Distribution of dopamine D₂ receptor mRNAs in the brain and the pituitary of female rainbow trout: An in situ hybridization study

Dopamine (DA) is a neurotransmitter involved in a wide variety of cerebral and peripheric functions such as control of movement, learning, behavior and regulation of the hypothalamic-pituitary axis. Disorders in the dopaminergic pathways are related to Parkinson's disease, schizophrenia, Tourette's syndrome, or pituitary tumors. Because of its involvement in human pathologic states, the dopaminergic systems have been studied extensively. In mammals, four main dopaminergic pathways are classically distinguished: the nigrostriatal pathway, associated with movements; the mesolimbic system, thought to be involved in drug abuse; the mesocortical systems, linked with cognitive and emotional functions; and the tuberoinfundibular system, involved in neuroendocrine control of pituitary functions.

The catecholaminergic systems have also largely been studied in fish, by using either antibodies (review in Meek, 1994; Smeets and Gonzalez, 2000) or riboprobes directed against the catecholamine synthesis enzymes, notably tyrosine hydroxylase (TH; Ma, 1997; Rink and Wullimann, 2001). However, antisera toward DA itself have also been used (Meek et al., 1989; Roberts et al., 1989; Linard et al., 1996; Pierre et al., 1997). In the forebrain, dopaminergic neurons are mainly located in the olfactory bulbs, the ventral regions of the preoptic area and the hypothalamus. This last structure is well-known in fish, and other nonmammalian vertebrates, to exhibit a strong density of dopaminergic cells, in the so-called paraventricular organ (PVO) (Meek, 1994; Meek and Nieuwenhuys, 1997), whose specific projections and functional significance are still a matter of debate. Until recently, it was believed that fish were lacking a nigrostriatal dopaminergic pathway, but recent evidence indicates that the dopaminergic cells of the posterior tuberculum could, in fact, represent the homologous of the tetrapod mesencephalic dopaminergic nuclei (Kapsimali et al., 2000, 2001; Rink and Wulliman, 2001).

DA acts through seven transmembrane domain, G-protein-coupled receptors. Two main families of DA receptors are characterized on the basis of their pharmacologic and biological properties and also on their sequence identities (review in Missale et al., 1998; Vallone et al., 2000). The D₁-like family includes D₁ and D₅ receptors, which are positively coupled to adenylyl cyclase, leading to an increase in cAMP (Kebabian and Calne, 1979; Monsma et al., 1990). The D₂-like family includes D₂, D₃, and D₄ subtypes, which are either negatively coupled to adenylyl cyclase, or act through other signal transduction pathways (Missale et al., 1998; Jackson and Westlind-Danielsson, 1994). In contrast to the well studied distribution of dopaminergic neurons, data regarding the localization of DA receptors in the brain of nonmammalian species are limited to birds (Schnell et al., 1999). In fish, DA receptors have been cloned in carp (D₁ and D₂ receptors Hirano et al., 1998), eel (D₁ receptor; Cardinaud et al., 1997), and fugu (D₂ receptor; Macrae and Brenner, 1995). To date, the central distribution of DA receptors is documented in only one fish species, the eel; and this study is limited to subtypes of the D₁ receptors (Kapsimali et al., 2000). Dopamine D₂R have also been cloned in Xenopus, but there is no information on their central distribution (Martens et al., 1993).

In fishes, the roles of DA have received attention mainly in the context of the neuroendocrine control of pituitary functions. It has notably been shown that DA inhibits prolactin release (James and Wigham, 1984; Johnston and Wigham, 1990). This catecholamine, respectively, stimulates GH through D₁-like receptors (Wong et al., 1993) and inhibits α-melanocyte stimulating hormone) α-MSH; Olivereau et al., 1987) release through D₂-like receptors (Omeljaniuk et al., 1989; Lamers et al., 1997), although a stimulation of α-MSH release by means of D₁-like receptors has been demonstrated in low-pH exposed tilapia (Lamers et al., 1997). DA has also been shown in several teleosts species to strongly inhibit basal and gonadotropin-releasing hormone (GnRH) -stimulated luteinising hormone (LH) secretion. This action is particularly well documented in goldfish (Chang et al., 1984; Peter et al., 1986; Sloley et al., 1991) but also in African catfish (De Leeuw et al., 1985; Van Asselt et al., 1988), coho salmon (Van der Kraak et al., 1986), and rainbow trout (Linard et al., 1995; Saligaut et al., 1998). It is assumed that hypophysiotrophic neurons in the anteroventral preoptic area are responsible for this inhibition (Peter and Paulencu, 1980; Kah et al., 1984b; Anglade et al., 1993; Linard et al., 1996). In fish farming, cotreatments with GnRH and DA antagonists are widely used to increase the GnRH-induced surge of LH to induce final maturation and ovulation in cyprinids (Lin et al., 1985; Peter et al., 1988). In vitro experiments have also evidenced the involvement of D₂-like but not D₁-like receptors in the dopaminergic inhibition of gonadotrophin secretion directly at the pituitary level (Chang et al., 1990; Vacher et al., 2000).

