The subcommissural organ expresses D2, D3, D4, and D5 dopamine ...

Cell Tissue Res (2004) 317: 65–77 DOI 10.1007/s00441-004-0900-z

REGULAR A RTICLE

Mercedes Tomé . Antonio J. Jiménez . Hans Richter . Karin Vio . F. Javier Bermúdez-Silva . Esteban M. Rodríguez . Jose Manuel Pérez-Fígares

The subcommissural organ expresses D2, D3, D4, and D5 dopamine receptors Received: 30 January 2004 / Accepted: 14 April 2004 / Published online: 9 June 2004 # Springer-Verlag 2004

Abstract Dopamine receptors have been found in certain populations of non-neuronal cells in the brain, viz., discrete areas of ciliated ependyma and the ependymal cells of the choroid plexus. We have studied the presence of both tyrosine-hydroxylase-immunoreactive nerve fibers and dopamine receptors in the subcommissural organ (SCO), an ependymal brain gland that is located in the roof of the third ventricle and that secretes, into the cerebrospinal fluid, glycoproteins that aggregate to form Reissner’s fiber (RF). Antibodies against D2, D3, D4, and D5 dopamine receptors were used in immunoblots of bovine striatum, fresh SCO, and organ-cultured SCO, and in immunocytochemistry of the bovine, rat, and mouse SCO. Only a few tyrosine-hydroxylase fibers appeared to reach the SCO. However, virtually all the secretory ependymal and hypendymal cells of the SCO immunoreacted with antibodies against D2, D4, and D5 receptors, with the last-mentioned rendering the strongest reaction, especially at the ventricular cell pole of the secretory ependymocytes, suggesting that dopamine might reach the SCO via the cerebrospinal fluid. The antibodies against the four subtypes of receptors revealed corresponding bands Financial support was provided by grants PI 030756 and Red CIEN, Instituto de Salud Carlos III, Spain (to J.M.P.F.), and 1030265 from Fondecyt, Chile (to E.M.R.) M. Tomé . A. J. Jiménez . J. M. Pérez-Fígares Departamento de Biología Celular, Genética y Fisiología, Facultad de Ciencias, Universidad de Málaga, Malaga, Spain H. Richter . K. Vio . E. M. Rodríguez (*) Instituto de Histología y Patología, Facultad de Medicina, Universidad Austral de Chile, Casilla de Correo, P.O. Box 567 Valdivia, Chile e-mail: [email protected] Tel.: +56-63-221207 Fax: +56-63-221604 F. J. Bermúdez-Silva Fundación Hospital Carlos Haya, Laboratorio de Investigación Hospital Civil, Malaga, Spain

in immunoblots of striatum and fresh SCO. Although the cultured SCO displayed dopamine receptors, dopamine had no apparent effect on the expression of the SCOspondin gene/protein or on the release of RF-glycoproteins (SCO-spondin included) by SCO explants, suggesting that dopamine affects the function(s) of the SCO differently from the secretion of RF-glycoproteins. Keywords Subcommissural organ . Ependymal gland . SCO-spondin . Dopamine . Dopamine receptors . Bovine . Rat (Wistar) . Mouse (C57BL/10J)

Introduction Dopamine receptors are widely expressed in neurons throughout the brain and also in certain populations of non-neuronal cells (Howard et al. 1998). The presence of dopamine D2 receptors has been demonstrated in oligodendrocytes (Howard et al. 1998) and astrocytes (Hansson and Ronnback 1988; Bal et al. 1994; Zanassi et al. 1999; Reuss and Unsicker 2000). Cells lining the ventricles have also been shown to express dopamine receptors. In the fetal primate brain, radial glia expresses D1 dopamine receptors (Wang et al. 1997), whereas during postnatal development and in mature rats, the ciliated ependymal cells lining the external wall of the lateral ventricles express D2 dopamine receptors (Howard et al. 1998). The ependymal cells of the choroid plexus express dopamine D1-like (D1 and D5) and dopamine D2-like (D2 and D4) receptors (Mignini et al. 2000); this finding has led these authors to suggest that dopamine modulates the production of the cerebrospinal fluid (CSF). Dopamine could reach the choroid plexus through dopaminergic nerve fibers (Mignini et al. 2000) and also via the CSF. Dopamine has been found to be present in the CSF of the sheep (Ruckebusch and Sutra 1984); the dopamine precursor LDOPA is the most abundant monoamine in the rat CSF (Rodríguez et al. 1999). Tyrosine-hydroxylase-immunoreactive fibers, most probably corresponding to dopamine fibers, have been observed to reach the secretory ependy-

