0306-4522/92 $5.00 + 0.00 Pergamon Press Ltd (c" 1992 IBRO

Neuroscience Vol. 50, No. 2, pp. 269 282, 1992 Printed in Great Britain

LIGHT A N D ELECTRON MICROSCOPIC LOCALIZATION OF RETROGRADELY TRANSPORTED N E U R O T E N S I N IN RAT NIGROSTRIATAL DOPAMINERGIC N E U R O N S M.-N. CASTEL,*J. WOULFE,tX. WANG,JrP. M. LADURON* and A. BEAUDET't'3~ *Rh6ne-Poulenc Rorer R-D, 13, Quai Jules Guesde-BP 14, 94403 Vitry-sur-Seine, Cedex, France tNeuroanatomy Laboratory, Montreal Neurological Institute, McGill University, 3801 University Street. Montreal, Quebec, Canada H3A 2B4 Abstract--We previously demonstrated the existence of a retrograde axonal transport of radioactivity to the substantia nigra pars compacta following injection of mono-iodinated neurotensin in rat neostriatum. In the present study, the topographical and cellular distribution of this retrogradely transported material was examined by light and electron microscopic autoradiography. Four and a half hours after unilateral injection of [125I]neurotensin in the caudoputamen~ retrogradely labelled neuronal cell bodies were detected by light microscopic autoradiography throughout the ipsilateral substantia nigra pars compacta as well as within the ventral tegmental area and retrorubral field. In semithin sections, silver grains were prevalent over the perinuclear cytoplasm of neuronal cell bodies but were also detected over neuronal nuclei. Analysis of electron microscopic autoradiographs revealed that the vast majority (> 85%) were associated with neuronal perikarya, unmyelinated and myelinated axons, dendrites and terminals. Within the soma, a significant proportion of silver grains (16% of somatic grains) was detected over the nucleus. However, the majority were identified over the cytoplasm where they often encompassed cytoplasmic organdies, including rough endoplasmic reticulum, mitochondria, Golgi apparatus, lysosomes, and multi-vesicular bodies. In dendrites and axons, a substantial percentage of silver grains (63 89%) was localized over the plasma membrane. A minor proportion (13 % of total) of the autoradiographic labelling was detected over myelin sheaths, astrocytes, and oligodendrocytes. The present results are consistent with previous light-microscopic evidence for internalization and retrograde transport of intrastriatal neurotensin within nigrostriatal dopaminergic neurons. They further suggest that retrogradely transported neurotensin may be processed along a variety of intracellular pathways including those mediating degradation in lysosomes and recycling in rough endoplasmic reticulum. The detection of specific autoradiographic labelling in the nucleus supports the concept that neurotensin alone, or complexed to its receptor, might be involved in the regulation of gene expression through direct or indirect interactions with nuclear DNA. Consequently, the retrograde transport of neurotensin in nigrostriatal dopaminergic neurons might provide a vehicle through which events occurring at the level of the axon terminal may initiate long-term biological responses.

Neurotensin is a tridecapeptide which was initially isolated from bovine hypothalamus 6 and later found to be present within selective neuronal populations widely distributed throughout the mammalian C N S . 1'17'27'32 Considerable anatomical, biochemical and pharmacological evidence has accumulated in support of a neurotransmitter role for neurotensin. 7'14'26'38'6° Accordingly, specific high-affinity neurotensin binding sites have been identified throughout the brain of numerous species wherein they have been shown to display a selective association with certain neuronal populations. Thus, neurotensin binding sites have been selectively localized to dopaminergic (DA) neurons in the rat substantia nigra pars compacta (SNC) and ventral tegmental area (VTA) 45'58which probably account for

++To whom correspondence should be addressed. Abbreviations: DA, dopamine/dopaminergic; HPLC, high performance liquid chromatography; SN, substantia nigra; SNC, substantia nigra pars compacta; SNR, substantia nigra pars reticulata; VTA, ventral tegmental area; 6-OHDA, 6-hydroxydopamine.

the numerous biochemical, behavioural, and electrophysiological effects demonstrated for neurotensin on midbrain D A neurons. 29'31'42The demonstration that S N C and VTA neurons also express the m R N A encoding the neurotensin receptor recently reinforced the concept of a selective association of neurotensin receptors with these cells. 16 Electron microscopic studies revealed that these receptors were more or less uniformly distributed along the somatodendritic plasma membrane of D A neurons in the midbrain ~t and cytotoxic lesion studies suggested that they were also present at the level of their axon terminals in the neostriatum. 2°'23'4°'48'53 In keeping with this interpretation is the demonstration of a presynaptic neurotensin-induced potentiation of basal and potassium-stimulated D A release from rat striatal slices ~2"18'24'43 and synaptosomal preparations. 28 Neurotensin receptors on nigrostriatal D A terminals have also been implicated in the neurotensin-induced stimulation of the phosphorylation of synaptosomal proteins demonstrated in the striatum? In vivo studies suggested that neurotensin receptors on nigrostriatal D A axon terminals were internalized 269

270

M-N. ('ASTEL et al.

and it was proposed that this internalization was a necessary step for the retrograde labelling observed in the midbrain tegmentum following unilateral injection of [~2~I]neurotensin into the neostriatum. 9 Film autoradiographic studies indeed indicated that this retrograde labelling was confined to the ipsilateral S N C and adjacent VTA, the two regions of origin of the D A innervation of the striatum. 6-Hydroxydopamine ( 6 - O H D A ) lesions of SNC neurons essentially abolished this labelling, thereby confirming the dopaminergic selectivity of the retrograde marker. 9 Intrastriatal injections of [~25I]neurotensin at concentrations above 0.2 pM failed to enhance levels of radioactivity in the ipsilateral SNC in keeping with a saturable, receptor-mediated effect. Accordingly, co-injection of [~25I]neurotensin with an excess of unlabelled neurotensin or of the biologically active fragment neurotensin8 ~3 prevented the appearance of radioactivity in the ipsilateral SNC. 9 The precise cellular mechanisms responsible for this retrograde labelling are still unclear. The time-course of the accumulation of radioactivity in the SNC and the absence of labelling after colchicine treatment are compatible with a rapid axonal transport mechanism. 9 High performance liquid chromatography (HPLC) analysis revealed that, 2 4 h after intrastriatal injection, an important fraction of the retrogradely transported radioactivity corresponded to intact neurotensin. The remainder was attributed to partially hydrolysed fragments of retrogradely transported neurotensin. 8 These biochemical results corroborate the hypothesis that the observed transport is initiated by a receptor-mediated internalization; they have been interpreted to suggest that it may subserve a transcellular transfer of information. In an attempt to elucidate the cellular mechanisms and morphological substrates mediating this process, and thereby provide insight into its functional significance, we have employed light and electron microscopic autoradiography to examine the topographic and cellular localization of retrogradely transported radioactivity within nigrostriatal neurons 4.5 h after the injection of [~25I]neurotensin into the neostriatum of the rat. EXPERIMENTAL PROCEDURES

