Mesencephalic dopaminergic neurons in primary cultures express functional neurotensin receptors

The cellular distribution and functional aspects of neurotensin (NT) binding sites in rat mesencephalic cells in primary culture were investigated by an original approach combining anatomical and biochemical studies. Using a double-labeling protocol combining 125I-NT receptor radioautography and tyrosine hydroxylase (TH) immunocytochemistry, we obtained the first direct visualization of NT binding sites on TH-immunoreactive neurons. Eighty percent of the TH neurons were endowed with NT binding sites, which can be observed on both cell bodies and processes. TH-immunoreactive neurons were characterized as dopaminergic neurons by their ability to take up dopamine in a benztropine- and nomifensine-sensitive manner. In the mesencephalic cultures, NT increased potassium-evoked release of tritiated dopamine, and the relative potencies of various NT-related peptides to increase dopamine release were in good agreement with their abilities to bind to NT sites. These results show for the first time that cultured rat mesencephalic dopaminergic cells express functional NT receptors. Finally, the specificity and distribution of NT receptors on dopaminergic neurons in primary culture are quite similar to what was observed in the adult rat brain using pharmacological and radioautographic approaches. These data indicate that NT can influence the activity of dopaminergic neurons at very early stages of the rat brain development.

lnstitut National de la Santk et de la Fiecherche Mbdicale, Unit6 339, H6pital Saint Antoine, 75012 Paris, France The cellular distribution and functional aspects of neurotensin (NT) binding sites in rat mesencephalic cells in primary culture were investigated by an original approach combining anatomical and biochemical studies. Using a double-labeling protocol combining 12SI-NT receptor radioautography and tyrosine hydroxylase (TH) immunocytochemistry, we obtained the first direct visualization of NT binding sites on TH-immunoreactive neurons. Eighty percent of the TH neurons were endowed with NT binding sites, which can be observed on both cell bodies and processes.
TH-immunoreactive neurons were characterized as dopaminergic neurons by their ability to take up dopamine in a benztropine-and nomifensine-sensitive manner.
In the mesencephalic cultures, NT increased potassium-evoked release of tritiated dopamine, and the relative potencies of various NT-related peptides to increase dopamine release were in good agreement with their abilities to bind to NT sites. These results show for the first time that cultured rat mesencephalic dopaminergic cells express functional NT receptors. Finally, the specificity and distribution of NT receptors on dopaminergic neurons in primary culture are quite similar to what was observed in the adult rat brain using pharmacological and radioautographic approaches.
These data indicate that NT can influence the activity of dopaminergic neurons at very early stages of the rat brain development.
Mesencephalic dopaminergic (DAergic) neurons have raised a considerable interest during the past 30 years, and dysfunctions of these neurons have been demonstrated or suspected in pathological states such as Parkinson's disease (Ehringer and Hornykiewicz, 1960) or schizophrenia (Matthysse and Kety, 1975). A better knowledge of the various factors regulating DAergic transmission and the anatomical bases of these regulations might lead to new therapeutic approaches for the treatment of these diseases.
The effects of NT on DAergic neurons suggest that these neurons possess NT receptors. This hypothesis has been indirectly supported by results of lesion studies showing that in the rat mesencephalon, densities of NT binding sites were greatly diminished after destruction of DAergic neurons with 6-hydroxydopamine (Palacios and Kuhar, 198 1;Quirion et al., 1985;HervC et al., 1986), and by a recent approach, using NT receptor radioautography and TH immunocytochemistry on adjacent brain sections (Szigethy and Beaudet, 1989). This association of NT binding sites with DA neurons seems also to occur in man, since decreases in nigral and striatal NT binding site densities have been reported in Parkinson's disease (Sadoul et al., 1984;Uhl et al., 1984;Rostene et al., 1988;Chinaglia et al., 1990).
In the aforementioned studies, the modulatory action of NT on DAergic neurons was investigated on adult brain, mainly in the rat. However, the presence of NT binding sites in the rat mesencephalon was detected very early during ontogeny (Palacios et al., 1988) as measurable amounts of NT sites could be observed in this region at prenatal day 18. The functional aspect of these sites, and particularly the existence of an effect of NT on DAergic neurons at an early developmental stage, is not known. Moreover, no evidence is available as to whether NT sites detected in the mesencephalon in the perinatal period are localized on DA neurons or on other types of cells.