66

mal cells of the subcommissural organ (SCO; Balaban et al. 1994; Jiménez et al. 2001), suggesting the possibility that this brain gland may be under the influence of dopamine neurons. The SCO is an ependymal gland located in the roof of the third ventricle, at the entrance of the cerebral aqueduct. This organ is an ancient and conserved phylogenetic brain structure (Oksche 1961; Rodríguez et al. 1992; Olsson 1993) that differentiates early in ontogeny (Naumann 1986; Schöbitz et al. 1986). With the exception of a few species, the SCO remains fully active during the entire life span (Rodríguez et al. 1992, 1998; Oksche et al. 1993; Meiniel et al. 1996). The SCO secretes high molecular weight N-glycosylated proteins (Nualart et al. 1991), which are released into the ventricular CSF where they condense to form a thread-like structure known as Reissner’s fiber (RF; Meiniel et al. 1996; Rodríguez et al. 1992, 1998). Since new molecules are constantly being added at its rostral end, RF is an ever-growing structure extending along the cerebral aqueduct, fourth ventricle, and central canal of the spinal cord (Rodríguez et al. 1992). A fraction of the secretory material released at the apical pole of the SCO cells remains soluble in the CSF (Rodríguez et al. 1993). A strong body of evidence indicates that SCO-spondin, the major protein component of RF, plays a role in the development of the central nervous system (CNS; Monnerie et al. 1995, 1997, 1998; Gobron et al. 1996, 2000; Meiniel 2001). Although the function of the SCO in adulthood is far from being fully known, the early view that it participates in the clearance of CSF monoamines (Hess and Sterba 1973) has gained strong support from recent investigations (Rodríguez et al. 1999; Rodríguez and Caprile 2001; Caprile et al. 2003). The elucidation of the mechanisms controlling the secretory activity of the SCO may contribute to the further unfolding of the overall function of this brain gland. Although several neural systems are apparently involved in the control of the secretory activity of the SCO (Bouchaud and Bosler 1986; Jiménez et al. 2001; Nürnberger and Schöniger 2001), the serotonergic input to the SCO originating in the raphe nuclei has been substantiated (Bouchaud and Bosler 1986; Rodríguez et al. 1992; Jiménez et al. 2001) and its role partially clarified. According to Møllgard et al. (1978), Bouchaud (1979), and Léger et al. (1983), serotonin plays an inhibitory control on the secretory activity of the SCO. Recently, H.G. Richter, M. Tomé, C.R. Yulis, K. Vio, A.J. Jiménez, J.M. Pérez-Fígares, and E.M. Rodríguez (in preparation) have shown that SCO cells express serotonin receptors, and that serotonin down-regulates the transcription of the SCO-spondin gene and decreases the amount of its translation products. Balaban et al. (1994) have proposed that the modest catecholaminergic input to the SCO might exert excitatory control in contrast to the inhibitory effect of the massive serotonergic innervation. The aim of the present investigation has been to gain information about the probable participation of dopamine in SCO regulation by searching for dopamine receptors in the secretory cells of the SCO

and by analyzing the effect of dopamine on the secretory activity of the SCO.

Materials and methods Animals Adult male and female mice of the C57BL/10J strain and adult female Wistar rats were used. The animals were maintained under controlled conditions of illumination (a light/dark ratio of 12:12) and temperature (22°C). Food was administered ad libitum. Principles of animal care and specific national (B.O.E. No. 256, October 25th, 1990) and European directives and laws (86/609/EEC, November 24th, 1986) were followed. Tissue blocks containing the bovine SCO were obtained at the slaughterhouses of Málaga (Spain) and Valdivia (Chile), with a postmortem time of approximately 10–20 min.