Materials The metallopeptidase inhibitor kelatorphan was generously provided by Bernard Roques (Paris V, France). Neurotensin (Sigma) was iodinated and purified by a modification of the lactoperoxidase-H202 protocol of Sadoul et al. 5~ lntrastriatal injection o f [125I]Tyr3 neurotensin Adult male Sprague-Dawley rats (Charles River; 185-200 g) were injected stereotaxicatly under sodium pentobarbital anaesthesia (65 mg/kg i.p.) with 2/~1 of the peptidase inhibitor kelatorphan (15mg/ml) into the right caudoputamen at a rate of 0.2/tl/min (co-ordinates A, -0.3 mm, L, 3.6 ram, V, - 6 mm with respect to bregma). Ten minutes later, 16pmol of [~25I]Tyr3 neurotensin

(74TBq/mol) dissolved in 3 Itl physiological saline ,acre infused at three different rostrocaudal levels into the right striatum (coordinates: A, t.6, -0.8, and -3.1 ram; L, 2.6, 4.1, and 5mm; V, -5.8, -6.4, --6.6mm with respecl to bregma). The animals were killed 4.5 h later, a duration chosen on the basis of previous time-course studies demonstrating that the maximal accumulation of radioactivity in the ipsilateral substantia nigra (SN) occurred 4 h subsequent to intrastriatal injection. Tissue preparation Animals were anaesthetized and perfused transaortically with a 3.5% solution of glutaraldehyde in 0.1 M P O 4 buffer. Following removal of each brain from the skull, the midbrain was blocked and postfixed overnight by immersion in the same fixative at 4°C. Serial sections were cut with a Vibratome (Lancer) in the coronal plane at alternating thicknesses of 30 and 60 pm and collected in 0.1 M PO 4 buffer. Autoradiographic processing The 30-pm sections were processed for topographic localization of retrogradely transported [a~SI]neurotensin in the midbrain. They were mounted onto gelatin-subbed slides, dried at room temperature, dehydrated in graded ethanols, defatted in xylene, and rehydrated. They were then coated by dipping into Kodak NTB-2 liquid emulsion diluted 1:1 with distilled water at 40°C, and exposed for six weeks at 4:(2 in light-tight boxes. These autoradiographs were developed at 20°C for 20 rain in Kodak D-19 developer, and fixed in a 30% solution of sodium thiosulfate. They were then counterstained with Cresyl Violet, dehydrated in graded ethanols, and coverslipped. The alternate 60-/~m sections were processed lbr cellular localization of the retrogradely transported radioactivity. They were postfixed for I h in 2% OsO4 in 0.1 M phosphate buffer containing 7% dextrose, dehydrated in graded ethanols, and flat-embedded in Epon between plastic coverslips. In some cases, sections were stained en bloc for 90 min with 2% uranyl acetate. Subsequent to polymerization of the embedding medium for 19 h at 60°C, the sections were re-embedded in Beem capsules. Regions of the SNC were then trimmed and blocked on the basis of the labelling observed in the film autoradiographs. Semithin sections (1 #m) were cut on a Reichert---Young ultramicrotome, collected on acid-washed glass slides, and dipped in Kodak NTB-2 liquid emulsion diluted 1:1 in distilled water at 40°C. Following seven weeks' exposure at 4°C in light-tight boxes, these semithin sections were developed in freshly prepared Kodak D-19 for 4rain at 17°C, counterstained with Toluidine Blue, and examined with a Leitz~)rthoplan microscope. Blocks from which the semithin sections were obtained were re-trimmed and ultrathin (80 nm) sections were cut using an ultramicrotome and collected on parlodion-coated glass slides. The slide-mounted ultrathin sections were then stained with uranyl acetate and Reynold's lead citrate, carbon-coated, and dipped in Itford L-4 liquid emulsion diluted 1:4 in distilled water, Subsequent to 2.5-4 months' exposure at 4°C in light-tight boxes, the autoradiographs were developed in Kodak D-19 developer for 90s at 18°C, fixed in 300 sodium thiosulfate for 10 rain at 4°C and rinsed in distilled water at 4°C. The parlodion membranes were floated off of the glass slides and the ultrathin sections were collected on bare 200-mesh copper grids. Subsequent to thinning of the parlodion membrane with amyl acetate, the electron microscopic autoradiographs were examined using a JEOL 1200 EX electron microscope. Quantitative analysis of electron-microscopic autoradiograms Ultrathin sections from four different animals were systematically scanned using the electron microscope and every silver grain or group of silver grains photographed at a

Retrograde transport of neurotensin

271

Fig. 1. Darkfield photomicrograph of a 30-#m-thick coronal section demonstrating the topographical distribution of labelled neurons in the ipsilateral mesencephalon 4.5 h following intrastriatal injection of [~25I]neurotensin. The midline of the section is at the right edge of the micrograph and the base of the brain is at the bottom (for orientation, see Fig. 2 in Castel et al.9). Labelled neurons are confined to the ipsilateral SNC and adjacent VTA. The pars reticulata (SNR) is essentially devoid of labelling. Note the presence of retrogradely labelled axons throughout the VTA. Scale bar = 0.5 mm. magnification of x 14,000. This yielded a sampling of approximately 2,100 silver grains whose cellular distribution was documented following visual analysis of the electron micrographs. Silver grains were assigned to cellular elements using a modification of the 50% probability circle method of Williams. 6~ A 50% resolution circle was centred over each grain and the structure(s) included within the circle documented. Those identified over a single structure were classified into one of the following compartments: neuronal soma, neuronal nucleus, dendrite, myelinated axon, unmyelinated axon, axon terminal, myelin, astrocyte or oligodendrocyte. Those overlying the apposing membranes of two or more adjacent elements were categorized as membrane-associated, based on previous autoradiographic line source analyses which had shown that more than 90% of such shared grains originated from one of the apposed membranes, n'39 Furthermore, shared grains associated with interfaces involving neuronal somata, dendrites or axons were systematically attributed to the somatic, dendritic, and axonal membranes, respectively, on the basis of correlative light and electron microscopic evidence for preferential intracytoplasmic labelling of these structures. In order to evaluate whether silver grains overlying dendritic membranes were selectively concentrated opposite any particular structure, their distribution frequency was compared with the occurrence frequency of structures apposing dendrites. A distribution of randomly distributed silver grains along dendritic membranes was generated by placing a transparent overlay on each electron micrograph displaying a labelled dendritic profile. The overlay consisted of a simple grid of parallel horizontal and vertical lines, each separated by 5cm (corresponding to approximately 1.5 #m), At each point of intersection of the grid with the dendritic membrane of a labelled dendrite, the structure apposing the dendritic membrane was noted. This random membrane distribution of silver grains was compared with the frequency with which real grains were associated with each type of apposing structure using Z 2 analysis. Estimation of nuclear relative to perikaryal surface area in labelled neurons was achieved in semithin sections using a computer-assisted morphometry (Bioquant)

