Target area-specific patterns of fluorescence have been observed. for these dopamine-containing neurons in experimental

Transcription

1 Proc. NatL Acad. Sci. USA Vol. 78, No. 2, pp , February 1981 Neurobiology Target neuron-specific process formation by embryonic mesencephalic dopamine neurons in vitro (neuronal recognition/fluorescence histochemistry/aggregating cell cultures/catecholamine neurons) L. M. HEMMENDINGER*t, B. B. GARBER*, P. C. HOFFMANN, AND A. HELLER *Department of Biology and *Department of Pharmacological and Physiological Sciences, University of Chicago, Chicago, Illinois Communicated by Hewson Swift, October 20, 1980 ABSTRACT Mesencephalic dopamine neurons from the embryonic mouse brain were dissociated, aggregated in vitro in the presence of dissociated cells from appropriate or inappropriate target neuron areas, and visualized by the Falck-Hillarp histofluorescence technique after exposure to 1 ImM exogenous dopamine. When aggregated with the surrounding rostral mesencephalic tegmentum cells only or with the addition of rostral tectum cells, the dopamine neurons formed a dense dendritic arborization, but no axons were observed. In the presence of dopamine-neuron target cells from the corpus striatum, a dense axonal plexus characteristic of that formed in this area in vivo was observed. In contrast, in aggregates formed with target cells from the frontal cortex, branching fluorescent axons bearing irregularly spaced and shapd varicosities were found coursing through the neuropil, as is characteristic of the dopaminergic innervation to the frontal cortex in viva Only proximal dendrites were observed in the presence of these axonal target cells. Dopamine neurons cultured with inappropriate target cells from the occipital cortex did not form either extensive axonal or dendritic processes. Thus, the presence, type, and distribution of dopamine neuronal processes are dependent on the presence of appropriate target cells. The formation of unique patterns of neuronal processes by dissociated neurons in vitro suggests that the information necessary for this differentiation is intrinsic to the dopamine neurons and their target cells. This system provides a useful model with which to study basic mechanisms underlying neuronal recognition. The development of a functional nervous system involves a series of complex events, including at least two types of neuronal recognition. First, cell bodies of common origin in the germinal neuroepithelium, destined to have common morphological properties and functions in the adult nervous system, associate to form discrete nuclei or layers. Second, processes emanating from these neurons must seek out and form specific functional connections with other cells. The mechanisms governing the establishment of fully differentiated neurons possessing appropriate terminal fields and synaptic connections are not well understood, because of the difficulties of studying these phenomena in the maturing intact brain. Dopamine-containing neurons whose cell bodies are located in the mesencephalon offer certain advantages for studying the regulation of neuronal recognition and differentiation. Such cells and their processes can be visualized selectively by using fluorescence histochemical techniques (1, 2). In addition, their terminal fields in the striatum and frontal cortex have distinct morphological patterns (3). The axons, after leaving the substantia nigra pars compacta and ventral tegmental area in the mesencephalon and ascending along a common pathway in the medial forebrain bundle and medial internal capsule (4, 5), form a dense plexus of brightly fluorescent axonal varicosities in the striatum (6). In the frontal cortex, however, single branching axons course through the neuropil and bear infrequent, irregularly shaped varicosities (7, 8). The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U. S. C solely to indicate this fact Target area-specific patterns of fluorescence have been observed for these dopamine-containing neurons in experimental systems. Intraocular double grafts of the intact embryonic substantia nigra with the