Immunocytochemistry The SCOs of six bovine brains were dissected out in a local slaughterhouse (GASA, Málaga) and fixed by immersion in PLP (a mixture of periodate, lysine, and paraformaldehyde) for 24 h at 4°C. Six male and female C57BL/10J mice and four female Wistar rats were deeply anesthetized with ether and fixed by vascular perfusion with PLP about 3–5 h after the beginning of the light period. Mouse and rat brains and bovine SCOs were cryoprotected by immersion in 10% (2–3 h), 20% (2–3 h), and 30% (overnight) sucrose solutions. Free-floating frozen sections (20 μm thick) were subsequently obtained. Before immunostaining, sections were washed in 0.01 M phosphate-buffered saline (PBS), pH 7.3, and sequentially treated with H2O2 to block endogenous peroxidase and an avidin–biotin blocking kit (SP-2001; Vector, Calif., USA) to block endogenous biotin and biotin-binding proteins. To study the expression of dopamine receptor subtypes, at least two different polyclonal antibodies raised in rabbit against synthetic peptides were used for each subtype (Fig. 1): (1) anti-D2 (sequence 1), against a peptide sequence of the extracellular domain of rat D2 (AB1558; Chemicon International, Calif., USA), 1:500 dilution; (2) anti-D2 (sequence 2), against a peptide sequence of the third intracellular loop of human D2 (324393; Calbiochem, Calif., USA), 1:5,000 dilution; (3) anti-D2 (sequence 3) affinity-purified antibody, against a peptide sequence of the third intracellular loop of rat D2 (donated by Dr. A. de la Calle, Department of Cell Biology, University of Málaga; Khan et al. 1998), 20 μg/ml; (4) anti-D3 (sequence 1), against a peptide sequence of the third intracellular loop of rat D3 (AB11786P; Chemicon International), 1:200 dilution; (5) anti-D3 (sequence 2) affinity-purified antibody, against a peptide sequence of the third intracellular loop of rat D3 (donated by Dr. A. de la Calle, Department of Cell Biology, University of Málaga; Khan et al. 1998), 7 μg/ml dilution; (6) anti-D4 (sequence 1), against a peptide sequence of the second extracellular loop of human D4 (324405; Calbiochem), 1:10,000 dilution; (7) anti-D4 (sequence 2) affinitypurified antibody, against a peptide sequence of the third intracellular loop of human D4 (donated by Dr. A. de la Calle, Department of Cell Biology, University of Málaga; Khan et al. 1998), 15 μg/ml; (8) anti-D5 (sequence 1), against a peptide sequence of the extracellular domain close to the amino terminal end of human D5 (324408; Calbiochem), 1:2,500 dilution; (9) anti-D5 (sequence 2) affinity-purified antibody, against a rat/human common peptide sequence of the intracellular domain close to the carboxy terminal end (donated by Dr. A. de la Calle, Department of Cell Biology, University of Málaga, Khan et al. 2000), 15 μg/ml; (10) anti-D5 (sequence 3), against a peptide sequence of the intracellular domain close to the carboxy terminal end of rat D5 (AB1791P; Chemicon International), 1:200 dilution. A monoclonal antibody against a peptide sequence of the rat D1 subtype was tested and found to react in rat brain tissue only; for this reason, it was not included in the

67 present report. Two other additional antibodies raised in rabbits were used: anti-tyrosine-hydroxylase (P40101-0; Pe-freez Biologicals, Roges, UK), 1:1,000 dilution; anti-RF-glycoproteins (AFRU: A, antibody; FR, fiber of Reissner; U, urea; Rodríguez et al. 1984). AFRU reacts with the two high-molecular-weight glycoproteins secreted by the SCO (Nualart et al. 1991), 1:1,000 dilution. Sections were sequentially incubated, under gentle agitation, in the primary antibody for 72 h, at 4°C, the biotinylated goat anti-rabbit secondary antibody (1:500; E-0432; Dako, Denmark), and extravidin conjugated with peroxidase (1:2,000; E-2886; Sigma, St Louis, Mo., USA); the two last-mentioned incubations were at room temperature (RT) for 1 h. The peroxidase reaction product was visualized with 0.05% 3.3′-diaminobenzidine tetrahydrochloride (DAB) containing 0.002% H2O2. Antibodies were diluted in 0.01 M PBS, pH 7.3, containing 1% bovine serum albumin, 5% goat normal serum, and 0.5% Triton X100 (Sigma). Immunostaining of rat and mouse brain cortex with the antibodies against the dopamine receptors was used as a positive control; omission of the primary antibody in the immunostaining procedure was used as a negative control.