programme. The nuclei and soma of 60 labelled neurons were circumscribed and the relative perimeters and surface areas were tabulated automatically. A ratio of nuclear to cell body surface area was calculated from these measurements.

RESULTS The t o p o g r a p h i c distribution of neurons retrogradely labelled in the mesencephalon 4.5 h subsequent to injection o f [~25I]neurotensin into the right striatum is illustrated in Fig. 1. Distinct accumulations o f silver grains were a p p a r e n t over neuronal cell bodies and intervening processes within the ipsilateral substantia nigra, VTA, and retrorubral field. The dopaminergic neuronal constituents o f these regions comprise the A9, A10, and A8 cell groups o f D a h l s t r o m and Fuxe, l° respectively. N o autoradiographic labelling was detected on the contralateral side. Within the SN, retrogradely labelled perikarya were p r o m i n e n t t h r o u g h o u t the rostrocaudal and lateromedial extents o f the SNC (Fig. 1). Small clusters o f retrogradely labelled cells were also detected in the pars reticulata (SNR). Within the VTA, labelled neurons were less n u m e r o u s and restricted mainly to the parabrachial pigmented subdivision located immediately ventral to the medial lemniscus (Fig. 1). In contrast, the more ventrally situated paranigral subdivision o f the VTA was essentially devoid o f autoradiographic labelling. Labelled neurons in the retrorubral field were sparsely distributed. These cells were continuous ventrally with those observed in the lateral aspect o f the caudalmost

Fig. 2. Distribution of retrogradely transported radioactivity in semithin (I/am) sections through the ipsilateral SNC processed 4.5 h following ntrastriatal injection of [12SI]neurotensin. Silver grains are mainly concentrated over the perinuclear cytoplasm of neuronal cell bodies. Some ire also apparent over nuclei (arrowheads in a and h) and proximal dendritic processes (arrows in c). Stray silver grains are detected over the neuropil, including myelinated fibre bundles (d). Scale bar = 20/am.

.q

2

~o

Retrograde transporl of neurotensin

273

Fig. 3~ Electron microscopic radioautograms from sections of the SNC 4.5 h subsequent to ipsilateral intrastriatal injection of [~25I]neurotensin. (a) Two silver grains overlie the nucleus of an oligodendrocyte. Another is associated with an adjacent unmyelinated axon. (b) Two silver grains are associated with the myelin sheath of a myelinated axon and another is localized at the interface between an axon terminal (AT) and a myelinated axon. (c) Two silver grains localized at the border of a myelinated axon and its myelin sheath. Scale bars = 0.5 ~m. SNC and extended medially, dorsal to the medial lemniscus, to merge with labelled neurons in the caudal VTA. Light-microscopic examination of semithin sections through the retrogradely labelled S N C revealed distinct accumulations of silver grains over a subpopulation of neuronal cell bodies (Fig. 2). These cell bodies were fusiform, round, oval, or multipolar in shape. Many displayed a large nucleus with a prominent, centrally located nucleolus. Morphometric analysis revealed that the average surface area of these nuclei was approximately one-third that of the entire cell soma (nucleo-cytoplasmic ratio; 1:2.4). Silver grains were most prevalent over the perinuclear

cytoplasm and, in favourable planes of section, were seen to extend into proximal dendritic processes (Fig. 2c). A small proportion, however, were also detected over the nucleus (Fig. 2a, b). Sparse autoradiographic labelling was also apparent over the neuropil, mainly in areas displaying bundles of myelinated axons (Fig. 2d) as well as over the nuclei and perinuclear cytoplasm of glial cells. Electron microscopic analysis of autoradiographic labelling in the SNC confirmed and extended the light microscopic results. Quantitative evaluation of the electron microscopic data revealed that more than 85% of the silver grains sampled from the SNC were associated with neuronal profiles. O f these, a sizeable

274

M.-N. CASTt!Let

al.

Fig. 4. Electron microscopic radioautograms from sections of the SNC 4.5 h following ipsilateral intrastriatal injection of [~SI]neurotensin.(a) Section through the cytoplasm of a neuronal cell body. Silver grains are visible over the rough endoplasmic reticulum. (b) Cross-section of a neuronal cell body displaying autoradiographic labelling over a multi-vesicularbody. (c, d) Two cross-sectionedneuronal cell bodies show labelling over the nucleus (Nu). (e) A silver grain overlies a bundle of cross-sectioned, unmyelinated axons. Scale bars = 0.5 ~trn. proportion (32% of the total number of grains counted) were identified over myelinated and unmyelinated axons (Figs 3b, c, 4e, Table 1). Labelling associated with unmyelinated axons was usually detected over the plasma membrane (Fig. 4e, Table 1) which was to be expected given the small diameter of these axons relative to the size of silver grains. A substantial proportion of the silver grains associated with myelinated axons was similarly detected over the membrane, at the interface with its myelin sheath (Fig. 3b, c, Table 1). Moreover, many were localized exclusively over the myelin sheath itself (Table 1). Consistent with this labelling of myelin, a small proportion of silver grains (4%) was detected in oligodendrocytes where they were often identified over the prominent nucleus (Fig. 3a). Approximately 13% of silver grains were detected over nerve cell bodies (Table 1). Seventy-seven per cent of these somatic grains overlaid the cytoplasm.

