Expression of the neurokinin 1 receptor in the rabbit retina

PII: S 0 3 0 6 - 4 5 2 2 ( 0 2 ) 0 0 4 0 8 - 6 Neuroscience Vol. 115, No. 4, pp. 1309^1321, 2002 H 2002 IBRO. Published by Elsevier Science Ltd All rights reserved. Printed in Great Britain 0306-4522 / 02 $22.00+0.00 www.neuroscience-ibro.com EXPRESSION OF THE NEUROKININ 1 RECEPTOR IN THE RABBIT RETINA G. CASINI,a A. SABATINI,a E. CATALANI,a D. WILLEMS,a L. BOSCOa and N. C. BRECHAb;c a b Dipartimento di Scienze Ambientali, Universita' della Tuscia, 01100 Viterbo, Italy Departments of Neurobiology and Medicine, Jules Stein Eye Institute, UCLA School of Medicine, Los Angeles, CA 90095-1763, USA c VAGLAHS, Los Angeles, CA 90073, USA Abstract8Substance P is the preferred ligand for the neurokinin 1 (NK1) receptor. In vertebrate retinas, substance P is expressed by amacrine, interplexiform and ganglion cells. Substance P in£uences the activity of amacrine and ganglion cells and it is reported to evoke dopamine release. We investigated NK1 receptor expression in the rabbit retina using a⁄nity-puri¢ed NK1 receptor antibodies. NK1 receptors were expressed by two distinct populations of retinal neurons. One is a population of ON-type bipolar cells characterized by axonal arborizations that rami¢ed in the inner plexiform layer near the ganglion cell layer. Double-label studies showed that NK1 receptor-expressing bipolar cells were distinct from rod bipolar cells and from other immunocytochemically identi¢ed types of cone bipolar cells. Their density was about 2250 cells/mm2 in the visual streak and 1115 cells/mm2 in ventral mid-periphery. They were distributed in a nonrandom pattern. In the outer plexiform layer, the dendrites of these bipolar cells converged into heavily immunostained clusters having a punctate appearance. The density of these clusters in mid-peripheral ventral regions (about 13 000 clusters/mm2 ) was similar to the reported cone density [Famiglietti and Sharpe (1995) Vis. Neurosci. 12, 1151^1175], suggesting these dendrites contact all cone photoreceptors. The second NK1 receptor expressing cell population corresponds to the tyrosine hydroxylase-containing amacrine cell population. NK1 receptor immunostaining was localized to the cell body and processes, but not to the processes that form the ‘rings’ that are known to encircle somata of AII amacrine cells. These ¢ndings show that NK1 receptor immunoreactivity is localized to a population of ON-type cone bipolar cells and to dopaminergic amacrine cells, suggesting that substance P acting on NK1 receptors in£uences multiple retinal circuits in the rabbit retina. H 2002 IBRO. Published by Elsevier Science Ltd. All rights reserved. Key words: substance P, dopamine, bipolar cells, amacrine cells, neuropeptide receptors, photoreceptors. borize in laminae 1, 3 and 5 of the inner plexiform layer (IPL). In addition, tachykinin immunoreactivity is also reported in ganglion cells of rat, rabbit, hamster and primate retinas (Brecha et al., 1987; Caruso et al., 1990; Takatsuji et al., 1991; Kolb et al., 1995; Cuenca and Kolb, 1998; Li et al., 1999). Functionally, SP has been reported to exert excitatory e¡ects on ganglion cells in the mud puppy, carp and rabbit retina, without a¡ecting ganglion cell receptive ¢eld properties (Dick and Miller, 1981; Glickman et al., 1982; Zalutsky and Miller, 1990a). In addition, SP has been found to be excitatory to presumptive GABA-expressing amacrine cells in the rabbit retina (Zalutsky and Miller, 1990a) and to evoke [3 H]dopamine release from the rat retina (Tsang, 1986). Finally, SP reduces calcium currents in isolated bipolar cells of the gold¢sh retina, indicating that SP may modulate the release of transmitter from these cells (Ayoub and Matthews, 1992). The cellular actions of the tachykinin peptides are mediated by speci¢c, high-a⁄nity receptors. The tachykinin peptides SP, NKA and NKB are the preferred ligands for the neurokinin (NK) receptors, termed NK1, NK2 and NK3, respectively. These receptors are G protein-coupled receptors (Otsuka and Yoshioka, 1993). NK1 and NK3 receptors have been localized in rat retina The tachykinin peptides, substance P (SP), neurokinin A (NKA), and neurokinin B (NKB), are structurally related and derived from two tachykinin genes. These peptides are widely distributed throughout the peripheral and central nervous systems (see Maggio, 1988; Otsuka and Yoshioka, 1993, for reviews) where they may act as neurotransmitters, modulators and growth factors (Jonsson and Hallman, 1982a,b; Pernow, 1983; Bartho and Holzer, 1985; Nilsson et al., 1985). The presence of tachykinin immunoreactivity and/or tachykinin mRNAs has been reported in the retinas of di¡erent mammalian species (Casini et al., 1997a). Tachykinin-expressing cells are mainly wide-¢eld amacrine and displaced amacrine cells. They are sparsely distributed in all retinal regions and their processes ar- *Corresponding author. Tel: +39-761-357040; fax: +39-761-357179. E-mail address: [email protected] (G. Casini). Abbreviations : CaBP, calbindin; Cy3, carboxymethylindocyanine; DAB, 3,3P-diaminobenzidine tetrahydrochloride ; GCL, ganglion cell layer; INL, inner nuclear layer; IPL, inner plexiform layer ; NK1, neurokinin 1; NK2, neurokinin 2; NK3, neurokinin 3; NKA, neurokinin A; NKB, neurokinin B; OPL, outer plexiform layer; PB, phosphate bu¡er; PKC, protein kinase C; SP, substance P; TH, tyrosine hydroxylase. 