Organ culture of bovine SCO Bovine SCO explants obtained according to Schöbitz et al. (2001) were used for in vitro experiments. The posterior commissure was removed leaving a tissue mostly consisting of ependymal and hypendymal secretory cells. About 10–15 explants were obtained from each SCO and cultured in a nutrient mixture F-12 mixed with

Fig. 1 The protein structure of the D2, D3, D4 and D5 dopamine receptors, indicating the location of the peptide sequences used to raise the antibodies utilized in the present investigation. D2: peptide sequence 1 was not provided by the supplier (Chemicon International); human sequence 2 was provided by Calbiochem, and the rat sequence 3 by Dr. A. de la Calle, Department of Cell Biology, University of Málaga. D3: rat sequence 1 for Chemicon antibody

Dulbecco’s modified Eagle’s medium (DMEM) at a ratio of 1:1 (DMEM-F12; D-047; Sigma), containing 1% penicillin–streptomycin (P3539; Sigma) and 5.6 mg/l amphotericin B (A9528; Sigma). The explants were cultured at 37°C, in an atmosphere of 95% O2 and 5% CO2.

Immunoblot analyses

Dopamine receptor subtypes Explants from 50 bovine SCOs were obtained. Explants obtained from groups of five SCOs were pooled. Ten pools were prepared; three of these pools were not cultured and used fresh (fSCO) to obtain a membrane preparation; the other pools were cultured for 15 days (three pools), 4 weeks (two pools), and 6 weeks (two pools). The incubation medium was changed every 3 days. At the end of the culture period, membrane protein extracts were obtained from each pool according to Khan et al. (1998), by using 0.05 M TRIS–HCl with 10% sucrose and a cocktail of proteases inhibitors (100 mM EDTA, 50 mM phenylmethyl sulfonyl fluoride, 2 mM pepstatin A, 2 mM leupeptin). The protein content was determined according to Bradford (1976). Samples (20 μg per well) were subjected to SDSpolyacrylamide gel electrophoresis (SDS-PAGE) with 10% polyacrylamide, followed by blotting to polyvinylidene difluoride membranes (PVDF; Bio-Rad Laboratories, Calif., USA). Blots were saturated with 0.01 M PBS containing 5% non-fat dry milk and

(not provided); rat sequence 2 was provided by Dr. A. de la Calle. D4: human sequence 1 for Calbiochem antibody provided by supplier and human sequence 2 by Dr. A. de la Calle. D5: human sequence 1 for Calbiochem antibody provided by supplier; rat/ human common sequence 2 provided by Dr. A. de la Calle; rat sequence 3 for Chemicon antibody (not provided by supplier)

68 0.1% Tween-20 (Sigma), for 90 min. Polyclonal antibodies against D2 [see (1) in “Immunocytochemistry” section, dilution 1:5,000], D3 [see (4), dilution 1:250], D4 [see (6), dilution 1:5,000; (7), dilution 1:500], and D5 [see (8), dilution 1:5,000; (10), dilution 1:500] dopamine receptor subtypes were used. Incubation in the primary antibody was overnight at RT. Goat anti-rabbit IgG labeled with horseradish peroxidase (HRP; A6154; Sigma) was used at a dilution of 1:5,000, for 1 h 30 min, in darkness. Immunoreactive polypeptides were detected by using an enhanced chemiluminescence (ECL) system (ECL Western Blotting Detection Reagents; RPN2109; Amersham, UK) as instructed by the manufacturer. The whole procedure was performed at RT. Membrane protein extracts from fresh bovine striatum were used as a positive control. Control blots were processed as above, but incubation in the primary antibody was omitted.

RF-glycoproteins Twelve pools (each pool containing the explants from four SCOs) were cultured in the absence (four pools) or presence (four pools) of 2.0 μM dopamine (4000; Calbiochem) or 2.0 μM serotonin (four pools; H-9523; Sigma) over 15 days. The incubation medium together with dopamine and serotonin was changed every 3 days. Ammonium bicarbonate protein extracts were prepared from each one of the 12 pools (Nualart et al. 1991), and their protein content was determined according to Bradford (1976). From each pool, 20 μl samples (corresponding to 0.1 SCO) were subjected to SDSPAGE with a 5%–15% polyacrylamide lineal gradient. Proteins were transferred to nitrocellulose membranes (Towbin et al. 1979) and, after the blockage of non-specific protein-binding sites, were sequentially incubated in AFRU (1:25,000 dilution) for 3 h, and goat anti-rabbit IgG-HRP (1:12,500) for 90 min, in darkness. Immunoreactive bands were visualized by using an ECL system (Super Signal, Pierce; Walker et al. 1995). After being developed and fixed, the autoradiographic films were digitalized (GS-800 Calibrated Densitometer, Biorad Laboratories, Calif., USA), and the relative band density was analyzed with Quantity One 4.4.0 software (Biorad Laboratories). By using the StadGraphic software, version 4.0, the data were subjected to the Kolmogorov–Smirnoff test and represented in a histogram. For each of the conditions, columns represent the mean ± SD. Data were expressed as the percentage of immunoreactive RF-glycoproteins content with respect to control values, which were regarded as 100%.