The remainder were detected over the nucleus (16%) or the perikaryal membrane (7%). In a relatively large proportion of intracytoplasmic grains, their resolution circles encompassed clearly identifiable organelles (Figs 4, 5). Among these, thc most notable were rough endoplasmic reticulum and mitochondria (Figs 4a, 5a, b). Less frequently, the resolution circle included lysosomes, Golgi apparatus or multivesicular bodies (Figs 4b, 5b). Although silver grains often impeded the detection of small, underlying organelles, small clear or large dense-cored vesicles were occasionally identified either within or in close proximity to the resolution circle (Fig. 5c). Of the silver grains detected over the nucleus, many were identified directly over or next to accumulations of heterochromatin (Fig. 4c, d). Silver grains detected over the perikaryal plasma membrane displayed no selective association with any particular element apposing the cell body. Concomitantly labelled

275

Retrograde transport of neurotensin Table 1. Distribution of retrogradely transported radioactivity in rat substantia nigra after injection of [~25I]neurotensinin neostriatum Structure Nerve cell bodies Cytoplasm Nucleus Membrane Dendrites Cytoplasm Membrane Myelinated axons Cytoplasm Membrane Myelin sheath Unmyelinated axons Cytoplasm Membrane Axon terminals Glia**

% Total number % Grains per of grains* structure 12.6 9.8 _+ 1.4 2.0 _+0.35 0.8 _+0.1 34.7 12.8 + 1.5 21.9 _+ 1.7 22.6 5.6 _+0.75 11.2 + 1.5 5.8 + 0.6 15.4 1.6 _+0.4 13.8 + 1.4 7.1 _+0.4 7.6 _+0.7

100 77 16 7

100 37 63 100 25 49 26 100 11 89 100 100

*Mean+S.E.M. of nine sections from four different animals. **Cumulative counts for astrocytes and oligodendrocytes.

elements included dendrites, axons, axon terminals, and astrocytes. Almost 35% of silver grains were associated with dendritic profiles (Table 1). Most of these dendrites were relatively large in cross-sectional diameter and displayed a variety of intracytoplasmic organelles including rough endoplasmic reticulum, Golgi apparatus, lysosomes, mitochondria, and multivesicular bodies (Fig. 6). Silver grains were also associated with small dendritic profiles, although much less frequently. In contrast to what was observed in cell bodies, only 37% of the labelling detected in dendrites was intracytoplasmic (Fig. 6a, b, Table 1). The remainder was identified at the periphery of dendritic profiles, concurrently overlying the plasma membrane of adjacent structures (Fig. 6c, d). A few grains were also observed over structures abutting the labelled dendrites, at a short distance from their plasma membrane. Intracytoplasmic grains frequently encompassed microtubules and/or mitochondria (Fig. 4a, b). Membrane-associated grains were detected opposite axons, axon terminals, glial cells, and other dendrites. Approximately 18% were found over synaptic contacts (Fig. 6d). Quantitative analysis of randomly-generated points along dendritic membranes revealed no significant differences between the frequency of occurrence of dendroaxonal, dendrodendritic, dendroglial, dendrosomatal and axodendritic appositions in the SNC and the relative proportions of silver grains detected at each of these interfaces (Fig. 7). Consequently, despite the apparent heterogeneity of the association of labelling with distinct apposing structures, these data are consistent with a non-selective distribution of the labelling along dendritic plasma membranes. Approximately 7% of the silver grains associated with neurons were associated with axon terminals

(Table l). These terminals were round, flattened~ or ovoid in shape and most contained numerous densely packed, small, clear vesicles. Some of these labelled terminals established synaptic contacts with labelled and unlabelled neuronal elements. Finally, a small proportion (3.6%) of silver grains was detected over astrocytes where they overlaid both cell bodies and processes.

DISCUSSION

The present electron microscopic data confirm and extend our previous film autoradiographic observations concerning the retrograde labelling of nigral neurons following unilateral injection of monoiodinated neurotensin into the rat neostriatum. Four and a half hours after [~25I]neurotensin injection, retrogradely transported radioactivity was confined to the side ipsilateral to the injection where it was prevalent within neuronal cell bodies in the SNC, the adjacent VTA, and, further caudally, the retrorubral field. Labelled neurons in each of these regions displayed a pattern of distribution similar to that demonstrated previously for dopaminergic mesostriatal neuronsJ 3'19 Moreover, on examination of thionine-stained semithin sections, labelled cell bodies exhibited morphological and cytological features characteristic of mesencephalic DA cells. '~ Electron microscopic analysis revealed that the majority of silver grains (>85%) were associated with neuronal structures including cell bodies, dendrites, axons, and axon terminals. Probability circle analysis was employed to assign the autoradiographic labelling to specific cellular sources. Silver grains overlying single neuronal or glial structures that were large relative to the size of the grains could be readily attributed to intracellular radioactive sources. Conversely, it was assumed on the basis of previous electron microscopic autoradiographic analyses ~39 that those overlying cellular membrane interfaces had arisen from membrane-associated radioactive sources. Furthermore, shared grains detected over the interfaces of neuronal somata, dendrites, and axons with other structures were systematically attributed to sources within the somatic, dendritic or axonal membrane itself on the basis of the present correlative light and electron microscopic evidence for a preferential internal labelling of perikarya, dendrites, and axons in the midbrain tegmentum. Therefore, the possibility that some of the silver grains ascribed to neuronal plasma membranes had in fact arisen from adjacent structures cannot be formally excluded.

Nature of retrogradely transported radioactivity We have previously demonstrated that 4 h subsequent to intrastriatal injection of [t25I]neurotensin, approximately 30% of the radioactivity detected in the ipsilateral SN corresponded to partially hydrolysed [~25I]neurotensin fragments containing the

276

M.-N. (ASTI~L et

al.

Fig. 5. Three retrogradely labelled neuronal cell bodies subsequent to ipsilateral intrastriatal injection ot [~25I]neurotensin. (a) One of the silver grains partially overlies the nuclear membrane (arrowheads) while the other is detected over rough endoplasmic reticulum (ER). (b) Two silver grains are seen over the nucleus (arrows) and the other over the Golgi apparatus (Gol). (c) Here again, one o f the grains lies over the nucleus (Nu) and the other is in the vicinity of a mitochondrion (arrow) and vesicular elements (arrowheads). Scale bars = 0.5 :tm.