Among the three D₂-like receptor subtypes, the D₂ receptor is the most expressed in the brain (Missale et al., 1998). We recently have cloned and characterized in rainbow trout two D₂R cDNA, differing by 23 amino acids all situated in the third intracytoplasmic loop (accession nos. AJ347728 and AJ347729; Vacher, 2001). Sequence identities and conserved positions of putative functional amino acids pointed out in fugu (Macrae and Brenner, 1995), carp (Hirano et al., 1998), Xenopus (Martens et al., 1993), turkey (Schnell et al., 1999), and mammals (Myeong et al., 2000) indicate that both trout sequences belong to the D₂R subtype. In the present study, we report for the first time the distribution of D₂ receptor mRNAs in the brain and pituitary of a teleost species, the rainbow trout. Because semiquantitative reverse transcriptase-polymerase chain reaction (RT-PCR) analyses indicated that the amount of D₂R mRNAs were highest in the brain of late vitellogenic females (Vacher, 2001), this study was conducted at this stage.

Abbreviations

MATERIALS AND METHODS

RESULTS

The dot-blot assay performed on in vitro-transcribed messengers demonstrated that the probe used for in situ hybridization equally recognized both rainbow trout D₂R messenger species (Fig. 1). The distribution of neurons expressing trout D₂R mRNAs has been studied on transverse and sagittal sections and will be presented from rostral to caudal. Sections were always alternatively hybridized with the sense and antisense probes to assess the specificity of the signal. The nomenclature used is that established in the rainbow trout by Meek and Nieuwenhuys (1997), with minor modifications. The data indicate that D₂R mRNAs are mainly present in the tel-, di-, and mesencephalon and poorly expressed in the caudal brain. The results are summarized on (Fig. 2).

Olfactory epithelium

The olfactory rosette consists of a row of lamellae comprising the sensory olfactory epithelium, which exhibited an intense hybridization signal, surrounding the lamina propria devoid of labeling (Fig. 3A,B). This columnar pseudostratified epithelium includes olfactory receptor neurons, basal cells, and supporting cells. However, the resolution of the autoradiographs did not allow precise identification of which of these cell types express the D₂R mRNA.

Diencephalon

Preoptic region.

Labeled cell bodies were observed in the majority of the subdivisions in the preoptic region (Fig. 2E,F). The most anterior positive cells were detected in the regio preoptica parvocellularis pars anterior, corresponding to the opening of the preoptic recess (Fig. 3G). At this level, few isolated larger cells intensely expressing the D₂R mRNAs were occasionally located close the ventral aspect of the brain (Fig. 3G). The population of positive cells in the regio preoptica parvocellularis pars anterior extended caudally up to the end of the large lateral extensions of the preoptic recess. The anterior ventral wall of the preoptic recess and the regions surrounding its lateral extensions, corresponding to the regio preoptica parvocellularis pars anteroventralis known for containing a high density of dopaminergic neurons (Linard et al., 1996), also exhibited numerous positive cells (Fig. 4A,B). The cells of the regio preoptica parvocellularis pars posterior consistently exhibited the highest density of silver grains over its entire extent (Fig. 4C–F). In this area, the labeling corresponded to densely packed small cells in subependymal position (Fig. 4D). In the caudal part of the preoptic region, scattered positive cells are detected in the nucleus suprachiasmaticus (Fig. 4C). Absolutely no hybridization signal was observed in the magnocellular nuclei.

Dorsal diencephalon and pretectal area.

In the habenula, only scattered cells in the lateral portions exhibited a slight signal, the densely packed cells of the medial habenula being devoid of silver grains. Cells of the ventral thalamus exhibited a consistent labeling (Figs. 2F, 4E,F). In the anterior pretectal area, both the nucleus pretectalis superficialis parvocellularis and nucleus pretectalis superficialis magnocellularis were negative, whereas some intensively labeled neurons were also observed in the central pretectal nucleus. In the posterior pretectal nucleus, a few large cells exhibited a strong staining.