intact striatum or cortex resulted in the proliferation of fibers into these two areas by dopamine-containing neurons. The morphological patterns of such fiber proliferation closely resemble those that develop in vivo (9). Even when completely dissociated into single cells, embryonic mesencephalic dopamine-containing neurons exhibit a preferential reassociation with one another when subsequently reaggregated in tissue culture (10). In this system, coculture of the dissociated dopamine cells with dissociated cells of one of the normal target areas, the corpus striatum, resulted in the appearance of punctate fluorescent varicosities that suggest target cell-dependent axonal proliferation (11). We have extended these studies of dissociated embryonic dopamine cells of the mesencephalon to include another target area, the frontal cortex, as well as in areas that do not receive dopaminergic innervation; i.e., the occipital cortex and tectum. We found the presence, type, and distribution of dopamine neuronal processes that form in such cultures to be dependent on the presence of appropriate target cells. MATERIALS AND METHODS Dissection of Embryonic Mouse Brains. Embryos were removed from pregnant C57BL/6J mice (The Jackson Laboratory) on the 14th day of gestation and staged according to Gruneberg (12). The day on which a vaginal plug was observed was designated day 0 of pregnancy. Dissections were done under sterile conditions at room temperature in Tyrode's solution. Dissection of the rostral mesencephalic tegmentum (RMT), containing the dopamine neurons, was based on Golden's description of the developing catecholamine systems in the embryonic mouse (13) and verified by observation of glyoxylic acid/paraformaldehyde-perfused (14) 14-day embryonic mouse brains. To obtain the RMT, the cerebral lobes, diencephalon, and tectum were removed. The locus coeruleus and the mesencephalic raphe nuclei were eliminated by cutting just caudal to the mesencephalic flexure. The corpus striatum (CS) was separated from the septum and dissected free from the surrounding cortex, and the rostral (frontal; FCX) and caudal (occipital; OCX) thirds of the lateral cortex were removed. Dissociation and Aggregation. Brain tissue was dissociated and aggregated as described (15). Each brain area was minced into 1- to 2-mm3 pieces and incubated first for 20 min in Ca2+- Abbreviations: RMT, rostral mesencephalic tegmentum; FCX, rostral (frontal) cortex; OCX, caudal (occipital) cortex; CS, corpus striatum. tpresent address: Department of Neurology, State University of New York, Stony Brook, NY twe regret the death of Dr. Beatrice B. Garber on April 26, She is missed by her friends and colleagues. To whom reprint requests should be addressed.

2 Neurobiology: Hemmendinger et al Mg2e-free Tyrode's solution and then for 35 min in 0.67% trypsin (wt/vol in Ca2+-Mg2+-free Tyrode's solution; GIBCO) at 370C, ph 7.4. After thorough washing in Ca2+-Mg2+-free Tyrode's solution, the tissue was dissociated in 1 ml of initial culture medium [100 ml of Eagle's minimal medium (Microbiological Associates) to which was added 10 ml of fetal bovine serum (GIBCO), 1 ml of penicillin-streptomycin (5000 units of penicillin and 5000 pug of streptomycin per ml), 1 ml of L-glutamine (29 mg/ml) (Microbiological Associates), and 2.5 ml of a solution of DNase (Millipore) in Eagle's medium (1 mg/ml)] by gentle flushing through a fine-bore Pasteur pipette. Each suspension was checked microscopically to verify complete dissociation, and the number of cells per ml was counted with a hemocytometer. Approximately 3-6 X 106 RMT cells were placed in each 25-ml Erlenmeyer flask containing 3.5 ml of initial culture medium (ph 7.4). Other cell types from the same brains were added to the RMT cells to give the following (approximate) mixtures: RMT/CS at 1:2, RMT/FCX at 2:1, RMT/ OCX at 2:1; and RMT/tectum at 1:1. The flasks were fitted with air-tight stoppers and set in a rotary incubator shaker [New Brunswick; diameter of rotation, 1.9 cm (0.75 in)] at 70 rpm and 350C. After 24 hr, all medium was removed and replaced with 3.5 ml of medium containing