Retrograde transport of neurotensin

Fig. 6. Cross-sections of dendrites retrogradely labelled in the SNC 4.5 h following ipsilateral intrastriatal injection of [125I]neurotensin. (a) One silver grain is localized over the plasma membrane (arrow) and the other is associated with a mitochondrion. (b) An intracytoplasmic grain is seen in the vicinity of a dendritic appendage filled with clear vesicles (arrowheads). (c) Two silver grains overlie the plasma membrane of a large dendritic shaft as well as an adjacent astrocytic leaflet (arrows). (d) The silver grain on the right overlies an axon terminal synapsing upon a cross-sectioned dendritic shaft. Its resolution circle encompasses both the m e m b r a n e of the terminal and that of the recipient dendrite as well as part of the synaptic specialization (arrow). The silver grain on the left encompasses the plasma membrane of an adjacent unmyelinated axon (arrowheads). Scale bars = 0.5 Izm.

277

278

M.-N. (IASTEL ~'[ ~11.

50 [ ] Real Grains

o')

._= (D 40 "o

~ 30 m 20 c E ®

10

g 0

GL

AX

SYN

TERM

DEN

SOM

Fig. 7. Histogram of the frequency of association of real vs randomly generated silver grains along dendritic membranes with specific apposing structures. Values are expressed as the mean percentage of the total grains detected over dendritic membranes __+S.E.M. GL, glial cells; AX, axons; SYN, synapses; TERM, axon terminals; DEN, dendrites; SOM, neuronal somata.

labelled Tyr-3 residue, or to the Tyr-3 residue itself: All of these fragments were therefore susceptible to covalent cross-linking by glutaraldehyde, provided that they were endowed with a free amino g r o u p : 6 Free tyrosine could also have been cross-linked through its phenolic rings which have been shown to react with glutaraldehyde. 22 Free radioactive iodine is not intermolecularly cross-linked to the tissue and is therefore likely to have been washed out during histological processing. Furthermore, given that no radioactive fragments eluted behind neurotensin in our HPLC system, radioactive molecules detected in the SNC are unlikely to correspond to neosynthesized proteins having incorporated the [~25I]Tyr moiety. The repartition of intact and partially metabolized forms of retrogradely transported neurotensin among distinct subcellular domains has not been reliably established. However, intrastriatal injections of [125I]neurotensin in the absence of kelatorphan, an inhibitor of neurotensin-degrading metallopeptidases, completely prevents the appearance of radioactivity in the ipsilateral SNC. This suggests that the shortened forms of neurotensin present in the SNC do not correspond to fragments internalized as such, but instead to fragments generated by the intracellular hydrolysis of intact neurotensin in the axon or in the cell body. The fact that the amount of intact neurotensin detected in the SNC decreases from 44 to 30% between 2.5 and 4 h following intrastriatal [~25I]neurotensin injection supports this interpretation. It also indicates that an important proportion of the radioactivity detected in axons corresponds to unmetabolized neurotensin in transit from the striatum to the SNC.

Axonal labelling The detection of silver grains over both myelinated and unmyelinated axonal profiles lends support to the

concept of an intra-axonal transport of neurotensin between the striatum and the substantia nigra. 8'9 Given the implication of receptor-mediated internalization in the initiation of this axonal transport, 33'53 the fact that many of the silver grains were not even in the vicinity of endosome-like vesicular profiles was quite unexpected. A possible interpretation for these results is the existence of a non vesicular transport mechanism. In favour of this interpretation is the demonstration of intracytoplasmic pools of immunoreactive neurotensin in neurotensin target cells within the ventral nucleus of the amygdala? A non-vesicular transport of intact neurotensin is predicated on the existence of mechanisms preventing its intracytoplasmic hydrolysis. The protection of neurotensin from intracytoplasmic peptidases might be fulfilled by its association with the neurotensin receptor in the form of a receptor-ligand complex. Indeed, previous autoradiographic studies have provided evidence for both the presence and the bidirectional transport 3° of neurotensin binding sites in axons. H Alternatively, the possibility exists that the label detected over vesicular compartments corresponds to intact neurotensin, whereas non-vesicular intra-axonal labelling mainly corresponds to degradation fragments. A substantial proportion of silver grains detected in axons were localized over the plasma membrane. In the case of unmyelinated axons, this localization could merely reflect the small diameter of these axons relative to that of silver grains. However, the consistent localization of silver grains over the plasma membrane and myelin sheath of large-diameter myelinated axons is more puzzling. A possible interpretation for these results is that retrogradely transported neurotensin or neurotensin degradation fragments are translocated through the axonal plasma membrane to the adjacent myelin sheath. It is interesting to recall in this context that binding sites