Posterior tuberculum and hypothalamus.

The anterior, lateral, and posterior parts of the nucleus lateralis tuberis were consistently weakly stained, whereas a strong hybridization signal was detected in several periventricular regions bordering the third ventricle and its lateral and posterior recesses (Fig. 2G,H). All levels of the paraventricular organ exhibited a high density of positive cells, in its anterior part along the brain ventricle, and more caudally after the closure of the ventricle (Fig. 5A). At this latter level, corresponding to the opening of the lateral recess, scattered large cells of the periventricular posterior tuberculum were heavily labeled (Fig. 5A,B). These cells were located along the midline, but also slightly laterally (Fig. 5B), especially in the anterior portion of this nucleus. More caudally, two small dorsal evaginations of the lateral recess were surrounded by cells showing a strong hybridization signal, whereas the neighbouring periventricular regions were negative (Fig. 5C,D). At the same level, a population of small densely packed cells located ventral to the wings of the lateral recess were strongly positive. At the level of the posterior recess, again the cells lining two small dorsal evaginations contained D₂R mRNAs and two diagonal bands of cells located ventral to the recess were also strongly positive (Fig. 5E,F). Finally, a moderate positive signal was observed in the nucleus of the saccus vasculosus and around the caudal aspect of the posterior recess (Fig. 5G,H), whereas the saccus vasculosus itself was negative.

Pituitary gland

In the anterior part of the pituitary, the adenohypophysis exhibited a weak but significant hybridization signal, whereas the neurohypophysis was negative (Fig. 6A,B). In all fish, the labeling was stronger in the proximal part of the pars distalis, which contains the gonadotrophs and the somatotrophs, than in the rostral pars distalis where the prolactin follicles are located. In the proximal pars distalis, the staining was not homogeneous, but preferentially located over certain cells (Fig. 6C). Figure D shows that, in the pars intermedia, numerous, but not all cells, exhibited silver grains.

Mesencephalon

In the optic tectum, the stratum periventriculare (SPV) exhibited a moderate density of D₂R-expressing cells (Figs. 2F–H, 7A,B); however, only part of the small densely packed cells were labeled (Fig. 7C). The small cells of the ventral part of the torus longitudinalis exhibited an intense hybridization signal along its entire extent (Fig. 7D).

In the rostral mesencephalon, D₂R mRNAs were detected in the pars centralis and the pars lateralis of the torus semicircularis. In these regions, two different cell types exhibiting either a small densely stained or a large paler nucleus were observed, but the hybridization signal only corresponded to the larger cells (Figs. 2G,H, 7E). Finally, scattered cells of the migrated posterior tubercular nucleus were consistently labeled. No significant hybridization signal could be detected in the caudal brain.

DISCUSSION

In mammals, the D₂R gene codes for one premessenger, which generates two mRNA species, long and short, resulting from an alternative splicing of exon 6 (Bunzow et al., 1988). We recently have cloned in the rainbow trout two cDNAs corresponding to the full-length coding sequences of a dopamine D₂R, most likely issued from a recent gene duplication (accession nos. AJ347728 and AJ347729; Vacher, 2001). These two sequences correspond to the mammalian long isoform, because no evidence for the presence of an alternative splicing of exon 6, generating a short isoform deleted from 29 amino acids in the third intracytoplasmic loop, has been found in trout (Vacher, 2001), similar to the fugu (Macrae and Brenner, 1995) and Xenopus (Martens et al., 1993). These cDNA sequences share 96% of identity and for our in situ hybridization experiments, we used an 810-bp probe that equally hybridizes with the two corresponding messenger species as shown by the dot-blot assay. This probe contains the sequence corresponding to the third intracytoplasmic loop known for being highly divergent between DA receptors from a single species (Missale et al., 1998). At the present stage, there is no information on the sequences of other DA receptor subtypes in rainbow trout, but phylogenetic analysis and sequence identities showed that our sequence clearly belongs to the D₂R subfamily (Vacher, 2001). To date, apart from the trout, D₂R have been cloned in only two fish species, namely the fugu (Macrae and Brenner, 1995) and the carp (Hirano et al., 1998), but the present study is the first to examine the expression sites of the corresponding messengers by in situ hybridization in a fish. The results indicate a widespread distribution of the D₂R in the tel-, di-, and mesencephalon, a pattern of expression in general agreement with what has been reported in mammals (Meador-Woodruff et al., 1989; Missale et al., 1998) and birds (Schnell et al., 1999).

