarries, L. and Beaudet( A., The use of radioautography for investigating transmitter-specific neurons. In A. BjOrklund and T. HOkfelt (Eds.), Handbook of Chemical Neuroanatomy, Vol. 1: Methods in Chemical Neuroanatomy, Elsevier, Amsterdam, 1983, pp. 286-364. 22 Descarries, L., Beaudet, A. and Watkins, K.C., Serotonin nerve terminals in adult rat neoeortex, Brain Research, 100 (1975) 563-588. 23 Domesick, V.B., Stinus, L. and Paskevich, T.A., The cytology of dopaminergic and nondopaminergic neurons in the substantia nigra and ventral tegmental area of the rat: a ltgfit- ,tnd electron-microscopic study, Neuroscie~.ce, 8 (1983) 743-765. 24 Fallon, J.H. and Moore, R.Y., Catecholamine innervation of the basal forebrain. IV. Topography of the dopamine projection to the basal forebrain and neostriatum, J. Comp. Neurol., 180 (1978) 545-580.

25 Fuxe. K., HOkfelt, T., Johansson, O., Jonsson, G., Lidbrink, P. and Ljungdahl, A., The origin of the dopamine nerve terminals in limbic and frontal cortex. Evidence of mesocortical dopamine neurons, Brain Research, 82 (1974) 349-355. 26 Halaris, A.E., Jones, B.E. and Moore, R.Y., Axonal transport in serotonin neurons of the midbrain raphe, Brain Research, 107 (1976) 555-574. 27 Herr6, D., Simon, H., Blanc, G., LeMoal, M., Glowinski, J. and Tassin, J.P., Opposite changes in dopamine utilization in the nucleus aceumbens and the frontal cortex after electrolytic lesion of the median raph6 in the rat, Brain Research, 216 (19~1) 422-428. 28 Herr6, D., Simon, H., Blanc, G., Lisoprawski, A., LeMoal, M., Glowinski, J. and Tassin, J.P., Increased utilization of dopamine in the nucleus accumbens but not in the cerebral cortex after dorsal raphe lesion in the rat, Neurosci. Lett., 15 (1979) 127-133. 29 H6kfelt, T., Everitt, B.J., Theodorsson-Norkeim, E. and Goldstein, M., Occurrence of neurotensin-like immunoreactivity in subpopulations of hypothalamic, mesencephalic and medullary catecholamine neurons, J. Comp. Neurol., 222 (1984) 543-559. 30 H6kfelt, T., Johansson, O., Fuxe, K., Goldstein, M. and Park, D., Immunocytochemical studies on the localization and distribution of monoamine neurons systems in the rat brain. I. Tyrosine hydroxylase in the mesencephalon and diencephalon, Med. Biol., 54 (1976) 427-453. 31 H6kfelt, T., Skirboll, L., Rehfeld, J.F., Goldstein, M., Markey, K. and Dann, O., A subpopulation of mesencephalic neurons projecting to limbic areas contains a cholecystokinin-like peptide: evidence from immunohistoehemistry combined with retrograde tracing, Neuroscience, 5 (1980) 2093-2124. 32 Hopwood, D., Theoretical and practical aspects of glutaraldehyde fixation, J. Histochem., 4 (1972) 267-303. 33 Joh, T.H. and Ross, M.E., Preparatir, n of catecholatninesynthesizing enzymes as immunoge,les for immunohistochemist, y. In A.C. Cuello (Ed.), Immunohistochemistry, IBI~O Handbook 5cries; Methods in the tqeurosciences. Wiley, Chichester, 1983, p p 121-138. 34 Kiss, J. and Halasz, B., Synaptic connections between serotoninergic axon terminals and tyrosine hydroxylase-immunorcactive neurons in the arcuate nucleus of the rat hypothalamus. A combination "~ electron microscopic autoradiography and immunocyt ~ceemistry, Brain Research, 364 (1986) 284-294. 35 L6ger, L. and Descarries, L., Serotonin nerve terminals in the locus coeruleus of adult rat: a .,~dioautographic study, Brain Research, 145 (1978) 1-13. 36 Liehtensteiger, W., Mutzner, U. and Langemann, H., Uptake of 5-hydroxytryptamine and 5-hydroxytryptophan by neurons of the central nervous system normally containing catecholamines, J. Neuro~hem., 14 (1967) 489-497. 37 Lindvall, O., Bj6rklund, A. and Divac, I.. Organization of catecholamine neurons projecting to the frontal cortex in the rat, Brahl Research, 142 (1978) 1-24. 38 Milhand, M. and Pappas, G.D., Post-synaptic bodies in the ha~enula and intclpeduncular nuclei of the rat, J. Cell Biol., 30 (1966) 437-441. 39 Milner. T.A.. Pickel. V.M.. Park, D . H . Joh. T.H. and Reis, D.J., Phenylethanolamine N-methyltransferase-containing neurons in the r , stral ventrolateral medulla of the rat: I. Normal ultrastructurc, Brai, Research. in press.