SCO-spondin gene expression To investigate the effect of dopamine on the expression of the SCOspondin gene of the bovine SCO, 12 pools of explants were prepared (each pool containing the explants from four SCOs, with each SCO rendering about 15 explants). Four of these pools were cultured for 15 days in the absence of dopamine, and the other pools were cultured for 15 days in the presence of 0.03 μM dopamine (four pools) or 2.0 μM dopamine (four pools). The incubation medium together with dopamine was changed every 3 days. At the end of the culture period, the explants were subjected to total RNA isolation as described below.

Analysis of the SCO-spondin mRNA by means of reverse transcriptase-polymerase chain reaction: materials and probes SuperScript II RNaseH-reverse transcriptase, Taq DNA polymerase, DNase I, 500-bp DNA-ladder, and Trizol reagent were all purchased from Gibco BRL (Life Technologies, Rockville, Md., USA). The following primers available in our laboratory were used: (1) bovine SCO-spondin, which amplifies a 233-bp fragment (Richter et al. 2001); (2) rat β-actin, which amplifies a 280-bp fragment (Rios et al. 2001).

Semiquantitative analysis of the SCO-spondin mRNA Total RNA was extracted by using Trizol reagent, according to the manufacturer’s instructions. Aliquots of 1 μg of each RNA sample were digested with DNase I, and cDNAs were synthesized by using random primers as described before (Torres-Farfan et al. 2003). A cDNA pool containing an aliquot of each experimental group was utilized to establish a lineal range of cycles for polymerase chain reaction (PCR) amplification of SCO-spondin and β-actin. Amplification was lineal from 26 to 35 cycles for SCO-spondin mRNA and from 22 to 30 cycles for β-actin mRNA. All PCRs were carried out in a Perkin–Elmer System 2400, under the conditions described by Richter et al. (2001). Aliquots of all PCR products were analyzed by electrophoresis on a 1.5% agarose-ethidium bromide gel. Gel images were captured by using the DigiDoc-It Imaging System (UVP, Upland, Calif., USA), and the density of each band was analyzed with the software Scion Image Beta 4.0.2 (Scion Corporation, USA, available at http://www.scioncorp.com). We measured the effect of the various dopamine treatments on SCOspondin mRNA expression in cultured bovine SCO explants (see above) by using 30 cycles, whereas the β-actin housekeeping gene was amplified during 26 cycles. Each SCO-spondin and β-actin cDNA sample was amplified in three separate assays to determine the ratio between SCO-spondin and β-actin. The data from SCOspondin gene expression were subjected to one-way analysis of variance (ANOVA) followed by the post-hoc Newman–Keuls comparison test (GraphPad Prism, version 3.02; GraphPad Software, USA), to determine the statistical significance of the relative values obtained for the SCO-spondin mRNA expression level. The data were represented as the mean ± SD of the results obtained as the ratio between SCO-spondin and β-actin mRNAs levels.

Quantification of RF-glycoproteins in conditioned medium by enzyme-linked immunosorbant assay The sandwich enzyme-linked immunosorbant assay (ELISA) method standardized by Estivill-Torrús et al. (1998) to quantify RFglycoproteins was used. The capture antibody was the purified IgG fraction of AFRU, used at a concentration of 10 μg/ml, and the bound antibody was a mixture of five monoclonal antibodies (Mabs) conjugated to horseradish peroxidase. Two of these Mabs were against bovine RF-glycoproteins (Pérez et al. 1996; Miranda et al. 2001), and three were against the intracellular secretion of the bovine SCO (Fernández-Llebrez et al. 2001). Peroxidase activity was revealed with 3-3′, 5-5′-tetramethyl benzidine (T-2885; Sigma), and absorbance at 450 nm was recorded. Bovine RF-glycoprotein solutions from 5,000 to 5 ng/ml were used to obtain the standard curve. The ELISA method was used to quantify the RF-glycoproteins released into the culture medium by bovine SCO explants cultured in the presence or absence of dopamine. Explants from ten SCOs were pooled and cultured for 1 day in DMEM-F12 supplemented with 10% fetal bovine serum (F-7524; Sigma). The explants were then grouped into 24 pools (four explants each) and further cultured for 8 h in 250 μl medium/pool in the absence (12 pools) or presence (12 pools) of 10.0 μM dopamine. This experiment was performed four times. Data were expressed as the percentage of RF-glycoprotein concentration (ng/ml) in the conditioned medium with respect to values obtained with the SCO cultured in the absence of dopamine, which were regarded as 100%. Data were represented in histograms; for each condition, the column and the bar represent the mean ± SD. One way ANOVA (StatGraphic software, version 4.0) was used to determine the statistical significance of the relative values obtained for the RF-glycoproteins released from the SCO explants.