Similar to what has been found in mammals (Coronas et al., 1997; Koster et al., 1999), a high density of D₂R mRNAs was observed on the olfactory epithelium of the rainbow trout, which consists of three main components: olfactory receptor neurons, supporting cells, and basal cells (Yamamoto, 1982). The resolution of the technique did not allow precise identification of these D₂R mRNA-expressing cells. However, as demonstrated in rat after bulbectomy, it is likely that these messengers correspond to the olfactory receptor neurons (Koster et al., 1999), which project in the glomeruli of the olfactory bulbs where they notably establish synapses onto the dendrites of dopaminergic juxtaglomerular interneurons. In rat, it is believed that DA from juxtaglomerular neurons modulates presynaptically the transmission of olfactory receptor neurons to mitral and tufted cells (Ennis et al., 2001). As periglomerular cells in fish also express TH and DA (Meek, 1994; Smeets and Gonzalez, 2000), it is possible that similar mechanisms occur in fish.

A striking feature of the localization of these D₂R mRNAs is the high correspondence with the distribution of dopaminergic neurons or their projections, suggesting both postsynaptic and presynaptic autoreceptor localization, similar to mammals (Meador-Woodruff et al., 1989; Weiner et al., 1991; Missale et al., 1998; Smeets and Gonzalez, 2000). This overlapping was particularly obvious in the olfactory bulbs, the ventral preoptic area, the periventricular regions of the mediobasal hypothalamus, and the posterior tuberculum all known for their high density of DA-producing cells (Roberts et al., 1989; Meek et al., 1989; Meek, 1994; Linard et al., 1996). Although double labeling will be necessary to ascertain this point, one can predict that part of the dopaminergic neurons express D₂R mRNAs. This expression is notably the case in the PVO/posterior tuberculum area, nucleus recessus lateralis, and nucleus recessus posterioris. Similar to the PVO, the small dorsal evaginations of the lateral or posterior recesses, and the ventral wing of the lateral recess are surrounded by an extremely high density of small DA-positive cerebrospinal fluid-contacting neurons (Smeets and Gonzalez, 2000). This finding strongly suggests that autoreceptors are expressed by the dopaminergic neurons of the preoptic area, PVO, and posterior hypothalamus.

In rat, a coexpression of D₁ and D₂ receptors has been described in striatal neurons (Lester et al., 1993), indicating that DA can modulate physiological mechanisms and behavior by means of separate receptor-specific pathways. The comparison between expression of D₁R in eel (Kapsimali et al., 2000) and D₂R in trout (present study) suggests that, in fish, these two types of DA receptors are both present in certain regions. In eel, the distribution of four D₁R subtypes (D₁A1, D₁A2, D₁B, and D₁C) has been studied, showing that the D₁C subtype is less expressed than the other forms. Expression territories of D₁A1, D₁A2, and D₁B often overlap with that of the trout D₂R in the forebrain. Indeed, similarities are found in the ventral telencephalon, where cells flanking the ventricle expressed D₁A and D₂ receptors. In the preoptic region, both D₁R in eel and D₂R in trout are detected from the rostral to the posterior parts of the anterior preoptic parvocellular nuclei and the same pattern of labeled cells surrounding the lateral extensions of the preoptic recess was described. Expression patterns are also comparable in the hypothalamic nuclei where cells located around the hypothalamic ventricle exhibited D₁R mRNAs in eel and D₂R mRNAs in trout. This finding is also true for the periventricular layer (stratum griseum periventriculare) of the optic tectum, which in the eel exhibits D1A1 and D1B mRNAs (Kapsimali et al., 2000). The presence of several DA receptor subtypes in the optic tectum is consistent with the well-documented rich innervation of most optic tectum layers by catecholaminergic fibers (Roberts et al., 1989; Meek et al., 1989; Meek, 1994). In contrast, the high density of D₂R mRNAs in the torus longitudinalis is surprising given the modest TH innervation of this nucleus in rainbow trout (Linard, 1996).