horse serum (GIBCO) in place of fetal bovine serum. Two-thirds of the medium was replaced every 2-3 days until the aggregates were collected. Fluorescence Histochemistry. After 7 days in culture, each set of aggregates was incubated with medium containing 1 AuM dopamine hydrochloride (Sigma) and 1 mm ascorbate for 10 min at 350C. The aggregates were then washed twice in 1 ml of Tyrode's solution and processed by the Falck-Hillarp histofluorescence method (1). Some aggregates were preincubated for 20 min with 10 AM benztropine mesylate or 10,uM desmethylimipramine hydrochloride. Paraffin sections (10 pm) were mounted with paraffin oil and viewed with a Leitz Orthoplan microscope equipped with a 200-W mercury lamp, two Schott BG 12 filters, and a 510-nm barrier filter. Sixteen-micrometer sections of adult C57BL/6J mouse brains were prepared by using the sucrose/phosphate/glyoxylic acid method of.de la Torre and Surgeon (16) and viewed as described above. The substantia nigra pars compacta and frontal cortex sections were obtained from a mouse pretreated with pargyline hydrochloride (Sigma) (100 mg/kg) 2 hr before sacrifice. This material was not exposed to exogenous dopamine. Fluorescent sections were photographed on Kodak SO-115 film through a x25 objective. High-Performance Liquid Chromatography. Monoamine levels were measured in RMT aggregates by using high-performance liquid chromatography with electrochemical detection (17). RESULTS Histochemical Visualization of Dopamine Neurons in RMT Aggregates. Under our culture conditions, the dopamine neurons were not visualized by the Falck-Hillarp histofluorescence method unless the aggregates were first exposed to 1,M exogenous dopamine for 10 min. Certain neuronal cell bodies and their processes took up and retained the exogenous dopamine, exhibiting very bright green fluorescence (Fig. 1A). These cells were ovoid and multipolar, measured 11 ± 2,m (mean + SEM; n = 115) in longest diameter, and resembled the fluorescence histochemical descriptions of the dopamine-containing neurons in the rat mesencephalon (3, 18). For comparison, adult mouse dopamine neurons in the substantia nigra pars compacta are shown at the same magnification in Fig. 1B. These adult cells measured Am (mean ± SEM; n = 19) in longest diameter. Proc. NatL Acad. Sci. USA 78 (1981) 1265 To establish that the fluorescence observed in the aggregates was due to a monoamine specific uptake, different brain regions were dissociated and aggregated under the conditions described above and exposed to either 1,uM dopamine or 1,uM norepinephrine for 10 min. After this treatment, fluorescent neurons were observed only in aggregates formed from mesencephalic areas that contained monoaminergic cell groups; i. e., the rostral or caudal mesencephalic tegmenta. Aggregates formed from areas that do not contain monoamine cell bodies-i.e., cortex, tectum, rhombic lip, medulla, or CS-showed no specific fluorescence after this treatment. Determination of endogenous monoamine levels in RMT aggregates showed that dopamine levels in these aggregates were 10 times greater than norepinephrine levels (4.5 ± 1. 1 pg per aggregate vs. 0.5 ± 0.1 pg per aggregate; n = 5), suggesting that the dissection eliminated most of the noradrenergic neurons of the locus coeruleus. Finally, no fluorescent neurons were observed in RMT aggregates that had been pretreated with 10,uM benztropine. This agent antagonizes dopamine uptake into dopamine cells but not into serotonin cells (19). However, fluorescence was not FIG. 1. (A) Section (10 Am) through a 7-day RMT aggregate that was exposed to 1,uM dopamine for 10 min and processed by the Falck-Hillarp histofluorescence technique. Note clusters of fluorescent dopamine-containing cells and the profusion of thick, varicose dendritic processes localized to the regionsimmediately adjacent to the fluorescent cell bodies. (B) Section (16 Am) through the substantia nigra pars compacta of an adult mouse brain (sucrose/phosphate/ glyoxylic acid technique). Note the multipolar cell bodies and the dense dendritic arborization surrounding them. Marker bars = 30 Aim.