Retrograde transport of neurotensin

279

for neurotensin~l and other neurotransmitters35 have neurons comprising the rat nucleus basalis of Meynert. 1~'59 Favouring the hypothesis of a nuclear been demonstrated on oligodendroglia and that evidence exists for a glial localization of the neuro- translocation of the neurotensin receptor ligand tensin-degrading metalloendopeptidase 24.11.2s'37'3~ complex is the considerable sequence homology Extra-axonal translocation of retrogradely trans- shared by the neurotensin receptor with several ported neurotensin might therefore render retro- nuclear binding proteins. 8 It is worthy of note that nuclear grains were often found over areas of chrogradely transported neurotensin susceptible to hydrolysis by glial cells. The mechanisms by which matin condensation, an association implicated~ in the such a precept might occur remain a matter of case of the insulin receptor-ligand complex, in the speculation. An attractive hypothesis is that the physiological substrates underlying insulin's longneurotensin receptor-ligand complex associates with term biological effects. 47We recently demonstrated an the axonal membrane. This association results in the increase in the expression of tyrosine hydroxylase incorporation of the receptor component into the mRNA in the SNC following the intra-striatal injecplasma membrane of the axon and the extra-neuronal tion of neurotensin.3 Such an effect might be interliberation of the intact neurotensin molecule. The preted as evidence for alterations in gene expression demonstration of specific, high-affinity neurotensin resulting from the nuclear translocation of retrobinding sites over axonal membranes It substantiates gradely-transported neurotensin. The existence of such a mechanism would provide an intracellular this hypothesis. substrate through which adaptive modifications in Perikarval labelling gene expression at the level of the cell body may be mediated by events occurring at the level of the axon Within nerve cell bodies, silver grains were frequently detected either directly over or in the terminal. immediate vicinity of intracellular organelles including vesicles, rough endoplasmic reticulum, mitochon- Dendritic labelling dria, Golgi apparatus, lysosomes, and multivesicular The localization of silver grains over axons and bodies. This pattern of intracytoplasmic labelling is nerve cell bodies is consistent with a retrograde consistent with contemporary models concerning the transport of neurotensin along axons and subsequent sequence of intracellular events following the endo- processing in neuronal perikarya. However, the cytosis of recepto~ligand complexes. 4's7 Subsequent detection of substantial autoradiographic labelling to internalization, the complexes are sequestered over dendritic profiles is more difficult to interpret. in an acidic endosomal compartment where they The presence of silver grains within dendrites implies undergo dissociation into separate ligand and recep- an intradendritic transport of retrogradely transtor components. The ligand and receptor are then ported neurotensin, or metabolites thereof, derived recycled in the rough endoplasmic reticulum or are from the perikaryon. A similar dendritic localization translocated to lysosomes for enzymatic degradation. has been reported for nerve growth factor in the Alternatively, the ligand or ligand-receptor complex peripheral sympathetic nervous system. ~6 In the premay be transported intracytoplasmically to interact sent study, most of the dendritic grains were associwith secondary intracellular targets. The present ated with the plasma membrane. The fact that these results implicate the nucleus as a potential candidate membrane-associated grains displayed no preferenfor one of these destinations. Indeed, not only was a tial localization opposite any one of the abutting significant proportion of silver grains associated with structures makes it unlikely for this association to the nucleus, but this proportion did not conform to reflect a specific transfer of information between the the nuclear/cytoplasmic surface area ratio, suggesting labelled dendrites and adjacent structures. They that the observed labelling was not attributable to could, however, correspond to neurotensin or neuropassive mobilization of the radioactivity in the course tensin fragments being non-selectively translocated to of fixation. Admittedly, the possibility that diffusion the extracellular space or to neighbouring elements. of small neurotensin degradation fragments from Such a mechanism would account for the presence of the cytoplasm across the nuclear membrane might label over nearby astrocytic processes. Translocation account for a fraction of the nuclear labelling cannot to glial cells could subserve extra-neuronal hydrolysis be entirely excluded. However, this labelling could of intact neurotensin. Consistent with this possibility also reflect an active translocation of neurotensin is the presence of high-affinity neurotensin binding itself, or of the neurotensin receptor-ligand complex sites demonstrated on ventral mesencephalic glial as has been described for a variety of polypeptide cells l~ as well as the relative abundance in the hormones and growth factors including insulin, SNC and VTA of astrocytes containing the thyroliberin, gonadoptropins, epidermal growth fac- neurotensin-degrading metalloendopeptidase 24.16 tor, and vasoactive intestinal peptide. 21"36"41'44"49"5°x'2 (Woulfe J. et al., unpublished observations). AlternaEvidence for the concept of a nuclear translocation of tively, it could provide for the catabolism and/or neurotensin itself is the demonstration of high-affinity recycling of short neurotensin degradation fragments neurotensin binding sites within the nuclei of mid- or free tyrosine. Indeed, astrocytes possess a wellbrain neurons and, more notably, those of cholinergic documented capacity for the accumulation of amino

280

M.-N CASTELet al.

acids from the extracellular space. 54'55The occurrence of silver grains over axon terminals could similarly reflect a translocation of neurotensin degradation products from dendritic compartments. Alternatively, it might imply the existence of an anterograde transport of neurotensin molecules from striatonigral neurotensinoceptive neurons. The later interpretation appears unlikely, however, in light of our previous demonstration that 6 - O H D A lesions of the dopaminergic neurons in the SNC essentially abolish nigral radioactivity subsequent to intra-striatal neurotensin injection. Moreover, Schotte and Leysen 52 recently demonstrated that virtually all of the high-affinity neurotensin binding sites in the striatum were localized on the axon terminals of dopaminergic afferents. CONCLUSIONS The present evidence for retrogradely transported neurotensin in nigrostriatal neurons is consistent with the existence of a dynamic intraneuronal process through which information concerning events at the level of the synapse might access the cell body. On the basis of the anatomical data described in the present study, we postulate a model whereby retrogradely transported neurotensin is processed through a variety of intraneuronal pathways. Subsequent to the binding of neurotensin to its receptor, the receptor-ligand complex is internalized at axon terminals within the striatum and transported retrogradely, through both vesicular and non-vesicular processes,

to nigrostriatal dopaminergic perikarya in the SNC. En route to the cell body, a proportion of intact neurotensin and/or neurotensin metabolites undergoes extraneuronal translocation into the extracellular space and possibly into myelin. At the level of the cell body, a proportion of the internalized complexes remains intact while the remainder undergo dissociation into independent ligand and receptor components. Neurotensin is subsequently sequestered in lysosomes where it is enzymatically degraded. The latter process generates partially hydrolysed forms of [~25I]neurotensin, some of which are delivered to the rough endoplasmic reticulum for recycling and the others transported into dendrites and eventually released into the extracellular space. A proportion of retrogradely transported neurotensin, or the intact receptor-ligand complex, is translocated to the nucleus. This retrogradely transported molecular information might be exploited by the cell to generate alterations in gene expression in response to distal synaptic events. A direct interaction of retrogradely transported neurotensin with genomic substrates effects long-term biological consequences, many of which may be related to the physiological mechanisms underlying neurogenesis, neural plasticity, and even long-term memory. 34

Acknowledgements--We wish to thank Beverley Lindsay,

Kathy Leonard and Charles Hodge for their excellent clerical assistance. This study was supported by the Medical Research Council of Canada.