69

Results

tyrosine-hydroxylase-immunoreactive nerve fibers establish synaptic contacts with the SCO secretory cells.

Only a few tyrosine-hydroxylase immunoreactive nerve fibers reach the SCO. Mouse Rat Periventricular hypothalamic neurons immunoreactive for anti-tyrosine-hydroxylase, located close to the walls of the third ventricle and ventral to the SCO, project numerous fibers that run dorsally along the subependymal region. Some of these fibers reach the basal region of the SCO cells (Fig. 2a–c). No evidence has been obtained that the Fig. 2a–e Immunostaining of rat and bovine sections with an antibody against tyrosine-hydroxylase (TH). a Frontal section through the rat diencephalon showing the intense immunoreaction in neurons and fibers from the periventricular hypothalamic areas (SCO subcommissural organ, PC posterior commissure, IIIV third ventricle). b Detailed magnification of the area framed in a showing the wealth of tyrosine-hydroxylaseimmunoreactive fibers (arrows) close to the ependyma (E) of the third ventricle. c Detailed magnification of the SCO region shown in a. Few immunoreactive fibers (arrows) are seen at the basal region of the SCO. d Bovine SCO (Hy hypendyma) showing scarce immunoreactive fibers at the base (arrow) and in between (arrowhead) the SCO ependymal cells (E). e Bovine posterior commissure showing numerous tyrosine-hydroxylaseimmunoreactive fibers (IIIV third ventricle). Bars 250 μm (a), 50 μm (b, c), 25 μm (d), 100 μm (e)

Although numerous tyrosine-hydroxylase-immunoreactive neurons were seen in the periventricular region of the hypothalamus, no immunoreactive fibers were seen in the SCO proper.

70

Bovine The posterior commissure displayed numerous tyrosinehydroxylase-immunoreactive fibers running parallel to the myelinated fibers (Fig. 2e). Very few of these fibers bent ventrally and penetrated the ependymal layer of the SCO (Fig. 2d).

Fig. 3a–e Frontal sections of bovine SCO immunostained with antibodies against RF-glycoproteins and D2, D4, and D5 dopamine receptors. a Immunostaining with an anti-RF-glycoproteins antibody (AFRU). The ependymal (E) and hypendymal (Hy) cells are strongly reactive. b Immunostaining with an antiD2 dopamine receptor (seq. 3 sequence 3 in Fig. 1). Hypendymal (Hy) and ependymal cells (E) are immunoreactive, especially at their free surface (arrow). c Immunostaining with an anti-D4 dopamine receptor (seq. 2 sequence 2 in Fig. 1). The ependymal (E) and hypendymal (Hy) cells of the SCO and secretory cells of the posterior commissure (PC, rectangle) are reactive, whereas the adjacent ciliated ependyma is not (open arrow). d Detail of an area similar to that framed in rectangle of c; immunostaining using an anti-D4 dopamine receptor (seq. 1 sequence 1 in Fig. 1). The SCO secretory cells located in the PC are strongly reactive (arrows). e Frontal section through the SCO immunostained with an anti-D5 dopamine receptor (seq. 2 sequence 2 in Fig. 1, Hy immunoreactive hypendymal cells). Some ependymal cells (E) are immunoreactive, especially at their free surface (arrow), whereas others are not (star). Bars 500 μm (a, c), 100 μm (b, e), 50 μm (d)

Immunodetection of dopamine receptors in tissue sections and blots of SCO D2 dopamine receptors Two of the antibodies used against D2 dopamine receptors (Fig. 1, sequences 2 and 3) immunostained the ependymal and hypendymal cells of the bovine SCO (Fig. 3b). In the ependymal cells, the reaction was strong at the ventricular cell pole (Fig. 3b). The same populations of cells were immunostained with the antiserum against RF-glycoproteins (AFRU, Fig. 3a). The cells of the mouse SCO reacted weakly with these antibodies; however, the reaction was stronger at the apical cell pole (Fig. 4a). Immunoreaction