3 1266 Neurobiology: Hemmendinger et al visibly decreased in RMT aggregates pretreated with 10 A.tM desmethylimipramine. At this concentration, desmethylimipramine blocks monoamine uptake into noradrenergic and serotonergic. cells without affecting uptake into dopamine cells (19). On the basis of these results, it is reasonable to assume that most of the fluorescent neurons observed in RMT aggregates after exposure to exogenous dopamine were dopaminergic neurons of the substantia nigra and ventral tegmental area. This method was therefore used routinely to visualize the embryonic dopamine neurons in vitro. Morphology of Neuronal Fiber Patterns Observed in the Presence of Target or Nontarget Cells. After 7 days in culture neuronal processes in the various types of aggregates were examined directly in the fluorescence microscope both by a single observer and by two independent observers in a blind study. At least six sets of each type of aggregate from separate experiments were examined in serial sections. The descriptions of the aggregates reported here are based on these observations. In 7-day RMT aggregates, a dense profusion of thick varicose Proc. Natt Acad. Sci. USA 78 (1981) processes localized to regions near the cell bodies was observed (see Fig. LA). These fibers resembled the varicose, dopaminecontaining dendrites of the substantia nigra pars compacta neurons described in the rat by Bjorklund and Lindvall (18). In 3- day RMT aggregates, however, only proximal dendrites were observed, showing that the formation of dense dendritic fields occurs essentially de novo in the aggregates. For comparison, the substantia nigra of, the adult mouse brain is shown in Fig. 1B Ḟluorescent cell bodies and proximal dendrites were observed in RMT-CS aggregates, but the profusion of dendrites observed in RMT aggregates was not seen. Dense patches of punctate fluorescence, presumably axonal and often located at a distance from the fluorescent cell bodies, were observed. This fluorescence often appeared to be pericellular, surrounding nonfluorescent ovoid areas 9,um in longest diameter (Fig. 2A). The fluorescent pattern of intact adult mouse CS is shown for comparison in Fig. 2B. Similarly, RMT-FCX aggregates lacked dense areas of dendritic arborization but contained fluorescent cell bodies, proximal dendrites, and an axonal plexus. In this case, however, the distribution of the fluorescence differed from the dense patches FIG. 2. (A) Section (10,um) through a 7-day RMT-CS aggregate that was exposed to 1,uM dopamine for 10 min and processed according to the Falck-Hillarp histofluorescence technique. A cell body is visible in the lower left side of the section and a patch of punctate fluorescence is located in the middle of the section, with nonfluorescent areas on either side. (B) Section (16,m) through the caudal portion of the CS of an adult mouse brain (sucrose/phosphate/glyoxylic acid technique), showing the fluorescent, punctate innervation pattern with nonfluorescent internal capsule fibers running through it. The caudal section is used for comparison because more cranial sections showed essentially confluent fluorescence except in the nonfluorescent internal capsule. Marker bars = 30,um. FIG. 3. (A) Section (10,um) through a 7-day RMT-FCX aggregate that was exposed to 1,uM dopamine for 10 min and processed according to the Falck-Hillarp histofluorescence technique. Note the long varicose axons that course for very long distances through the aggregate. This particular section was obtained from the periphery of the aggregate and does not contain a cell body. (B) Catecholaminergic innervation of the FCX of an adult mouse brain (16-I.m section, sucrosephosphate-glyoxylic acid technique). Note the long, beaded axons. Marker bars = 30 j±m.

4 Neurobiology: Hemmendinger et al FIG. 4. Section (10-pm) through a 7-day RMTOCX aggregate thatwas exposed to 1,uM dopamine for 10 min and processed according to the Falck-Hillarp histofluorescence technique. Three cell bodies and their proximal dendrites can be seen at the top of the section. Otherwise, no dopamine-specific fluorescent processes were observed. Marker bar = 30 gtm. of punctate fluorescence seen in RMT-CS aggregates and mimicked the pattern observed in situ in the FCX (Fig. 3A). Thin axons bearing irregularly spaced and shaped 1- to 2-,um varicosities were found throughout each aggregate, especially in the periphery. These fibers could often be traced for some distance. They did not appear to be pericellular but rather coursed through the neuropil. The fiber patterns in aggregates formed with nontarget areas varied. In RMT-OCX aggregates, cell bodies and proximal dendrites were observed