REFERENCES

1. Alexander M. J., Miller M. A., Dorsa D. M., Bullock B. P., Meltoni R. H. Jr, Dobner P. R. and Leeman S. E (1989) Distribution of neurotensin/neuromedin N mRNA in rat forebrain: unexpected abundance in hippocampus and subiculum. Proc. hath. Acad. Sci. U.S.A. 86, 5202-5206. 2. Bayer V. E., Towle A. C. and Pickel V. M. (1991) Vesicular and cytoplasmic localization of neurotensin-like immunoreactivity (NTLI) in neurons postsynaptic to terminals containing NTLI and/or tyrosine hydroxylase in the rat central nucleus of the amygdala. J. Neurosci. Res. 30, 398-413. 3. Burgevin M.-C., Castel M. N., Quarteronnet D., Chevet T. and Laduron P. M. (1991) Neurotensin injected into the rat striatum increases tyrosine hydroxylase messenger RNA in the substantia nigra. Ann. N.Y. Acad. ScL (in press). 4. Burwen S. J. and Jones A. L. (1987) The association of polypeptide hormones and growth factors with the nuclei of target cells. Trends biochem. Sci. 11, 159-162. 5. Cain S. T., Abramson M. A. and Nemeroff C. B. (1988) Neurotensin stimulates the phosphorylation of caudate nucleus synaptosomal proteins. Ann. N.Y. Acad. Sci. 537, 488-490. 6. Carraway R. and Leeman S. E. (1973) The isolation of a new hypotensive peptide, neurotensin, from bovine hypothalami. J. biol. Chem. 248, 6854-6861. 7. Carraway R. and Leeman S. E. (1976) Characterization of radioimmunoassayable neurotensin in the rat. Its differential distribution in the central nervous system, small intestine and stomach. J. biol. Chem. 251, 7045-7052. 8. Castel M. N., Faucher D., Cuin6 F., Dub&iat P,, Boireau A. and Laduron P. M. (1991) Identification of intact neurotensin in the substantia nigra after its retrograde axonal transport in dopaminergic neurons. J. Neurochem. 56, 1816-1818. 9. Castel M. N., Malgouris C., Blanchard J. C. and Laduron P. M. (1990) Retrograde axonal transport of neurotensin in the dopaminergic nigrostriatal pathway in the rat. Neuroscience 36, 425-430. I0. Dahlstrom A. and Fuxe K. (1964) Evidence for the existence of monoamine containing neurons in the central nervous system. I. Demonstration of monoamines in the cell bodies of brainstem neurons. Acta physiol, scand., Suppl. 62, 1-55. l 1. Dana C., Vial M., Leonard K., Beauregard A., Kitabgi P., Vincent J. P., Rost6ne W. and Beaudet A. (1989) Electron microscopic localization of neurotensin binding sites in the midbrain tegmentum of the rat. I. Ventral tegmental area and interfascicular nucleus. J. Neurosci. 9, 2247-2257. 12. DeQuidt M. and Emson P. C. (1983) Neurotensin facilitates dopamine release in vitro from rat striatal slices. Brain Res. 274, 376-380. 13. Deutch A. Y., Goldstein M., Baldino F. Jr and Roth R. H. (1988) Telencephalic projections of the A8 dopamine cell group. Ann. N.Y. Acad. Sci. 537, 27-50.

Retrograde transport of neurotensin

281

14. Dobner P. R., Barber D. L., Villa-Komaroff L. and McKiernan C. (1987) Cloning and sequence analysis of cDNA for the canine neurotensin/neuromedin N precursor. Proe. nam. Acad. Sci. U.S.A. 84, 3516-3520. 15. Domesick V. B. (1988) Neuroanatomical organization of dopamine neurons in the ventral tegmental area. Ann. N.Y. Acad. Sci. 537, 10 26. 16. Elde R., Schalling M., Ceccatelli S., Nakanishi S. and Hokfelt T. (1990) Localization of neuropeptide receptor mRNA in rat brain: initial observations using probes for neurotensin and substance P receptors. Neurosci. Lett. 120, 134-- 138. 17. Emson P. C., Goedert M., Horsfield P., Rioux F. and St-Pierre S. (1982) The regional distribution and chromatographic characterization of neurotensin-like immunoreactivity in the rat central nervous system. J. Neurochem. 38, 992 999. 18. Faggin B. H., Zubrieta J. K., Rezvani A. H. and Cubeddu L. X. (1990) Neurotensin-induced dopamine release in vivo and m vitro from substantia nigra and nucleus caudate. J. Pharmac. exp. Ther. 252, 817 825. 19. Fallon J. H. (1988) Topographic organization of ascending dopaminergic projections. Ann. N.Y. Acad. Sci. 537, I 9. 20. Goedert M., Pittaway K. and Emson P. C. (1984) Neurotensin receptors in the rat striatum. Brain Res. 299, 164 168. 21. Goldfine I. D., Clawson G. A., Smuckler E. A., Purello F. and Vigneri R. (1982) Action of insulin at the nuclear envelope. Molee. cell. Biochem. 48, 3 14. 22. Habeeb A. F. S. A. and Hiromoto R. (1968) Reaction of proteins with glutaraldehyde. Archs Biochem. Biophys. 126, 16 23. 23. Herv6 D., Tassin J. P., Studler J. M., Dana C., Kitabgi P., Vincent J. P., Glowinski J. and Rost6ne W. (1986) Dopaminergic control of 125 I-labelled neurotensin binding site density in corticolimbic structures of the rat brain. Proc. nam. Acad. Sci. U.S.A. 83, 6203 6207. 24. Hetier E., Boireau A., Dubedat P. and Blanchard J. C. (1988) Neurotensin effects on evoked release of dopamine in slices from striatum, nucleus accumbens, and prefrontal cortex in rat. Naunyn Schmiedeberg's Arch. Pharmac. 337, 13 17. 25. Horsthemke B., Hamprecht B. and Bauer K. (1983) Heterogeneous distribution of enkephalin-degrading peptidases between neuronal and glial cells. Biochern. biophys. Res. Commun. 115, 423 429. 26. Iversen L. L., Iversen S. K., Bloom F. E., Douglas C., Brown M. and Vale W. (1978) Calcium-dependent release of somatostatin and neurotensin from rat brain in vitro. Nature 273, 161 163. 27. Jennes L., Stumpf W. and Kalivas P. W. (1982) Neurotensin: topographical distribution in rat brain immunochemistry. J. comp. Neurol. 210, 211-224. 28. Jiang D. H., Feng D. N., Zhu L. H. and Cheng Y. (1988) Neurotensin potentiates endogenous dopamine release from striatal synaptosomes of rat. Life Sei. 43, 27-33. 29. Kalivas P. W. (1985) Interactions between neuropeptides and dopamine neurons in the ventromedial mesencephalon. Neurosci. Behav. Rev. 9, 573--587. 30. Kessler J. P. and Beaudet A. (1989) Association of neurotensin binding sites with sensory and visceromotor components of the vagus nerve. J. Neurosci. 9, 466 472. 31. Kitabgi P. (1989) Neurotensin modulates dopamine neurotransmission at several levels along brain dopaminergic pathways. Neurochem. Int. 14, 111 --119. 32. Kobayashi R. M., Brown M. R. and Vale W. (1977) Regional distribution of neurotensin and somatostatin in rat brain. Brain Res. 126, 584 588. 33. Laduron P. M. (1984) Axonal transport of muscarinic receptors in vesicles containing noradrenaline and dopamine-Bhydroxylase. Fedn Eur. biochem. Soes Lett. 165, 128-132. 34. Laduron P. M. (1987) Axonal transport of neuroreceptors: possible involvement in long-term memory. Neuroscienee 22, 767 779. 35. Larocca J. N., Ledeen R. W,, Dvorkin B. and Makman M. H. (1987) Muscarinic receptor binding and muscarinic receptor-mediated inhibition of adenylate cyclase in rat brain myelin. J. Neurosci. 7, 3869 3876. 36. Laverri6re D., Gourdji R., Picart R. and Tixier-Vidal A. (1981) Thyroliberin is rapidly transferred to the nucleus of GH3 pituitary cells at both 4 degree C and 37 degree C. Biochem. biophys. Res. Commun. 103, 833 840. 37. Lentzen H. and Palenker J. (1983) Localization of the thiorphan-sensitive endopeptidase, termed enkephalinase A, oi1 glial cells. Fedn Eur. hiochem. Socs Lett. 153, 93 97. 38. Maeda K. and Frohman L. A. (1981) Neurotensin release by rat hypothalamic fragments in vitro. Brain Res. 210, 261 269. 39. Marcel D., Pollard H., Verroust P., Schwartz J. C. and Beaudet A. (1990) Electron microscopic localization of immunoreactive enkephalinase (EC 3.4.24.11) in the neostriatum of the rat. J. Neurosci. 10, 2804 2817. 40. Masuo Y., P61aprat D., Montagne M. N., Scherman D. and Rostene W. (1990) Regulation of neurotensin-containing neurons in the rat striatum and substantia nigra. Effects of unilateral nigral lesions with 6-hydroxydopamine on neurotensin content and its binding site density. Brain Res. 510, 203 210. 41. Millar R. P., Rosen H., Badminton M., Pasqualini C. and Kerdelhue B. (1983) Luteinizing hormone-releasing hormone (LIt-RH) binding to purifed rat pituitary nuclei. Fedn Eur. biochem. Socs Lett. 153, 382-386. 42. Nemeroff C. B. (1986) The interaction ofneurotensin with dopaminergic pathways in the central nervous system: basic neurobiology and implications for the pathogenesis and treatment of schizophrenia. Psychoneuroendocrinolgoy II, 15 37. 43. Okuma Y., Fukuda Y. and Osumi Y. (1983) Neurotensin potentiates the potassium-induced release of dopamine from rat striatal slices. Eur. J. Pharmac. 93, 27 33. 44. Omary M. B. and Kagnoff M. F. (1987) Identification of nuclear receptors for VIP on the human colonic adenocarcinoma. Science 238, 1578 1581. 45. Palacios J. M. and Kuhar M. J. (1981) Neurotensin receptors are found on dopamine-containing neurons in rat brain: an autoradiographic study. Nature 294, 587 589. 46. Peters T. and Ashley C. A. (1967) An artefact in radioautography due to binding of free amino acids to tissue by fixatives. J. Cell Biol. 33, 53-60. 47. Podlecki D. A., Smith R. M., Kao M., Tsai P., Huecksteadt T., Brandenburg D., Lasher R. S., Jarett L. and Olefsky J. M. (1987) Nuclear translocation of the insulin receptor. A possible mediator of insulin's long term effects. J. biol. Chem. 262, 3362 3368. 48. Quirion R., Chiueh C. C. and Pert A. (1985) Comparative localization of neurotensin receptors on nigrostriatal and mesolimbic dopaminergic terminals. Brain Res. 327, 385 389. NSC 50/2--B