1309 1310 G. Casini et al. using immunohistochemical (Casini et al., 1997b, 2000a; Oyamada et al., 1999; Chen et al., 2000) and in situ hybridization (Kondoh et al., 1996; Shughrue et al., 1996) techniques. NK1 receptors are mainly distributed to amacrine, displaced amacrine, and interplexiform cells. In addition, some ganglion cells may also express these receptors (Casini et al., 1997b, 2000a; Oyamada et al., 1999). Double-label immuno£uorescence studies have demonstrated the expression of NK1 receptors by GABA- and tyrosine hydroxylase (TH)-immunoreactive amacrine cells of the rat retina (Casini et al., 1997b). These ¢ndings are consistent with previous observations showing SP actions on GABA- and dopamine-containing amacrine cells (Tsang, 1986; Zalutsky and Miller, 1990a). NK3 receptors have been detected in OFF-type cone bipolar cells and in large-sized, TH-immunoreactive amacrine cells (Casini et al., 2000a; Chen et al., 2000). The cellular expression of NK receptors has only been de¢ned in the rat retina. In the present study, we evaluated the localization of NK1 receptors in the rabbit retina to better de¢ne the pattern of NK1 receptor expression in the mammalian retina. Preliminary results have been presented in abstract form (Casini et al., 2000b). EXPERIMENTAL PROCEDURES Animals and tissue preparation All procedures for the care and handling of animals were approved both by the Animal Research Committees of VAGLAHS and UCLA, and by the Italian National Committee (Law on Animal Care # 116/1992). All e¡orts were made to minimize animal su¡ering and to reduce the number of animals used. New Zealand albino rabbits were obtained from commercial sources and killed with an overdose of chloral hydrate (i.p.) or sodium pentobarbital (i.v.). The eyes were removed, the anterior segments were cut away, and the eyecups, containing the retinas, were immersion ¢xed in 4% paraformaldehyde for 2 h. Retinas were then dissected and transferred to 25% sucrose in 0.1 M phosphate bu¡er (PB, pH 7.4) for storage until cryosectioning. Sections (10 Wm) were cut in a plane either perpendicular (coronal sections) or parallel (horizontal sections) to the vitreous surface, mounted onto gelatin-coated slides and stored at 320‡C. Antibodies Two di¡erent polyclonal antibodies (Vigna et al., 1994) directed to the same sequence in the C-terminus of the rat NK1 receptor were purchased from Chemicon (Temecula, CA, USA). The ¢rst is an antiserum raised in rabbit (cat. No. 5060) and it was used at 1/100 dilution; the second is an antiserum raised in guinea-pig (cat. No. 5800) and it was used at 1/1000 dilution. Double-labeling immuno£uorescence experiments were performed using the NK1 receptor antisera in conjunction with the following antibodies: (a) a mouse monoclonal antibody (MAb115A10, from Dr. S.C. Fujita, of the Mitsubishi Kasei Institute of Life Sciences, Tokyo, Japan) which stains bipolar cells (Onoda and Fujita, 1987; Greferath et al., 1990) used at 1/5000; (b) a mouse monoclonal antibody (clone MC5, Sigma, St. Louis, MO, USA) against protein kinase C (PKC), which stains rod bipolar cells (Greferath et al., 1990; Young and Vaney, 1991; Casini et al., 1996), used at 1/1000; (c) a polyclonal antiserum made in rabbit (from Dr. K.-W. Koch, of the Institut fu«r Biologische Informationsverarbeitung, Ju«lich, Ger- many) against recoverin (Lambrecht and Koch, 1992), which stains two populations of cone bipolar cells (Euler and Wa«ssle, 1995; Massey and Mills, 1996), used at 1/1000; (d) a mouse monoclonal antibody (Sigma) directed to the 28-kDa subunit of calbindin (CaBP), which stains one population of cone bipolar cells (Massey and Mills, 1996), used at 1/1000; (e) a mouse monoclonal antibody (Becton Dickinson, San Jose, CA, USA) directed to the carbohydrate epitope CD15, which stains a population of ON-type cone bipolar cells (Brown and Masland, 1999), used at 1/10; (f) a mouse monoclonal antibody (clone 2/40/15; Boehringer-Mannheim, Indianapolis, IN, USA) directed to TH (Rohrer et al., 1986), which stains dopaminecontaining amacrine cells in the rabbit retina (Casini and Brecha, 1992a,b), used at 1/200. Immunohistochemical procedures Coronal and horizontal cryostat sections were washed in 0.1 M PB and incubated overnight in NK1 receptor antiserum diluted in 0.1 M PB containing 0.5^1% Triton X-100 at 4‡C. Sections were then washed in 0.1 M PB and incubated in the presence of secondary antibodies conjugated with Alexa Fluor £uorescent dyes (Alexa Fluor 546 and Alexa Fluor 488, Molecular Probes, Eugene, OR, USA) or with carboxymethylindocyanine (Cy3, Jackson Immuno Research, West Grove, PA, USA) at a dilution of 1/100 in 0.1 M PB containing 0.5% Triton X-100 for 2 h at room temperature. Subsequently, the sections were washed in 0.1 M PB and coverslipped in a 0.1 M PB^glycerin mixture. Horizontal sections were incubated with the guinea-pig antibodies directed to the NK1 receptor, then they were washed and incubated in biotinylated goat anti-guinea-pig IgG (Vector Laboratories, Belmont, CA, USA) at a dilution of 1/50 in 0.1 M PB containing 0.5% Triton X-100 for 2 h at room temperature. Following washing in 0.1 M PB, the sections were incubated in an avidin^biotin^peroxidase solution (Vectastain ABC Kit, Vector, Burlingame, CA, USA) for a subsequent 2 h. The sections were washed, preincubated in 0.05% 3,3P-diaminobenzidine tetrahydrochloride (DAB, Sigma) in 0.1 M Tris bu¡er for 10 min and incubated in the same solution containing 0.05% hydrogen peroxide for an additional 10 min. Finally, they were treated with 0.05% osmium tetroxide in H2 O for 30^40 s to intensify staining, washed, dehydrated and coverslipped with Permount. Speci¢city of immunostaining was evaluated by preadsorbing the primary antibodies with 10 WM of the NK1 receptor peptide fragment NK1393 407 overnight at 4‡C. Immunostaining was not apparent in sections incubated with the preadsorbed guinea-pig antibodies. Non-speci¢c immunostaining was observed in photoreceptor outer segments and, occasionally, in microglial cells located in the IPL and in the ganglion cell layer (GCL) with the preadsorbed rabbit antibodies. Both coronal and horizontal retinal sections were processed for double-label immunohistochemical staining by indirect immuno£uorescence technique (Casini et al., 1997b). Sections were washed in 0.1 M PB and then incubated overnight in a mixture containing one of the NK1 receptor antibodies and one of the other antibodies diluted in 0.1 M PB containing 0.5% Triton X-100 at 4‡C. In the case of NK1^recoverin double-labeling experiments, the NK1 receptor antiserum raised in guinea-pig was used. Sections were then washed in 0.1 M PB and incubated in the presence of the appropriate secondary antibodies conjugated