83 40 Mugnaini, E. and Oertel, W.M., An atlas of the distribution of GABAergic neurons and terminals in the rat CNS revealed by GAD immunohistochemistry. G A B A and neuropeptides in the CNS, Part I. In A. Bj6rklund and T. H6kfelt (Eds.), Handbook of Chemical Neuroanatomy. Vol. 4, Elsevier, Amsterdam, 1985, pp. 436-608. 41 Parent, A., Descarries, L. and Beaudet, A., Organization of ascending serotonin systems in the adult rat brain. A radioautographie study after intraventricular administration of [3H]5-hydroxytryptamine, Neuroscience, 6 (1981) 115-138. 42 Phillipson, O.T., The cytoarchitecture ofthe interfascicular nucleus and ventral tegmental area of Tsai in the rat, J. Comp. Neurol., 187 (1979) 85-98. 43 Pickel, V.M., Immunocytochemical methods. In L. Heimer and M.J. Robarts (Eds.), Neuroanatomical Tract-Tracing Methods, Plenum Press, New York, 1981, pp. 483-509. 44 Pickel, V.M., Joh, T.H., Chan, J. and Beaudet, A., Serotoninergic V'rminals: ultrastructural and synaptic interaction with catecholamine-containing neurons in the medial nuclei of the solitary tracts, J. Comp. Neurol.. 225 (1984) 291-301. 45 Pickel, V.M., Joh, T.H. and Reis, D.J,, Monoamine-synthesizing enzymes in central dopaminergic, noradrenergic and serotoninergic neurons. Immunoeytochemical localization by light and electron microscopy, I. Histochem. Cytochem., 24 (1976) 792-806. 46 Pickel, V.M., Joh, T.H. and Reis, D.J., A serotoninergie innervation of noradrenergic neurons in nucleus locus coeruleus: demonstration by immunocytochemical localization of the transmitter specific enzymes tyrosine and tryptophane hydroxylase, Brain Research, 13 (1977) 197-214. 47 Ruda, M.A. and Gobel, S.. Ultrastruetural characterization of axonal endings in the substantia gelatinosa which take up [3H]serotonin, Brain Research, 184 (1980) 57-83. 48 Saavcdra, J.M., Distribution of serotonin and synthesizing

49

50

51

52

53

54

55

56

enzymes in discrete areas of the brain, Fed. Proc.. 36 (1977) 2134-2141. Shaskan, E.G. and Snyder, S.H., Kinetics of serotonin accumulation into slices from rat brain: relationship to catecholamine uptake, J, Pharmaeol. Exp. Ther., 175 (1970) 404-418. Simon, H., Le Moal, M. and Calas, A., Efferents and afferents of the ventral tegmental A~0 region studied after local injection of [3H]leucine and horseradish peroxidase, Brain Research, 178 (1979) 17-40. Steinbusch, H.W.M., Serotonin-immunoreactive neurons and their projections in the CNS. Classical transmitters and transmitter receptors in the CNS, Part iI. In A. Bj6rklund, T, H6kfelt and M.J. Kuhar (Eds.), Handbook of Chemical Neuroanatomy, Vol. 3, Elsevier, Amsterdam, 1984. pp. 68-125. Sternberger, L.A., Immunocytochemistry. In S. Cohen and R.T. McClusky (Eds.), Boric and Clinical Immunology, 2nd edn., Wiley, New York. 1979, Swanson, L.W., The projections of the ventral tegmental area and adjacent regions: a combined fluorescent retrograde tracer and immunofluorescence study in the rat, Brain Res. Bull., 9 (1982) 321-353. Swanson, L.W. and Hartman B.K., The central adrenergic system: an immunofluorescence study of the location of cell bodies and their efferent connections in the rat utilizing dopamine fl-hydroxylase as a marker, J. Comp, Neurol., 163 (1975) 467-506. Ungerstedt, U., Stereotaxic mapping of the monoamine pathways in the rat brain, Acta Physiol. Scand., 367 (1971) 1-48. Wiklund. L.. Descarries, L. and Mollgard, K., Serotoninergic axon terminals in the rat dorsal accessory olive: normal ultrastructure and light microscopic demonstration of regeneration after 5.6-dihydroxytryptaminc lesioning, J. NeuroQ'lol.. 11)( 1981 } i11119-11127.