282

M.-N. CASTEL et al.

49. Rakowicz-Szulczynska E. M., Rodeck U., Herlyn N. and Koprowski H. (1986) Chromatin binding of epidermal growth factor, nerve growth factor and platelet-derived growth factor in cells bearing the appropriate surface receptors. Proc. natn. Acad. Sci. U.S.A. 83, 3728-3732. 50. Rao C. V. and Milta S. (1979) Gonadotropin and prostaglandin binding sites in nuclei of bovine corpora lutea. Biochem. biophys. Acta 51t4, 454--466. 51. Sadoul J. L., Mazella J., Amar S., Kitabgi P. and Vincent J. P. (1984) Preparation of neurotensin selectively iodinated on the Tyr-3 residue: biological activity and binding properties on mammalian neurotensin receptor. Biochem hiophys. Res. Commun. 120, 812 819. 52. Schotte A. and Leysen J. E. (1989) Autoradographic evidence for the localization of high affinity NT binding sites on dopaminergic nerve terminals in the nigrostriatal and mesolimbic pathways in rat brain. J. chem. Neuroanat. 2, 253-257. 53. Schotte A., Rost6ne W. and Laduron P. M. (1988) Different subcellular localization of neurotensin-receptor and neurotensin-acceptor sites in the rat brain dopaminergic system. J. Neurochem..~0, 1026-1031. 54. Schousboe A. and Divac I. (1979) Differences in glutamate uptake in astrocytes cultured from different brain regions. Brain Res. 177, 407~-409. 55. Schousboe A., Fosmark H. and Svenneby G. (1976) Taurine uptake in astrocytes cultured from dissociated mouse brain hemispheres. Brain Res. 116, 158-164. 56. Schwab M. and Thoenen H. (1977) Selective trans-synaptic migration of tetanus toxin after retrograde axonal transport in peripheral sympathetic nerves: a comparison with nerve growth factor. Brain Res. 122, 459-474. 57. Shephard V. L. (1989) Intracellular pathways and mechanisms of sorting in receptor-mediated endocytosis. Trends pharmac. Sci. 10, 458-462. 58. Szigethy E. and Beaudet A. (1989) Correspondence between high affinity neurotensin binding sites and dopaminergic neurons in the rat substantia nigra and ventral tegrnental area: a combined radioautoradiographic and immunohistochemical light microscopic study. J. comp. Neurol. 279, 128--137. 59. Szigethy E., Leonard K. and Beaudet A. (1990) Ultrastructural localization of [~2~I]neurotensin binding sites to cholinergic neurons of the rat nucleus basalis magnocellularis. Neuroscience 36, 377-391. 60. Uhl G. R. and Snyder S. H. (1977) Neurotensin receptor binding, regional and subcellular distributions favour transmitter role. Eur. J. Pharmac. 41, 89-91. 61. Williams M, A. (1969) The assessment of electron microscopic autoradiography. In Advances in Optical and Electron Microscopy (eds Barer R. and Cosslett V. E.), Vol. 3, pp. 219-272. Academic Press, London. 62. Yankner B. A. and Shooter E. M. (1979) Nerve growth factor in the nucleus: interaction with receptors on the nuclear membrane. Proc. natn. Acad. Sci. U.S.A. 76, 1269-1273. (Accepted 23 April 1992)