with Alexa Fluor 546, with Alexa Fluor 488 or with Cy3 at a dilution of 1/100 in 0.1 M PB containing 0.5% Triton X-100 for 2 h at room temperature. Subsequently, the sections were washed in 0.1 M PB and the slides were coverslipped in a 0.1 M PB^glycerin mixture. Control experiments were performed to ascertain that secondary antibodies speci¢cally recognized the appropriate antigen^ antibody complex (Goehler et al., 1988; Casini et al., 1997b). Attempts were also made to immunostain whole retinas. However, satisfactory immunostaining could not be achieved, presumably due to the lack of penetration of the NK1 receptor antibodies. The whole-mount retinal preparations were understained compared to horizontal sections through the retina, even NK1 receptors in rabbit retina after incubation periods in the primary antibody lasting up to 4 weeks in the presence of 3% Triton X-100. Quantitative analysis The quantitative evaluation of the NK1 receptor immunostaining was made on horizontal sections treated with DAB. The retinal areas chosen for analysis ranged in size from 10 000 to 30 000 Wm2 . Measurements were obtained with a system of computer-assisted image analysis using Image-Pro Plus (Media Cybernetics, Silver Spring, MD, USA). Cell density was expressed as the number of NK1 receptor-immunoreactive cells per mm2 of retinal surface. Nearest-neighbor analysis (Wa«ssle and Riemann, 1978) was performed in each retina in the same ¢elds used for calculating NK1 receptor-immunoreactive cell density. The x^y coordinates relative to each immunoreactive cell in the ¢eld were obtained using Image-Pro Plus and, from these coordinates, nearest-neighbor distances were calculated using an automated procedure. The distribution of nearestneighbor distances was then compared with the normal Gaussian distribution expected for the same values of the mean and standard deviation of the nearest-neighbor distances if the NK1 receptor-immunoreactive cells were non-randomly distributed (Wa«ssle and Riemann, 1978). The ratio of the mean/standard deviation of the nearest-neighbor distances was calculated to obtain a measure of regularity in the pattern of distribution of NK1 receptor-immunoreactive somata (Eberhardt, 1967). Figure preparation Immuno£uorescent material was observed with a conventional Zeiss £uorescence microscope and with a Leitz laser confocal scanning microscope. Specimens treated with DAB were observed with a Zeiss Aristoplan microscope equipped with Nomarski optics. Images from the £uorescence microscope were photographed using Kodak T-Max 400 (negative ¢lm) or Provia 400 (color slides). Bright-¢eld images were photographed using Kodak Elite Chrome 100 ¢lm. The photographic negatives and slides were scanned at 2400 dpi with a SprintScan 35/LE scanner (Polaroid, Cambridge, MA, USA) and saved as TIF ¢les. Both the images scanned from photographic ¢lms and those acquired through the confocal microscope were adjusted to the ¢nal size, corrected for contrast and brightness and labeled using Adobe Photoshop 5.0 (Adobe Systems, Mountain View, CA, USA). Final images were saved at a minimum of 300 dpi. 1311 IPL (IPL laminae are numbered according to Cajal, 1893). No speci¢c immunostaining was detected either in the outer nuclear layer or in the GCL. These observations indicate that NK1 receptors are expressed by bipolar cells and large-sized amacrine cells. NK1 receptor-immunoreactive bipolar cells NK1 receptor-immunoreactive bipolar cells were observed in all retinal regions. Their dendritic arborizations in the OPL were characterized by intensely immuno£uorescent puncta, while their axonal processes passed through the INL and terminated in the IPL near the GCL, where they formed a narrow band of varicose arborizations (Fig. 2). In horizontal retinal sections treated with DAB, the arborizations of NK1 receptorimmunoreactive bipolar cells were observed in greater detail both in the OPL and in the IPL (Fig. 3). It was not possible to de¢ne the dendritic coverage of individual NK1 receptor-immunoreactive bipolar cells, because the ¢ne immunoreactive processes converged to join together and formed prominently immunostained clusters, which had a punctate appearance. These processes formed a continuous network throughout the OPL (Fig. 3A^C). In some instances, processes could be observed to run within the OPL contacting several darkly immunostained clusters (inset of Fig. 3B). The organization of the NK1 receptor-immunoreactive network in the OPL suggests that dendrites of NK1 receptor-expressing bipolar cells converge to form clusters at the point where they contact photoreceptor terminals. Dendrites converging into a single cluster may originate from the same soma or from di¡erent somata. In addition, single dendrites appear to contact more than one cluster. In confocal images from horizontal sections, immunoreactive somata displayed three or four primary dendrites that entered the OPL and branched into second-order collaterals (Fig. 4A). Higher-order dendrites were only rarely observed. As RESULTS NK1 receptor immunostaining patterns The same immunostaining pattern was obtained using both NK1 receptor antisera. Speci¢c NK1 receptor immunoreactivity was localized to cell bodies in the inner nuclear layer (INL) and to processes in the outer plexiform layer (OPL) and the IPL in all retinal regions. The immunostaining outlined the cell bodies and their thickest processes, suggesting NK1 receptor immunoreactivity is localized at or near the plasma membrane. Small NK1 receptor-immunoreactive somata were located in the distal INL near the OPL, and large NK1 receptor-immunoreactive amacrine cells were in the proximal INL adjacent to the IPL (Fig. 1). The cell bodies in the distal INL originated ¢ne processes that entered the OPL and a single smooth process directed through the INL into the IPL. The immunostained cell bodies in the proximal INL gave rise to thick processes that were distributed to lamina 1 of the Fig. 1. Photomicrograph showing the NK1 receptor immunoreactivity in a coronal section of the rabbit