In order to study those different aspects, we thus decided to use primary cultures of dissociated brain cells. Indeed, numerous studies had shown that primary cultures of rat or mouse embryonic mesencephalic cells contained DAergic neurons that develop morphological and biochemical characteristics similar to those found in vivo (Berger et al., 1982;Barbin et al., 1985;Dal Toso et al., 1988;Engele et al., 1989). Furthermore, the presence of high-affinity NT binding sites was detected in primary cultures of embryonic mouse whole brain cells (Checler et al., 1986) and mouse mesencephalic cells (Chabry et al., 1990) and recently NT sites were also found in rat mesencephalic cells (Dana et al.,199 1). Since primary cultures of embryonic brain cells have been used over the last 10 years to investigate various functional aspects of neurotransmission (Bockaert et al., 1986;Thomas, 1986) we hypothesized that such cultures might pro- Embryonic rat brain dissection and cell culture conditions. Mesencephalic cells in culture were prepared from embryonic day 15 Wistar rats, as previously described (Berger et al., 1982), with several modifications. Mesencephalic tegmentum was dissected in phosphate-buffered saline (PBS: NaCl, 137 mM; Na,HPO,, 21 mM; KH,PO,, 29 mM; KCl, 1.2 mM; pH 7.3). Cells were mechanically dissociated in serum-free medium (SFM) consisting of a mixture of Dulbecco's modified essential medium and Ham-F1 2 (1: 1, v/v, both from GIBCO), supplemented with 14 mM.glucose, 15 mM NaHCO,, 5 mM HEPES, 0.05 U/ml of penicillin, and 20 &ml of streptomycin (Seromed). Cells were then collected by centrifugation (500 x g, 5 min), resuspended in SFM complemented with 10% fetal calf serum (Boehringer) at a concentration of 1 million cells/ml, and plated at a density of 0.5 million cells per well, in 24 well Costar multiwell plastic culture plates previously coated with gelatin (250 &ml, 30 min, room temperature) and polyomithine (MW = 40,000, 1.5 pg/ml, overnight at room temperature). For radioautography and immunocytochemistry, cells were grown on glass slides added in the wells just before the coating procedure. For 'H-DA release experiments, the cells were plated at a density of 1 million cells per well. The cultures were incubated at 37°C in a water-saturated 95% air, 5% CO, atmosphere. Medium was totally removed at day 5 and replaced by f;esh medium containing cytosine arabinoside (Arae; 20 PM) to limit the proliferation of glial cells (Greene and Rein, 1976). Half of the medium was then replaced every day from day 6 by medium without AraC, and cells were used after 9 d in culture.
Bindinaorocedures. Cells were incubated 60 min at 37°C with 0.3 nM Y-Tyr3-KT [IZ51-NT, 2000 Ci/mmol, prepared as previously described by Sadoul et al. (1984)] in SFM supplemented with 0.2% BSA and containing 0.3 mM phenylmethylsulfonyl fluoride and 1 mM o-phenanthroline to prevent degradation of the ligand (Checler et al., 1986). Nonspecific binding, determined in the presence of 1 KM unlabeled NT, rdutinely represented less than 15% of the total binding. At the end of the incubation, the cells were rapidly washed twice with cold SFM/BSA. Cells were scraped off in 0.1 N NaOH, and radioactivity was estimated with an LKB gamma counter. Competition experiments were performed by incubating 1251-NT (0.3 nM) in the presence of increasing amounts of unlabeled NT or related analogs. NT, Acetyl-NT,_,, (AcNT,,,), and neuromedin N were from Neosystem Laboratdries (Strasbourg, France). NT, ,, was donated bv Dr. P. Kitabai (Centre CNRS. Nice. France). Radioautography. Incubation step with lZ51-NT and washes were performed as described above. The cells were then fixed with 3.5% glutaraldehyde in 0.1 M sodium phosphate buffer (pH 7.4) for 30 min (Dana Double-labeling experiments. After the lz51-NT dinding procedure (incubation step and rinses), cells were fixed in 1% glutaraldehyde and dehydrated as for radioautography. TH immunocytochemistry was then performed as described above. After visualization of the DAB staining, glass slides bearing the cells were removed from the wells and dipped in Ilford K, emulsion, and the radioautographic protocol was used as before.