and, occasionally, there was a beaded axon, but this was a rare finding. Otherwise, these aggregates were devoid of fluorescent fibers (Fig. 4). RMT-tectum ag-gregates were indistinguishable from RMT aggregates, exhibiting a profusion of varicose fibers, presumably dendritic, in the vicinity of the fluorescent cell bodies (Fig. 5). Thin beaded fi-bers were not observed. Despite the difference in fiber patterns among aggregate types, the mean number of fluorescent cells observed in all FIG. 5. Section (10,um) through a 7-day RMT-tectum aggregate that was exposed to 1,jM dopamine for 10 min and processed according to the Falck-Hillarp histofluorescence technique, showing a cluster of fluorescent cell bodies with a profusion of dendritic processes surrounding them. Note the similarity of this fluorescence pattern to that of the RMT aggregate shown in Fig. LA. Marker bar = 30 pm. Proc. NatL Acad. Sci. USA 78 (1981) 1267 types ofaggregates was relatively the same, ranging from 58 to 100, and showed no systematic variation between types. In addition, varying the proportion of RMT to corpus striatum cells from 1:2 to 1:1 did not alter the pattern of fluorescent processes observed. DISCUSSION Dopamine neurons cultured in vitro in a three-dimensional aggregate system have the ability to take up and store exogenous dopamine. To the extent of our observations, the characteristics of these uptake and storage properties are similar to those in both adult and developing systems in vivo (20, 21). Dopamine or norepinephrine is taken up and can be visualized histochemically. The uptake mechanism for dopamine is sensitive to 10,uM benztropine but apparently insensitive to 10 AuM desmethylimipramine, and the dopamine neurons can take up and store exogenous dopamine in excess of their endogenous levels. Similar results have been obtained with primary monolayer cultures of murine mesencephalic dopamine neurons (22). The aggregate system provides an opportunity to study, under a variety of defined conditions, the formation of neuronal processes essentially de novo [after dissociation, neurons are devoid of processes (23)]. The presence or absence of target cells is a major determining factor in the formation of axons by dopamine neurons. Dopamine cells cultured with cells from-the CS or FCX differentiate so as to provide axonal proliferation characteristic of each of these areas, which appears to be essentially identical to that observed in vivo (see Figs. 2 and 3). In the absence of such axonal targets-i.e., RMT aggregates or coaggregates of RMT cells with cells from the OCX or tectumfluorescent axonal processes are not observed. The conditions required for the dopamine neurons to form dendritic processes are less easily understood. Dense patches of fluorescence, clearly suggesting areas of extensive dendritic arborization, are seen only when the dopamine cells are cultured in the presence of mesencephalic cells only (see Figs. 1A and 5). Although some dendritic processes are observed in the presence of telencephalic cells from the CS, FCX, or OCX, the profusion of dendrites seen in aggregates constituted of mesencephalic cells only is not observed. The formation of dendritic fields by dopamine neurons in the mesencephalic aggregates is interesting because a variety of morphological and biochemical evidence suggests that the dopamine-containing dendrites of the substantia nigra pars compacta neurons may function as presynaptic elements (18, 24-29). In this sense, the formation of a dendritic field by these neurons may represent a situation analogous to axon formation in the presence of an appropriate neuronal target. Why telencephalic cells should prevent the establishment of an extensive dendritic field by the dopamine neurons requires further investigation. Not only is the formation of axonal processes by the dopamine cells dependent on the presence of target cells, but the patterns of distribution of these axons are dependent on the area from which the target cells are obtained; i.e., CS or FCX. This does not necessarily imply that a particular dopamine neuron can form different types of axonal fields dependent on the target cells available. The dopamine neurons in these aggregates were obtained from both the substantia nigra and the ventral tegmental area, and it has been shown in vivo that individual dopamine cells from these nuclei project to specific fields and do not provide collaterals to a variety of target areas (3, 30). It may be that, in the presence of CS or FCX, only those individual dopamine neurons destined to innervate each of these forebrain structures differentiate and develop axons in the aggregate system. The possibility that selective cell death could occur in the