retina. Immunostaining was localized to the plasma membrane of bipolar cells and large amacrine cells. The bipolar cell somata in the distal INL give rise to dendrites in the OPL (arrow) and a single process directed through the INL and into the IPL (arrowheads). The amacrine cell somata in the proximal INL give rise to thick processes (double arrows) in the IPL adjacent to the INL. Scale bar = 10 Wm. 1312 G. Casini et al. Fig. 2. Photomicrographs showing NK1 receptor-immunoreactive bipolar cells in coronal retinal sections. These cells were more densely distributed in central retina (A) than in the periphery. Immunostaining in the distal IPL is due to NK1 receptor-immunoreactive amacrine cell processes. In the OPL, NK1 receptor immunostaining was characterized by highly immuno£uorescent puncta (short arrows in B), while process arborizations in the IPL were characterized by varicosities (long arrows in B). Scale bar = 15 Wm. shown in Fig. 4B, some dendrites appeared to interconnect two NK1 receptor-immunoreactive somata, while other dendrites originating from the same soma converged into the same cluster of processes in the OPL. The axonal arborizations of NK1 receptor-immunoreactive bipolar cells formed a narrow band at a level corresponding to lamina 4/5 of the IPL. Overall, these terminal arborizations made up an intricate network of varicose processes where it was di⁄cult to discriminate individual terminal arbors (Fig. 3D, E). The confocal image of Fig. 4C shows the terminal arborizations of immunostained bipolar cell axons: in this image, two di¡erent terminal arbors could be observed; however, in adjacent optical sections the distinction between the two arbors disappeared. Characterization of NK1 receptor-immunoreactive bipolar cells To con¢rm that these NK1 receptor expressing cells were bipolar cells, double-label immuno£uorescence was performed using antibodies directed to the NK1 receptor in conjunction with a monoclonal antibody (MAb115A10) that recognizes an antigen expressed by all bipolar cells (Onoda and Fujita, 1987; Greferath et al., 1990). All NK1 receptor-immunoreactive cells with somata in the distal INL also displayed MAb115A10 immunoreactivity (Fig. 5), indicating they are bipolar cells. To characterize the NK1 receptor-expressing bipolar cells and to determine whether they constitute a unique bipolar cell type, a series of double-label immuno£uorescence experiments (Fig. 6) was performed using NK1 receptor antibodies in conjunction with PKC, recoverin, CaBP, or CD15 antibodies. Rod bipolar cells can be immunostained using antibodies directed to PKC (Greferath et al., 1990; Young and Vaney, 1991; Casini et al., 1996). Double-label studies showed that NK1 receptor and PKC immunoreactivities are in di¡erent bipolar cells, indicating that the NK1 receptor-immunoreactive bipolar cells are not rod bipolar cells (Fig. 6A). Colocalization of NK1 receptor and PKC immunostaining was not observed in bipolar cell bodies (Fig. 6B) or in axonal terminals (Fig. 6C). In particular, the NK1 receptor-immunolabeled terminal varicosities were located in a band just distal to the PKC-labeled terminals (Fig. 6A), and they also appeared to be considerably smaller than those of rod bipolar cells (Fig. 6A, C). These observations indicate that the NK1 receptor-immunoreactive bipolar cells are an ON-type cone bipolar cell. A recoverin-immunoreactive cone bipolar cell type arborizes in the proximal IPL in both the rabbit and rat retina (Euler and Wa«ssle, 1995; Massey and Mills, 1996). Recoverin immunostaining was observed in photoreceptors and in cone bipolar cells; however, none of the recoverin-immunoreactive bipolar cells contained NK1 receptor immunoreactivity in either coronal (Fig. 6D) or horizontal sections (Fig. 6E). The recoverin-immunoreactive cone bipolar cells that arborize in the proximal IPL also contain CaBP immunoreactivity (Massey and Mills, 1996). Double-label studies showed that NK1 receptor and CaBP immunoreactivities occur in di¡erent bipolar cell types (Fig. 6F). The axonal terminals of CaBP-immunoreactive bipolar cells arborized in the IPL slightly proximal to the axonal terminals of NK1 receptor-immunoreactive bipolar cells (Fig. 6F) and co-occurrence of CaBP and NK1 immunostaining was not observed in the terminal varicosities (Fig. 6G). Another type of ON-type cone bipolar cell has been NK1 receptors in rabbit retina Fig. 3. NK1 receptor immunostaining in horizontal sections. In some cases (as in A and D), the sections were cut in a tangential plane relative to the retinal surface. (A) NK1 receptor-immunoreactive bipolar cell somata could be seen in the distal INL and these cells give rise to ¢ne dendrites that extend within the OPL and form a network characterized by intensely immunoreactive puncta. The ‘punctate’ network of NK1 receptor-immunostained dendrites extended throughout the OPL (B). In some cases, single dendrites appeared to contact several puncta (B, inset). A higher power view of the NK1 receptor-immunoreactive dendrite network can be seen in C: the darkly immunoreactive puncta appeared to be the site of convergence of di¡erent dendrites. This pattern suggests that the immunoreactive dendrites may form clusters that are likely to contact photoreceptor terminals. The photomicrograph in D shows a NK1 receptor-immunoreactive bipolar cell soma (arrow) that appears to be very small with respect to an immunoreactive amacrine cell body in the proximal INL. Long processes originating from the NK1 receptor-immunoreactive amacrine cell can also be seen. In the proximal IPL, NK1 receptor-immunostained, varicose axonal terminal arborizations of bipolar cells formed a continuous meshwork (also shown in E at higher magni¢cation) where it was not possible to recognize single terminal arbors. Scale bars = 20 Wm. Calibration in A also applies to B and D. 1313 1314 G. Casini et al. about the same level in the IPL (Fig. 6H). In horizontal sections, some rare somata and perhaps dendrites (Fig. 6I) as well as rare axonal varicosities in the IPL (Fig. 6J) were seldom observed to display both