Release of tritiated dopamine OH-DA). 'H-DA release experiments were performed as described by Mount et al. (1989). Briefly, the cells were incubated for 20 min at 37°C with 50 nM 'H-DA (New Eneland Nuclear, 37 Ci/mmol), and release experiments were subsequentlfperformed at 20°C. A basal release fraction of 5 min was first recovered, followed by a 5 min high potassium fraction (20 mM KCl) in the presence or absence of NT analogs. During these two periods, o-phenanthroline (0.1 mM) was added in the medium to prevent peptide degradation (Checler et al., 1986). At the end of the experiment, residual intracellular radioactivity was extracted from the cells by adding 0.5 ml of NaOH 0.1 N in the wells, and 'H-DA release in the basal and evoked fractions was expressed as a percentage of the total intracellular 'H-DA content at the beginning of the corresponding release period.

Results
Characterization and visualization of high-ajinity NT binding sites in mesencephalic cultures As shown in Figure 1, an apparent single class of high-affinity lZSI-NT binding sites was detected in rat mesencephalic cells grown for 9 d in primary culture (Fig. 1A). The apparent dissociation constant (&) was 0.30 f 0.04 nM (n = 6), and the maximal number of sites (B,,,) was 2.84 f 0.30 fmol/well (n = 6). NT, AcNT,-,,, and neuromedin N were potent competitors of '251-NT binding, while NT,-, , and levocabastine were inactive (Fig. 1B).
Radioautograms taken from cells incubated with Y-NT showed that some individual cells were labeled ( Fig. 2A,B). Silver grains were found on both cell bodies and their proximal processes. The number of cells labeled by lZ51-NT was 687 f 30 cells/well (n = 8), which represented 0.14% of the initial number of cells. Such a cellular accumulation of grains was not observed on cells incubated with the radioligand in the presence of 1 PM unlabeled NT, where only scarce silver grains were found (nonspecific binding; Fig. 2C9.

Characterization and visualization sf DAergic neurons
In a second series of experiments, TH immunocytochemistry revealed the presence, in the mesencephalic cultures, of TH-IR cells (Fig. 3). These cells were bi-or multipolar, with various et al., 1989). On the mesencephalic cultures, this procedure led to covalent cross-linking of 50% of bound1251-NT to, or nearby, its binding perikaryal shapes, and endowed with long processes bearing site. Cells were then dehydrated in graded ethanol/water baths (70:30, varicosities ( Fig. 3A-C). The number of TH-IR cells was 407 2 x 5 min; 90: 10, 2 x 5 min; lOO:O, 2 x 10 min) at room temperature f 13 cells/well (n = 8). The neurochemical identity of these to remove unfixed liaand. The slides were dinned in Ilford K, nuclear TH-IR cells was further investigated in 3H-DA uptake experiemulsion diluted 1: 1: and exposed for 3 weeis at 4°C. The radioautoments. Double-labeling exneriments confirmed that TH-IR cells  of the control values, while 84% (84.01 * 4.28%, n = 4) of the 3H-DA uptake was still observed in the presence of the noradrenaline uptake inhibitor desipramine (5 PM). Similarly, 86% of the 3H-DA uptake remained in the presence of the 5-HT uptake inhibitor fluoxetine (1 PM, data not shown). These data demonstrated that the majority of TH-IR cells in these cultures represented DA neurons.

Combined 12'I-NT radioautography and TH immunocytochemistry
The relationship between 1251-NT-labeled cells and DAergic neurons was then investigated in double-labeling experiments combining '**I-NT radioautography and TH immunocytochemistry. Unfortunately, we observed that the high aldehyde concentration (3.5% glutaraldehyde) previously used in radioautography to cross-link 1251-NT destroyed the immunocytochemical signal for TH. It was therefore necessary to find compatible fixation conditions. Decreasing the glutaraldehyde concentration from 3.5% to 1% significantly decreased neither the amount of cross-linked lZ51-NT (50.1 + 1.3%, n = 6, and 47.5 f 1.7%, n = 6, respectively), nor the number of lZ51-NTlabeled cells (682 + 45 vs. 7 11 + 62 cells/well, n = 6). Similarly, the number of TH-IR cells was not modified when 4% paraformaldehyde, 0.08% glutaraldehyde was replaced by 1% glutaraldehyde alone (430 f 26, n = 6, and 417 + 34 cells/well, n = 6, respectively) in the immunocytochemical procedure. Thus, 1% glutaraldehyde was used in the double-labeling protocol to cross-link lZSI-NT to its binding sites.