71

was also found in other brain regions in which the presence of the D2 receptor has been reported, such as the striatum and brain cortical neurons (Fig. 4d). Immunoblot analysis with the anti-D2 dopamine receptor antibody (Fig. 1, sequence 1) revealed two bands of 44 and 66 kDa in extracts of both fresh bovine SCO and striatum (Fig. 5a). Both polypeptides continued to be expressed by the bovine SCO cells after 15 days of organ culture (Fig. 5a). D3 dopamine receptors Immunoreactivity for this receptor with antibodies against sequences 1 and 2 was not detected in brain sections containing the bovine, mouse, or rat SCO. A weak immunoreaction was detected in the rat brain cortex by using the antibody against sequence 2. However, in immunoblotting, the anti-D3 dopamine receptor antibody

Fig. 4a–f Immunocytochemical localization of D2 (seq. 2 sequence 2 in Fig. 1), D4 (seq. 1 sequence 1 in Fig. 1), and D5 (seq. 2 sequence 2 in Fig. 1) dopamine receptor proteins in frontal sections of the mouse SCO (a–c) and mouse brain cortex (d–f). a The ependymal cells of the SCO display a moderate reaction with antiD2 dopamine receptor in their cytoplasm, and a strong reaction at their free surface (black arrow). b The SCO ependymal cells display a moderate reaction with anti-D4 dopamine receptor (black arrow). c Immunostaining with anti-D5 dopamine receptor antibody revealed

(Fig. 1, sequence 1) reacted strongly with a band of 52 kDa in extracts of fresh bovine SCO and striatum (Fig. 5b), and a weak band of the same molecular weight was seen in extracts of bovine SCO maintained in organ culture for 15 days (not shown). D4 dopamine receptors Antibodies against the sequences 1 and 2 (see Fig. 1) of the D4 dopamine receptor selectively immunoreacted with the secretory cells of the bovine SCO distributed in the ependymal and hypendymal layers and throughout the posterior commissure (Fig. 3c, d), fully resembling the distribution of the cells immunostained with AFRU (Fig. 3a). The antibody against sequence 1 rendered a stronger reaction than that of the antibody against sequence 2. The rat and mouse SCO reacted with the antibody against sequence 1 of the D4 dopamine receptor

a strong reaction of the ependymal cells of the SCO in their cytoplasm and an even stronger reaction at their free apical surface (black arrow). The neighboring ciliated ependyma is not reactive with anti-D2, anti-D4 or anti-D5 antibodies (open arrows in a–c). d-f Cortical neurons immunoreacted with D2, D4 and D5 dopamine receptor antibodies. Whereas reaction for D2 and D5 was preferentially seen in cell bodies (d, f), the peripheral region and cell processes of cortical neurons are visualized with the anti-D4 antibodies (e, arrow). Bars 100 μm (a–c), 25 μm (d, e), 10 μm (f)

72

Fig. 5a–d Immunoblot analyses of membrane preparations from bovine striatum (Str), freshly collected SCO (fSCO) and SCO explants organ cultured for 15 days (15 div), with antibodies against D2, D3, D4, and D5 dopamine receptors (numbers left molecular masses in kiloDaltons). a Immunoblots with anti-D2 dopamine receptor (seq. 1 sequence 1 in Fig. 1). b Immunoblots with anti-D3 dopamine receptor (seq. 2 sequence 2 in Fig. 1). c Immunoblots with two antibodies against D4 dopamine receptor (seq. 1, seq. 2

sequences 1 and 2, respectively, in Fig. 1). Both antibodies revealed the same immunoreactive band (44 kDa) in the striatum and fresh SCO preparations; the cultured SCO did not show any band. d Immunoblots with two antibodies against D5 dopamine receptor (seq. 1, seq. 3 sequences 1 and 3, respectively, in Fig. 1). One of the antibodies (seq. 1) revealed a wide range of immunoreactive bands in the three membrane preparations, whereas the other (seq. 3) stained only two bands of 34 and 44 kDa

(Fig. 4b). This antibody also reacted in other brain regions, such as the striatum and brain cortex (Fig. 4e). The immunoblot analysis with the antibodies against sequences 1 and 2 (see Fig. 1) of the D4 dopamine receptor revealed a band of 44 kDa in extracts of both fresh bovine SCO and striatum (Fig. 5c). The antibody against sequence 1 reacted stronger with the 44-kDa band than did the antibody against sequence 2 (Fig. 5c). The extracts of bovine SCO maintained in organ culture for 15 days did not show immunoreactive bands with these antibodies (Fig. 5c).