NK1 and CD15 immunoreactivities. Quantitative analysis of NK1 receptor-immunoreactive bipolar cells The density of NK1 receptor-immunoreactive bipolar cell somata was determined in three areas of the visual streak and in 10 areas of the mid-peripheral ventral retina. The density of dendritic clusters was determined in nine areas of the mid-peripheral ventral retina. NK1 receptor-expressing bipolar cell bodies in the visual streak had an average density of 2248.9 Z 832.1 cells/ mm2 , while in the ventral retina the density was 1115.1 Z 344.1 cells/mm2 . Nearest-neighbor analysis (Wa«ssle and Riemann, 1978) performed in the same areas in the ventral retina used for determining cell density indicated that NK1 receptor-immunoreactive bipolar Fig. 4. Confocal images from NK1 receptor-immunostained horizontal sections. A and B are extended focuses of 3- and 5-, respectively, 1-Wm-thick optical sections, while C is a single optical section. A shows a NK1 receptor-immunoreactive bipolar cell originating four primary dendrites. Two of these dendrites could be seen to give rise to second-order rami¢cations. (B) The ‘Nomarski-like’ appearance of this confocal image, which allows a better view of the ¢ne processes, was obtained using Adobe Photoshop as follows: the original gray-scale image (background) was duplicated and put in a new layer (layer 1). Subsequently, layer 1 was shifted 2 pixels up and 2 pixels right with respect to the background. Then, the fusion mode ‘di¡erence’ was applied to layer 1. Finally, the brightness of layer 1 was increased until the desired e¡ect was obtained. Several NK1 receptor-immunostained bipolar cell somata and dendrites can be seen in this image. Some of the dendrites can be observed to apparently connect two di¡erent cell bodies (arrows), while other dendrites originating from the same cell body converge into the same clusters (arrowheads). This characteristic pattern of the NK1 receptor-immunoreactive dendrites in the OPL did not allow the identi¢cation of the dendritic ¢elds of single immunoreactive bipolar cells. (C) Immunostained varicosities of NK1 receptor-immunoreactive bipolar cell terminal arborizations can be seen in the proximal IPL. Scale bars = 10 Wm. recently detected with antibodies directed to the carbohydrate epitope CD15 in the rabbit retina (Brown and Masland, 1999). NK1 receptor and CD15 immunoreactivities were in di¡erent cell bodies, indicating two separate cell populations with terminal axons arborizing at Fig. 5. Photomicrographs showing that the NK1 receptor-immunoreactive bipolar cell in A (arrow) is also labeled by MAb115A10, a bipolar cell marker (B, arrow). In this section, NK1 receptor immunostaining was visualized using Cy3-conjugated secondary antibodies, while MAb115A10 immunoreactivity was detected with Alexa Fluor 488-conjugated secondary antibodies. Scale bar = 10 Wm. NK1 receptors in rabbit retina 1315 Fig. 6. Confocal images from coronal and horizontal sections double labeled with a NK1 antiserum in conjunction with antibodies directed to di¡erent markers of bipolar cells of the rabbit retina. In all images, green immuno£uorescence (from Alexa Fluor 488-conjugated secondary antibodies) identi¢es NK1 receptor immunostaining, while red immuno£uorescence (from Alexa Fluor 546-conjugated secondary antibodies) identi¢es PKC (A^C), recoverin (D, E), CaBP (F, G) or CD15 (H^J) immunostaining. All images are 1-Wm-thick confocal sections. The arrows in I and J point to yellow-colored pro¢les displaying both NK1 receptor and CD15 immunolabeling. See the text for description. Scale bars = 10 Wm. Calibration in A also applies to D, E, F, H, I and J. cells are non-randomly distributed (Fig. 7). The ratio of the mean/standard deviation of nearest-neighbor distances is an index of mosaic regularity (Eberhardt, 1967). The higher this ratio, the greater the regularity. The values calculated for NK1 receptor-immunoreactive bipolar cells ranged from 3.1 to 3.8, indicating a fairly regular pattern. Density measurements of the NK1 receptor-immunoreactive dendritic clusters in ventral retina gave an average of 13 586.5 Z 2260.6 clusters/mm2 . NK1 receptor-immunoreactive amacrine cells NK1 receptor-immunoreactive amacrine cells were characterized by a medium to large soma (see Figs. 1 and 3D for a comparison of bipolar and amacrine cell sizes). These cells were sparsely distributed throughout the proximal INL, and they were characterized by one to three thick primary processes that arborized into second- and higher-order rami¢cations within lamina 1 1316 G. Casini et al. Fig. 7. Nearest-neighbor distance histograms for NK1 receptor-immunoreactive bipolar cells in one of the areas selected for analysis in mid-peripheral ventral retina. The dashed curve illustrates the relative frequencies of distances to nearest neighbors that would be expected if the NK1 receptor-immunoreactive bipolar cells were randomly distributed (according to a Poisson probability rule) with density equal to that observed experimentally. The solid curve illustrates the normal Gaussian distribution expected for the same values of nearest-neighbor distance if these cells were non-randomly distributed (Wa«ssle and Riemann, 1978). The experimental values show better correlation with the Gaussian curve than with the random one. of the IPL. Some of the immunoreactive ¢bers displayed varicosities (Fig. 8). Overall, the arborizations of the NK1 receptor-expressing amacrine cells formed a loose meshwork that covered the entire retinal area. The large soma size, the sparse distribution and the pattern of process arborization suggest that these cells are TH-containing amacrine cells (Casini and Brecha, 1992a,b). To assess whether NK1 receptors are expressed by TH-containing amacrine cells, NK1 receptor antibodies were used in conjunction with monoclonal antibodies directed to TH. All NK1 receptor-immunoreactive amacrine cells contained TH immunoreactivity (Fig. 9A, B) and, conversely, all TH-immunolabeled amacrine cells contained NK1 receptor immunoreactivity. The prominent TH-immunoreactive