The results of the double-labeling experiments are presented in Figure 5 and Table 1. Using this approach, 1251-NT binding sites could be shown on TH-IR cells (Fig. 5A,B). Grain accumulation was present on both cell bodies and processes. Cell counting revealed that, over a total number of 400 f 39 TH-IR cells/well, 322 + 30 TH-IR cells were labeled with lZSI-NT (Table 1). In the mesencephalic cultures, 80% ofthe TH-IR cells thus expressed lZSI-NT binding sites. The remaining 20% of TH-A Figure 2. Radioautographs of lZ51-NT total (A, B) and nonspecific (C) binding to rat mesencephalic cells in culture. Cells were counterstained with cresyl violet. Note the cell-like appearance of the silver grain accumulation in A and B, and the lack of such an image in C. Scale bar, 10 pm. IR cells appeared to be devoid of silver grain accumulation (Fig.  5C). According to our previous experiments with single labeling in which the number of lZ51-NT-labeled cells exceeded that of TH-IR cells, lz51-NT sites were also detected on TH-negative cells (Fig. 5D). These cells represented 54% of the total number of cells (701 2 67 cells/well) labeled by the radioligand (Table  1).

Effect of NT and related peptides on 3H-DA release from mesencephalic cultures
The previous anatomical studies clearly demonstrated the presence of l*SI-NT sites on TH-IR cells characterized as DAergic  neurons. In order to check whether these binding sites represented physiologically functional NT receptors, the effect of NT on DA release was tested. As shown in Figure 6, NT was able to increase potassium-evoked release of previously uptaken 3H-DA. The minimal effective concentration was lo-lo M, and the maximal effect was obtained at lo-' M NT. This maximal effect represented a 54% increase (54.0 & 5.0%, p < 0.01, n = 22) of the potassium-induced DA release. AcNT,-,, (EC,, = 0.28 nM) was as potent as NT (EC,, = 0.35 nM), while NT,_,, was inactive up to 1O-6 M (Fig. 6). When the potency of several NT fragments or analogs to stimulate DA release was compared with their ability to displace lZSI-NT binding in these cultures, a linear relationship (r = 0.989) was found between the two sets of data ( Fig. 7). No effect of NT or related peptides (lo-lo to 1O-5 M) on basal DA release could be observed under these conditions.

Discussion
We show in this article that rat embryonic DAergic neurons in primary cultures express functional NT receptors. This demonstration was achieved by the combination of (1) an original double-labeling protocol, combining both lZSI-NT receptor radioautography and TH immunocytochemistry, which provided the first direct visualization of NT binding sites on TH-IR neurons; (2) the identification of the TH-IR neurons present in the mesencephalic cultures as DAergic neurons; and (3) pharmacological experiments showing the ability of NT and its analogs to increase potassium-evoked release of 3H-DA.
We have shown previously that postfixation of 1251-NT-labeled tissue sections with high concentrations of glutaraldehyde ensured irreversible cross-linking of the radioligand to, or nearby, its binding sites (Moyse et al., 1987). We thus applied this approach for the electron microscopic localization of lZ51-NT binding sites in the rat mesencephalon (Dana et al., 1989). In the present work, we demonstrate that this procedure can be used on cells in culture to identify the cell populations expressing l*sI-NT binding sites.
The technical difficulties encountered on tissue sections by others (Szigethy and Beaudet, 1989) to combine, in the same biological sample, NT receptor radioautography and TH immunocytochemistry were overcome in cultured cells. Indeed, Figure 5. Double-labeling experiments associating both '**I-NT receptor radioautography and TH immunocytochemistry. Some TH-IR cells possessed silver grain accumulations (A, B) indicating expression of NT binding sites by these cells, while some other TH-IR cells were devoid of silver grains (C'). On the other hand, NT binding sites could be observed in some TH-negative cells (D). Scale bar, 10 pm.
TH antigenicity could be retained in spite of the various histological procedures needed for NT cross-linking onto its binding sites. The main reason for it, in our opinion, is that slicing unfixed or only lightly fixed brain, as needed to retain NT receptor binding on brain sections (Moyse et al., 1987;Dana et al., 1989;Szigethy and Beaudet, 1989), damages the tissue and NT AcNTB-13 NTl-11 may favor the loss of intracellular antigens during the binding protocol.
In the present case, the cells remained in a suitable medium until the fixation and no slicing step was required. These conditions may favor the maintenance of intracellular antigens.