In the telencephalic hemispheres of the rainbow trout, the D₂R mRNAs are widely distributed, notably in the area ventralis telencephali pars ventralis and dorsalis. This finding is interesting in view of the fact that this region is now considered to be the homologous of the tetrapod striatum (Reiner and Northcutt, 1992), which expresses both D₁R and D₂R (Lester et al., 1993). In the eel, the dorsal subdivision of the ventral telencephalon also strongly expresses D₁R (Kapsimali et al., 2000). In the zebrafish, tracing studies have shown that DiI or biocytin application in the pars dorsalis of the ventral telencephalon results in staining of TH-positive neurons of the posterior tuberculum (Rink and Wullimann, 2001). In addition, the orphan receptor Nurr1, which is involved in the differentiation and maintenance of the substantia nigra/ventral tegmental area in mammals (Castillo et al., 1998), is coexpressed with TH in the posterior tuberculum of developing and adult medaka (Kapsimali et al., 2001). This latter structure consistently exhibits large TH/DA-positive neurons in several teleost species (Meek, 1994), including the rainbow trout. Immunohistochemistry of TH and DA indeed results in staining of large cells adjacent and slightly caudal to the DA cerebrospinal fluid-contacting neurons of the PVO (Linard, 1996). In this context, the finding that a good proportion of the large cells of the posterior tuberculum express D₂R messengers is particularly relevant and provides further indication that the dopaminergic neurons of the posterior tuberculum are homologous to those of the substantia nigra/ventral tegmental in tetrapods. This homology could suggest that, as shown in mammals, D₂R could serve as autoreceptors, notably by reducing DA secretion (Tang et al., 1994).

DA inhibits γ-aminobutyric acid (GABA) release in the striatum of rat by means of D₂-like receptors (Girault et al., 1986). This action could be direct, because immunohistochemical studies evidenced the expression of D₂R by GABAergic neurons (Delle Donne et al., 1997). The distribution of glutamic acid decarboxylase (GAD) mRNA, the rate-limiting enzyme of GABA synthesis, recently has been published in trout (Anglade et al., 1999) and indicates a large overlap between the expression of GAD and D₂R messengers. GAD is also strongly expressed in the internal cell layer of olfactory bulbs and to a lesser extent in the mitral cell layer (Anglade et al., 1999), showing a very high similarity with D₂R localization; however, a large proportion of these cells also contain DA (Meek, 1994; Smeets and Gonzalez, 2000). A high density of GAD-positive neurons was found in the ventral subdivisons of the telencephalon. In the preoptic region, both D₂R and GAD mRNA are present in the majority of preoptic nuclei, but the highest intensity was found in the anterior and posterior pars parvocellularis. A correlation between the expression of GAD and D₂R is also obvious in the regio preoptica parvocellularis pars posterior and in the hypothalamus (Anglade et al., 1999). This finding suggests that, as in mammals, DA could modulate the activity of GABAergic neurons through D₂ receptors. In fish, GABA stimulates LH release (Sloley et al., 1992; Kah et al., 1992; Mañanos et al., 1999), and, in goldfish, there is indication that DA inhibits GABA-stimulated GTH-II release, evidencing functional DA-GABA interactions (Trudeau et al., 1993).

The overall expression of D₂R mRNAs in the pituitary was found to be similar to that described in mammals where the highest expression is also found in the pars intermedia (Meador-Woodruff et al., 1989). The presence of D₂R mRNAs, but also D1A1 and D1A2 mRNAs in eel (Kapsimali et al., 2001), in the pars intermedia is consistent with the heavy innervation of this lobe by dopaminergic fibers in teleosts (Kah et al., 1986; Kapsimali et al., 2001), notably in trout (Linard, 1996). The dopaminergic regulation of α-melanotrophin-stimulating hormone (α-MSH) is well known in fish, as in other vertebrates: DA decreases α-MSH release through D₂R in goldfish and tilapia (Olivereau et al., 1987; Omeljaniuk et al., 1989; Lamers et al., 1997) but increases α-MSH release through D₁R after acid stress in tilapia (Lamers et al., 1997). One may suggest, therefore, that D₂R mRNAs are expressed by MSH cells, although the pars intermedia also contain somatolactin cells in teleosts.