processes in lamina 1 of the IPL were also NK1 receptor-immunoreactive, although most of the ¢nest, varicose processes were devoid of NK1 receptor immunostaining, including the TH-containing ¢bers which form the characteristic ‘rings’ (Fig. 9C-E) that encircle and make synaptic contacts with AII amacrine cell bodies (Pourcho, 1982; Voigt and Wa«ssle, 1987; Kolb et al., 1990, 1991; Casini et al., 1995). In addition, the less intensely staining TH-immunolabeled processes in laminae 3 and 5 of the IPL were not NK1 receptor-immunoreactive. DISCUSSION The present study has investigated the localization of NK1 receptors, whose preferred ligand is SP (Otsuka and Yoshioka, 1993), in rabbit retinas. Together, the results demonstrate that NK1 expression is con¢ned to a population of ON-type bipolar cells and to TH-containing, dopaminergic amacrine cells. Bipolar cells expressing NK1 receptors The bipolar cells displaying NK1 receptor immunoreactivity appear to form a distinct population of bipolar cells in the rabbit retina. These cells are di¡erent from rod bipolar cells, therefore they must be a type of cone bipolar cell. Interestingly, in rat retina NK1 receptors are mainly expressed by amacrine cells and not by bipolar cells (Casini et al., 1997b, 2000; Oyamada et al., 1999). Cone bipolar cells can be functionally divided into two main types, ON and OFF, according to their response to light (Famiglietti and Kolb, 1976; Nelson and Kolb, 1983). Morphologically, ON and OFF types are distinguished by the location of their terminal arborizations in the IPL: ON-type bipolar cells arborize in sublamina b (proximal half), while OFF-type bipolar cells arborize in sublamina a (distal half) of the IPL (Famiglietti and NK1 receptors in rabbit retina Fig. 8. NK1 receptor-immunoreactive amacrine cells in a horizontal section. In this photomicrograph, the immunostaining outlines the shapes of two large-sized cell bodies and their processes. Some ¢ne immunoreactive processes were observed to carry varicosities. See the text for further description. Scale bar = 15 Wm. Fig. 9. Photomicrographs showing NK1 receptor expression in TH-containing amacrine cells. In all images, red immuno£uorescence (from Alexa Fluor 546-conjugated secondary antibodies) identi¢es NK1 receptor immunostaining, while green immuno£uorescence (from Alexa Fluor 488-conjugated secondary antibodies) identi¢es TH immunoreactivity. A depicts a largesized amacrine cell displaying TH immunostaining that also bears NK1 receptor immunostaining on its plasma membrane. Note that small TH-containing varicosities in the distal IPL are not labeled by NK1 receptor antibodies (arrowheads), while one larger varicosity is double-labeled (arrow). A NK1 receptor-immunoreactive bipolar cell can also be seen. B shows some NK1 receptor-immunoreactive bipolar cells together with another TH-containing amacrine cell that is also NK1 receptorimmunolabeled. Note that TH-containing varicosities in laminae 3 and 5 (numbers on the right indicate IPL laminae) are not immunostained, and the NK1 receptor immunostaining in the proximal IPL is con¢ned to bipolar axon terminals. C^E are photomicrographs from a double-labeled horizontal section, with E being a composite image of C and D. The arrows indicate a TH-containing ‘ring’ structure that lacks NK1 receptor immunostaining. Similarly, NK1 receptor immunoreactivity was absent in the thinnest TH-immunoreactive ¢bers (arrowheads). Scale bars: A, 15 Wm; B, 20 Wm; C^E, 10 Wm. 1317 1318 G. Casini et al. Kolb, 1976; Nelson et al., 1978; Peichl and Wa«ssle, 1981; Euler and Wa«ssle, 1995; Hartveit, 1997). On the basis of their axonal arborization pattern in the IPL, the NK1 receptor-expressing bipolar cells can be classi¢ed as ON-type cone bipolar cells. Multiple types of cone bipolar cells have been identi¢ed by Golgi, histochemical, immunohistochemical and cell-labeling methods in the rabbit retina (Famiglietti, 1981; Mills and Massey, 1992; Jeon and Masland, 1995; Massey and Mills, 1996; Merighi et al., 1996; Brown and Masland, 1999; McGillem and Dacheux, 2001). These bipolar cells di¡er in the size of their dendritic ¢elds in the OPL (narrow- or wide-¢eld), and in the level of their axonal terminal arborizations in sublamina a or b of the IPL. The NK1 receptor-immunoreactive bipolar cells do not correspond to any of the ON-type cone bipolar cells previously identi¢ed in the rabbit retina. For example, the sparsely distributed, wide-¢eld, ON-type cone bipolar cells that correspond most closely to Wb cone bipolar cells (Famiglietti, 1981) have a cell density of less than 70 cells/mm2 in the peripheral retina (Jeon and Masland, 1995). In contrast, the NK1 receptor-immunoreactive bipolar cells have a cellular density that is greater than 1000 cells/mm2 in the ventral mid-peripheral retina. NK1 receptor-immunoreactive bipolar cells are also di¡erent from CaBP-immunoreactive narrow-¢eld, ON-type cone bipolar cells (Massey and Mills, 1996). The CaBP-immunoreactive bipolar cells, which also express recoverin immunoreactivity (Massey and Mills, 1996), correspond best to type Nb1 or Nb2 bipolar cells of Famiglietti (1981) or to type CBmb5 of McGillem and Dacheux (2001). Antibodies directed to CD15 label a distinct population of narrow-¢eld ON-type cone bipolar cells, although weak CD15 immunostaining is also seen in other bipolar cell bodies (Brown and Masland, 1999). The very rare co-occurrence of NK1 receptor and CD15 immunostaining in some bipolar cells shows that there is little or no overlap between the NK1 receptor- and prominently staining CD15-immunoreactive bipolar cell populations. Therefore, we interpreted the CD15- and the NK1 receptor-immunoreactive bipolar cells as constituting two separate ON-type bipolar cell populations. The CD15immunoreactive bipolar cells have been suggested to correspond to type CBmb4 (McGillem and Dacheux, 2001). On the basis of the level of their axonal arborizations in the IPL, NK1-expressing bipolar cells may be another type of CBmb4 or they may correspond to type CBmb4^5 (McGillem and Dacheux, 2001). Taken