In the rat mesencephalic cultures, 80% of the TH-IR cells I possessed lZ51-NT sites. These TH-IR cells may be identified as DAergic neurons since they were able to take up 3H-DA in a benztropine-and nomifensine-sensitive manner (Prochiantz et al., 1979). The functional aspect of NT sites was attested by the ability of several NT-related peptides to increase the release of 3H-DA in this system and by the relationship found between the relative activities of these peptides and their affinities for the NT binding sites. Such an action of NT on the regulation of DA release has been observed in the adult brain (see Ervin andNemeroff, 1989, andFaggin andCubeddu, 1990, for references). Moreover, it should be noticed that the binding characteristics of these NT sites in cultured mesencephalic cells are similar to those previously obtained for the high-affinity NT sites in the adult rat brain (Granier et al., 1982;Sadoul et al., 1984;Kitabgi et al., 1987;Moyse et al., 1987). Taken together, these results clearly show that, even at this early developmental stage, NT binding sites detected on mesencephalic DAergic neurons are already under a mature form, and that binding of the ligand is coupled to a biological effect. These sites may therefore be considered as functional NT receptors. The presence of NT receptors on both cell bodies and processes suggests that the peptide may act at the level of these various elements of the DAergic neurons. We cannot, however, determine whether the NT-induced increase in DA release observed under our experimental conditions reflects an action of NT on dendrites, axon terminals, or both. It also remains possible that some of the binding sites visualized in this study represent sites in transport (Dana et al., 1989).
Interestingly, both the high percentage of DAergic neurons expressing NT receptors and the cellular distribution of these receptors on cell bodies and processes, found in this work, are quite similar to results obtained in previous radioautographic studies performed on adult rat brain (Palacios and Kuhar, 198 1;Quit-ion et al., 1985;Her& et al., 1986;Szigethy and Beaudet, 1989). The presence of NT sites on both cell bodies and terminals of the DAergic neurons was indirectly suggested by the concomitant decreases of NT binding site densities observed after destruction of DAergic neurons by 6-hydroxydopamine, in mesencephalic regions such as SN or VTA and in DA projection areas such as striatum or nucleus accumbens (Palacios and Kuhar, 198 1;Quirion et al., 1985;Her& et al., 1986). Dendritic and perikaryal localizations of NT sites were also observed in SN and VTA by light microscopy (Moyse et al., 1987;Szigethy and Beaudet, 1989) and in VTA by means of electron microscopic determination (Dana et al., 1989). The good correspondence between the results of the present study and those obtained in the adult rat brain suggests that mesencephalic DA neurons in culture do not substantially differ from DA cells "in situ" in their ability to express NT receptors. Although obtained on primary cultures of embryonic cells, our data might thus also be considered to provide an anatomical support for the various effects of NT, observed in adult brain, on the different elements of DA neurons, that is, excitation of SN and VTA neurons (Pinnock, 1985;Seutin et al., 1989), alteration of dendritic nigral DA release (Myers and Lee, 1983) and of DA release in striatum or nucleus accumbens (De Quidt and Emson, 1983;H&tier et al., 1988;Blaha et al., 1990), as well as behavioral responses following intra-VTA injections (Kalivas et al., 1983;Cador et al., 1985).
Finally, a small percentage (20%) of TH-IR neurons were unlabeled by '251-NT. As there is no way to distinguish between mesolimbic, mesocortical, and nigrostriatal DA neurons in these cultures, we cannot conclude that DA neurons devoid of NT sites belong to one or another DAergic population. This percentage of unlabeled TH-IR neurons appears, however, to be quite similar to that of DA neurons exhibiting NT immunoreactivity in the adult rat brain (Hokfelt et al., 1984). It would thus be of interest to know whether these unlabeled TH-IR neurons represent colocalized DA-NT neurons. Similarly, further studies will be necessary to characterize the phenotype of the TH-negative cells that were labeled with lZ51-NT in the mesencephalic cultures, and to elucidate the functional significance of this observation.
In conclusion, the present study clearly shows that mesencephalic DAergic neurons at early stages of development express functional NT receptors distributed both on cell bodies and processes of the neurons. These data thus enlighten the anatomical bases governing the regulation of the DAergic transmission by NT at the level of the DAergic neuron itself. Moreover, the original double-labeling approach developed in this study is applicable to other brain regions, other peptide receptors, and various types of cells, allowing the precise determination of the cells expressing a given receptor. Due to the complexity of the brain circuitry, this kind of approach could improve our understanding of the interactions between neuromediators and the physiological meaning of these interactions in brain functions or pathology.