of the secretory cells of the rat and mouse SCO, with a stronger reaction at their apical region (not shown). Brain cortical neurons of the mouse showed strong immunoreactivity with anti-D5 antibody (Fig. 4f). Immunoblotting of extracts of fresh bovine SCO and striatum with the antibody against sequence 1 (see Fig. 1) of the D5 dopamine receptor revealed several bands with different degrees of reactivity and ranging between 34 and more than 200 kDa in molecular weight. Three of these bands of 44, 52, 94 kDa and one of more than 200 kDa were strongly reactive and present in both tissue samples (Fig. 5d). These bands, although less evident, were also present in the bovine SCO organ cultured for 15 days (Fig. 5d). The antibody against sequence 3 (see Fig. 1) of the D5 dopamine receptor reacted only with two bands of 34 and 44 kDa in extracts of fresh bovine SCO and striatum (Fig. 5d). No immunoreaction was detected in sections of bovine SCO and rat and mouse brain when incubation in the primary antibody was omitted. No bands were visualized in blots of bovine SCO and striatum prepared for immunoblotting but without incubation in the primary antibody.

D5 dopamine receptors Antibodies against the sequences 1 and 2 (see Fig. 1) of the D5 dopamine receptor immunolabeled a population of secretory cells of the bovine SCO, with the reaction being stronger at the ventricular cell pole (Fig. 3e). The antibody against sequence 2 strongly stained the secretory cells of the mouse SCO (Fig. 4c); the reaction was also stronger at the apical cell pole (Fig. 4c). The antibody against sequence 1 gave a weak reaction throughout the cytoplasm

73

Expression of the SCO-spondin gene and protein in bovine SCO organ cultured in the presence of dopamine Expression of the SCO-spondin gene No significant differences were found between the levels of SCO-spondin mRNA in explants cultured for 15 days in the absence or presence of dopamine, at two different concentrations (Fig. 6a). Expression of the SCO-spondin protein Immunoblots with AFRU as the primary antibody revealed an intense band of 540 kDa corresponding to the predicted size for SCO-spondin protein. The levels of SCO-spondin protein, as established by semiquantitative immunoblotting, in explants cultured for 15 days in the presence or absence of dopamine were not significantly affected (Fig. 6b). However, the serotonin treatment used as a control (H.G. Richter, M.Tomé, C.R. Yulis, K. Vio, A. J. Jiménez, J.M. Pérez-Fígares, E.M. Rodríguez, in preparation) showed a significant reduction of SCOspondin protein expression in explants cultured for 15 days in the presence of serotonin compared with untreated explants (Fig. 6b). Release of RF-glycoproteins (SCO-spondin included) by bovine SCO explants into the culture medium Explants of bovine SCO cultured in the absence of dopamine released RF-glycoproteins into the culture medium, resulting in a mean concentration of 600 ng/ml (±2 SD; Fig. 6c). This rate of release was constant in the various experiments performed and was regarded as the 100% point of the basal capacity of the SCO explants to release their secretory proteins. When the explants of bovine SCO were further cultured for 8 h in the presence of dopamine, no significant changes in the concentration of RF-glycoproteins in the conditioned medium were detected (Fig. 6c).

Discussion Presence of dopamine receptors in the SCO revealed by immunochemical methods

Fig. 6a–c Effect of dopamine on the synthesis and release of RFglycoproteins in bovine SCO explants. a RT-PCR for SCO-spondin mRNA expression. No significant differences were seen in the SCOspondin mRNA levels between explants cultured in the absence (control, n=4) and presence of 0.03 μM (n=4) and 2.0 μM (n=4) dopamine. Vertical lines mean standard deviation. b Detection of intracellular RF-glycoproteins by immunoblots with AFRU as primary antibody (OD optical density). The representative immunoblots show the 540-kDa band corresponding to SCO-spondin protein. As predicted (H.G. Richter, M. Tomé, C.R. Yulis, K. Vio , A.J. Jiménez, J.M. Pérez-Fígares, and E.M. Rodríguez, in preparation), the semiquantitative analysis revealed a statistically significant decrease of SCO-spondin protein levels in serotonin (5HT)-treated explants (asterisk P