together, these observations indicate that NK1 receptor is expressed by a previously unidenti¢ed type of ON-type cone bipolar cell in the rabbit retina. These cells also meet the criteria to be considered a distinct population. Indeed, NK1 receptor-immunoreactive cell bodies are present in all retinal regions, while dendritic clusters as well as axonal terminal arbors are in all retinal areas, thus indicating complete retinal coverage by these cells. In addition, cell densities were higher in the visual streak than in the peripheral retina. Finally, the analysis of nearest-neighbor distances indicated that the immunore- active cell bodies are arranged in a regular mosaic. Interestingly, the average density of immunoreactive dendritic clusters in ventral mid-periphery (about 13 000/mm2 ) was similar to the reported cone density of 10 000^12 000 cones/mm2 in this region (Famiglietti and Sharpe, 1995), suggesting that all cones are contacted by the dendrites of NK1 receptor-expressing bipolar cells. Amacrine cells expressing NK1 receptors TH-containing amacrine cells display NK1 receptor immunoreactivity, similar to previous observations in the rat retina (Casini et al., 1997b). In rat retinas, THcontaining amacrine cells are composed of two di¡erent populations that can be distinguished morphologically and neurochemically: the small-sized TH-immunoreactive cells are adrenergic amacrine cells, while the largesized ones are dopaminergic (Versaux-Botteri et al., 1986), and both populations were found to express NK1 receptors (Casini et al., 1997b). However, there is a di¡erence in the NK1 receptor expression pattern between rabbit and rat retinas. While in rabbits all dopaminergic amacrine cells display NK1 receptors, only 71% of the rat large TH-immunostained amacrine cells are labeled with antibodies directed to NK1 receptor (Casini et al., 1997b). In addition, in the rat retina, NK1 receptor immunoreactivity is localized to numerous amacrine cells, including GABA-immunoreactive amacrine cells, that ramify in all laminae of the IPL (Casini et al., 1997b). Functional implications The cellular expression of NK1 receptors in the rabbit retina is consistent with SP having a direct action on ON-type cone pathways and indirect actions, mediated by the dopamine-containing amacrine cells, on the rod pathway (Fig. 10). In a rabbit eyecup preparation, application of SP at nanomolar concentrations increases the excitability of the majority of ON, OFF, and ON^OFF center ganglion cells, and of some amacrine cells (Zalutsky and Miller, 1990a). In particular, these earlier data suggested both direct and indirect actions of SP onto ganglion cells. Our ¢ndings, however, indicate that direct e¡ects of SP on ganglion cells are unlikely, and ganglion cell activity may be regulated by SP through cone and rod pathways. In addition, possible targets of SP actions were identi¢ed in GABAergic and in putative AII amacrine cells (Zalutsky and Miller, 1990a). Although NK1 receptors are expressed by GABAergic amacrine cells in rat retinas (Casini et al., 1997b), our observations seem to exclude the possibility that in rabbit retinas GABA-containing amacrine cells may respond to SP. On the other hand, the presence of NK1 receptors in dopaminergic amacrine cells may indicate indirect actions of SP onto AII amacrine cells. Indeed, dopaminergic amacrine cells are likely to be involved in the modulation of AII amacrine cells (Jensen, 1989; Hampson et al., 1992), which are important interneurons of the rod pathway (see Wa«ssle and Boycott, 1991; Bloom¢eld and Dacheux, 2001, for NK1 receptors in rabbit retina 1319 Fig. 10. Wiring diagram showing the retinal circuitry mediating SP neurotransmission via NK1 receptors. SP released by SPcontaining amacrine cells acts at NK1 receptors expressed by ON-type cone bipolar cells (CBC, on the right) and therefore it is likely to in£uence ganglion cell (GC) activity. NK1 receptors are also expressed by TH-containing, dopaminergic amacrine cells and mediate SP-induced excitation of these cells resulting in dopamine (DA) release and modulation of AII amacrine cells. The AII amacrine cells are important interneurons in the rod pathway. They receive signals from rod bipolar cells (RBC) through chemical synapses and contact cone bipolar cells (CBC, on the left) through electrical synapses. From these cone bipolar cells, the rod signal reaches the ganglion cells. The peptidergic control of the rod pathway and of dopamine release, in addition to SP, may also include somatostatin (SRIF). SRIF released by SRIF-containing displaced amacrine cells (Rickman et al., 1996) may in£uence rod bipolar cells through somatostatin subtype 2A receptors (sst2A, Johnson et al., 1998; Fontanesi et al., 2000) and dopaminergic amacrine cells through somatostatin subtype 1 receptors (sst1, Cristiani et al., 2000). While SP may act as a light signal and induce dopamine release, SRIF may act as a dark signal and inhibit dopamine release. Note that, although synapses are schematically represented in this diagram, both SP and SRIF may also act in a paracrine manner. reviews). SP acting at NK1 receptors on dopaminergic amacrine cells could indirectly a¡ect the £ow of visual information through the rod pathway at the level of AII amacrine cells (Fig. 10). Finally, Zalutsky and Miller (1990a) reported lack of SP e¡ects on horizontal cells and ruled out possible actions of SP at the level of ON bipolar cells. While the absence of SP in£uence on horizontal cells is consistent with our ¢ndings, the present investigation demonstrates that ON-type cone bipolar cells are primary targets of SP actions in the rabbit retina. SP may play a role in cellular processes underlying light adaptation by its in£uence on dopamine release from dopamine-containing amacrine cells. Earlier studies have demonstrated that extracellular dopamine levels increase in the retina with increasing light intensity (Djamgoz and Wagner, 1992; Boelen et al., 1998). If SP evokes dopamine release in the rabbit retina like in the rat retina (Tsang, 1986), this peptide