The pars distalis of teleosts also receives a dense dopaminergic innervation, and, in the goldfish, DA-positive terminals have been evidenced at the electron microscope level in close vicinity with most cell types (Kah et al., 1984a, 1986). In agreement, DA is involved in the control of several anterior pituitary cell types in teleosts, which is in accordance with our finding of a weak but significant expression of D₂R mRNAs in the pars distalis. The best-documented action of DA in the fish pituitary concerns the LH cells, whose basal and GnRH-stimulated secretion is strongly inhibited by DA in several species (goldfish, Chang and Peter, 1983; Chang et al., 1983; trout, Linard et al., 1995; Saligaut et al., 1998; Vacher et al., 2000; eel, Dufour et al., 1988; catfish, Van Asselt et al., 1990). In goldfish, catfish, and trout, pharmacologic studies have demonstrated that this action is mediated by D₂-like receptors. In vitro, on primary culture of pituitary cells, we evidenced a direct effect of bromocriptine on LH and follicle-stimulating hormone release (Vacher et al., 2000); thus, we would have expected that both gonadotroph cell types carry D₂R. Therefore, that we failed to clearly evidence D₂R mRNAs on the gonadotrophs was disappointing; however, we cannot exclude the possibility that DA acts through D₃R, because bromocriptine has similar affinity for both subtypes (Seeman and Van Tol, 1994). However, expression of D₂-like receptors in the gonadotrophs is likely because no effects of DA by means of D₂R have been reported on either the growth hormone or thyrotrophin cells of teleosts. In the trout, DA was shown to stimulate growth hormone release (Agustsson et al., 2000), but in the goldfish, this action is mediated by means of D1-like receptors (Wong et al., 1992). DA also acts on prolactin release as domperidone, a specific D₂-like antagonist, increases prolactin secretion in some teleost species (James and Wigham, 1984; Olivereau et al., 1988). In the present study, a low but consistent expression of D₂R messengers was detected in the rostral pars distalis, which contains mostly prolactin follicles and a few LH cells. Expression of D₂R by prolactin cell is classically described in mammals (Schimchowitsch et al., 1986; Meador-Woodruff et al., 1989; Mengod et al., 1989).

In trout, there is evidence that the dopaminergic turnover in the anterior brain changes during the reproductive cycle (Saligaut et al., 1992), indicating that DA may also influence the reproductive process at the brain level. In addition, we have demonstrated that the dopaminergic inhibitory tone is correlated with the estrogenic environment, suggesting some pre- or postsynaptic regulation of DA neurons by sex steroids, notably estradiol (Saligaut et al., 1998; Vacher et al., 2000). In support of this assumption, a significant expression of D₂R was observed in the main neuroendocrine regions, the preoptic area and the mediobasal hypothalamus, both known for exhibiting a high density of estrogen receptors in trout (Anglade et al., 1994; Menuet et al., 2001). We found a striking similarity between the distribution of D₂R messengers and estrogen receptors in the regio preoptica parvocellularis pars anterior and around the lateral extensions of the preoptic recess, indicating that estrogen receptors could be present in neurons expressing the D₂R. These cells may correspond to GABA neurons in the dorsal part of the parvocellular preoptic nucleus (Anglade et al., 1999) and to GABA and/or dopaminergic neurons around the lateral extensions of the preoptic recess (Linard et al., 1996; Anglade et al., 1999). Therefore, the possibility exists that estrogens modulate D₂R expression in trout. In mammals, it is known that alternative splicing of the sixth exon can generate two D₂R isoforms differing by 29 amino acids in the third intracytoplasmic loop. It is also known that estradiol does not modify the overall amount of D₂R but, by facilitating retention of exon 6, increases the ratio of the long isoform over the short one (Guivarc'h et al., 1995, 1998). In trout (Vacher, 2001) and Xenopus (Martens et al., 1993), no evidence for the presence of alternative splicing has been found, in contrast with mammals and turkey (Schnell et al., 1999); therefore, the potential estrogen regulation on D₂R expression will have to be investigated in more details.

In the present study, some isolated perikarya strongly expressed D₂R mRNAs in the ventral telencephalon/anterior preoptic area and, based on their localization and shape, these isolated cells could correspond to GnRH neurons, which have been extensively studied in trout (Bailhache et al., 1994; Navas et al., 1995). Although double labeling will be necessary, this observation could provide neuroanatomic support to the well-documented DA/GnRH interactions in teleosts. In the goldfish, apomorphine (D₁/D₂ agonist) and SKF 38398 (D₁ agonist) but not bromocriptine and LY-171555 (D₂-agonists) significantly reduced spontaneous GnRH release from preoptic-anterior hypothalamus slices in vitro (Yu and Peter, 1992). On the other hand, D2 agonists but not D1 agonists significantly reduced GnRH release from pituitary fragments (Yu and Peter, 1992). Although in situ hybridization provides information on the localization of cells expressing D₂R mRNAs, this technique does not reveal the precise final subcellular location of the functional protein, and whether D₂R can be exported to pituitary GnRH terminals remains to be proven.

In conclusion, this study is the first report on the distribution of D₂R messengers in the brain and pituitary of a teleost species. The results indicate that these receptors have a widespread expression and provide a useful morphologic basis for a better understanding on the functions and evolution of the dopaminergic systems in lower vertebrates.