5 1268 Neurobiology: Hemmendinger et al absence of the appropriate target cells and thus account for the distinctive patterns of fluorescence seems unlikely, since the mean numbers of fluorescent cells per aggregate were quite similar, regardless of the other cell types with which they were cultured. The underlying mechanisms responsible for the target area-specific patterns of axonal morphology in the RMT coaggregates with CS and FCX remain to be elucidated. The marked contrast in the density and morphological appearance of the innervation in RMT-CS and RMT-FCX coaggregates, replicating the differences observed in vivo, may be related either to target area-specified differences in the number of terminals formed by the amine neurons or to more complex interactions that result in patterns that, at least as observed histochemically, appear to be target-area specific. The similarity of the fluorescence patterns of. the aggregates to those that develop in vivo suggests that the patterns in the aggregates represent an innervation by dopamine neurons of appropriate target cells. Electron microscopic studies will, however, be necessary to characterize the ultrastructure of the fluorescent varicosities observed histochemically. Although morphologically normal patterns of fluorescence are also seen in model systems using intraocular grafts (9), the normal structure and intercellular relationships in the grafted explants are preserved. The results presented here from a dissociated cell model provide evidence to suggest that such morphologic relationships are not an absolute requirement for target area-specific process formation. The embryonic dopamine neurons and their target cells, when.removed from their normal environment in the developing brain and dissociated by using trypsin into single cells, retain the ability to interact and form a morphologically characteristic pattern of dopamine fibers. This suggests that the information. necessary for these interactions is intrinsic to the individual neurons themselves. We gratefully acknowledge the technical assistance of Eligia Dimapilis and Francis Karapas. Benztropine mesylate was a gift of Merck Sharp & Dohme, and desmethylimipramine hydrochloride was supplied by Merrell National Laboratories. This research was supported by U. S. Public Health Service Grants MH and NS L. M. H. was supported as a Predoctoral Trainee on Grant GM Falck, B., Hillarp, N.-A., Thieme, G. & Torp, A. (1962)J. Histochem. Cytochem. 10, Lindvall, 0. & Bjorklund, A. (1974) Histochemistry 39, Proc. Natl. Acad. Sci. USA 78 (1981) 3. Moore, R. Y. &. Bloom, F. E. (1978) Annu. Rev. Neurosci. 1, Moore, R. Y., Bhatnagar, R. K. & Heller, A. (1971) Brain Res. 30, Ungerstedt, U. (1971) Acta PhysioL Scand. 367, H6kfelt, T. & Ungerstedt, U. (1969) Acta PhysioL Scand. 76, Berger, B., Tassin, J. P., Blanc, G., Moyne, M. A. & Thierry, A. M. (1974) Brain Res. 81, Lindvall, 0., Bjorklund, A., Moore, R. Y. & Stenevi, V. (1974) Brain Res. 81, Olson, L., Seiger, A., Hoffer, B. & Taylor, D. (1979) Exp. Brain Res. 35, Levitt, P., Moore, R. Y. & Garber, B; B. (1976) Brain Res. 111, Garber, B. B., Powell, G. M., Hoffmann, P. C. &Heller, A. (1978) in Fourth International Catecholamine Symposium, eds. Usdin, E., Kopin, I. J. & Barchas, J. (Pergamon, New York), Vol. 2, pp Gruneberg, H. (1943)J. Hered. 34, Golden, G. S. (1973) Dev. BioL 33, Loren, I., Bjorklund, A., Falck, B. & Lindvall, 0. (1976) Histochemistry 49, Garber, B. B. & Moscona, A. A. (1972) Dev. BioL 27; de la Torre, J. C. & Surgeon, J. W. (1976) Histochemistry 49, Keller, R., Oke, A., Mefford, I. & Adams, R. N. (1976) Life Sci. 19, Bjorklund, A. & Lindvall, 0. (1975) Brain Res. 83, Berger, B. & Glowinski, J. (1978) Brain Res. 147, Iversen, L. L. (1975) in Handbook of Psychopharmacology, eds. Iversen, L. L., Iversen, S. D. & Snyder, S. H. (Plenum, New York), Vol. 3, pp Coyle, J. T. & Henry, D. (1973)J. Neurochem. 21, Prochiantz, A., DiPozzio, U., Kato, A., Berger, B. & Glowinski, J. (1979) Proc. NatL Acad. Sci. USA 76, Garber, B. B., Huttenlocher, P. R., & Larramendi, L. M. H. (1980) Brain Res., 201, Geffen, L. B., Jessell, T. M., Cuello, A. C. & Iversen, L. L. (1976) Nature (London) 260, Korf, J., Zieleman, M. & Westerink, B. H. C. (1976) Nature (London) 260, Cuello, A. C. & Iversen, L. L. (1978) in Interactions Between Putative Neurotransmitters in the Brain, eds. Garattini, S., Pujol, J. F. & Samanin, R. (Raven, New York), pp Kebabian, J. W. & Saavedra, J. M. (1976) Science 193, Phillipson, 0. T. & Horn, A. S. (1976) Nature (London) 261, Hattori, T., McGeer, P. L. & McGeer, E. G.. (1979) Brain Res. 170, van der Kooy, D. & Kuypers, H. G. J. M. (1979) Science 204,