may also act as a light signal. Light adaptation likely involves the integrated action of multiple systems including at least two peptidergic systems (Fig. 10). The somatostatin subtype 1 receptor is expressed by dopamine-containing amacrine cells of the rabbit retina (Cristiani et al., 2000), and somatostatin’s action in the rabbit eyecup preparation is consistent with a role for this peptide in light adapta- tion (Zalutsky and Miller, 1990b). Furthermore, in the chick retina, somatostatin acts as a dark signal (Ishimoto et al., 1986; Yang et al., 1997). Based on these observations, dopamine levels may be up-regulated by a stimulatory action of SP acting at NK1 receptors and downregulated by an inhibitory action of somatostatin at somatostatin subtype 1 receptors. The interaction between the light signal (SP) and the dark signal (somatostatin) would therefore modulate dopamine release and light adaptation. In rat retina, there is a similar, but not identical, expression pattern of NK1 and somatostatin receptors on dopamine-containing amacrine cells (Casini et al., 1997b; Helboe and Moller, 1999; Johnson et al., 1999) and, therefore, it is likely that there are some differences in the cellular mechanisms mediating light adaptation in the rabbit and in the rat retina. Peptide receptor expression in rat and in rabbit retinas The data reported in the present study indicate that the distribution pattern of NK1 receptors in rat retinas (Casini et al., 1997b, 2000a; Oyamada et al., 1999) is not conserved in the rabbit. In rat and in rabbit retina, NK1 receptors are expressed by TH-immunoreactive amacrine cells. However, there are unexpected di¡erences between the two species: in the rat retina, NK1 receptors are 1320 G. Casini et al. mainly expressed by amacrine cells, while in the rabbit retina they are expressed by a population of cone bipolar cells and by the population of dopaminergic amacrine cells. These ¢ndings, together with previous observations on the di¡erential retinal localization of somatostatin receptors (Johnson et al., 1998, 1999; Helboe and Moller, 1999; Cristiani et al., 2000; Fontanesi et al., 2000), indicate a surprising diversity in peptide receptor expression in the retinas of di¡erent mammalian species. A di¡erent organization of peptidergic systems may determine species-speci¢c characterization of retinal function. These characterizations are likely to be related to behavioral habits and be important, for example, in determining di¡erences for adaptation to nocturnal vs. diurnal life. Acknowledgements&We thank Dr. S.C. Fujita, of the Mitsubishi Kasei Institute of Life Sciences, Tokyo, Japan, for providing MAb115A10 and Dr. K.-W. Koch, of the Institut fu«r Biologische Informationsverarbeitung, Ju«lich, Germany, for providing the recoverin antiserum. This work was supported by the Italian Board of Education, by the National Institutes of Health (Grant No. EY 04067) and by a VA Career Scientist Award. REFERENCES Ayoub, G.S., Matthews, G., 1992. Substance P modulates calcium current in retinal bipolar neurons. Vis. Neurosci. 8, 539^544. Bartho, L., Holzer, P., 1985. Search for a physiological role of substance P in gastrointestinal motility. Neuroscience 16, 1^32. Bloom¢eld, S.A., Dacheux, R.F., 2001. Rod vision: pathways and processing in the mammalian retina. Prog. Retinal Eye Res. 20, 351^384. Boelen, M.K., Boelen, M.G., Marshak, D.W., 1998. Light-stimulated release of dopamine from the primate retina is blocked by 1^2-amino-4phosphonobutyric acid (APB). Vis. Neurosci. 15, 97^103. Brecha, N.C., Johnson, D., Bolz, J., Sharma, S., Parnavelas, J.G., Lieberman, A.R., 1987. Substance P-immunoreactive retinal ganglion cells and their central axon terminals in rabbit. Nature 327, 155^158. Brown, S.P., Masland, R.H., 1999. Costrati¢cation of a population of bipolar cells with the direction-selective circuitry of the rabbit retina. J. Comp. Neurol. 408, 97^106. Cajal, S.R., 1893. La re¤tine des verte¤bre¤s. Cellule 9, 119^257. Caruso, D.M., Owczarzak, M.T., Pourcho, R.G., 1990. Colocalization of substance P and GABA in retinal ganglion cells: a computer-assisted visualization. Vis. Neurosci. 5, 389^394. Casini, G., Brecha, N.C., 1992a. Postnatal development of tyrosine hydroxylase immunoreactive amacrine cells in the rabbit retina. I. Morphological characterization. J. Comp. Neurol. 326, 283^301. Casini, G., Brecha, N.C., 1992b. Postnatal development of tyrosine hydroxylase immunoreactive amacrine cells in the rabbit retina. II. Quantitative analysis. J. Comp. Neurol. 326, 302^313. Casini, G., Brecha, N.C., Bosco, L., Rickman, D.W., 2000a. Developmental expression of neurokinin-1 and neurokinin-3 receptors in the rat retina. J. Comp. Neurol. 421, 275^287. Casini, G., Grassi, A., Trasarti, L., Bagnoli, P., 1996. Developmental expression of protein kinase C immunoreactivity in rod bipolar cells of the rabbit retina. Vis. Neurosci. 13, 817^831. Casini, G., Rickman, D.W., Brecha, N.C., 1995. The AII amacrine cell population in the rabbit retina : identi¢cation with parvalbumin immunoreactivity. J. Comp. Neurol. 355, 1^11. Casini, G., Rickman, D.W., Sternini, C., Brecha, N.C., 1997b. Neurokinin 1 receptor expression in the rat retina. J. Comp. Neurol. 389, 496^507. Casini, G., Sabatini, A., Trasarti, L., Catalani, E., Marroni, P., Bosco, L., Brecha, N.C., 2000b. Neurokinin 1 (NK1) receptor expression in rabbit retina. Soc. Neurosci. Abstr. 26, 2152. Casini, G., Trasarti, L., Andol¢, L., Bagnoli, P., 1997a. Morphologic maturation of tachykinin peptide-expressing cells in the postnatal rabbit retina. Dev. Brain Res. 99, 131^141. Chen, L.W., Wei, L.C., Liu, H.L., Duan, L., Ju, G., Chan, Y.S., 2000. Retinal dopaminergic neurons (A17) expressing neuromedin K receptor (NK(3)): a double immunocytochemical study in the rat. Brain Res. 885, 122^127. Cristiani, R., Fontanesi, G., Casini, G., Petrucci, C., Bagnoli, P., 2000. Expression of somatostatin subtype receptor 1 (sst1 ) in the rabbit retina. Invest. Ophthalmol. Vis. Sci. 41, 3191^3199. Cuenca, N., Kolb, H., 1998. Circuitry and role of substance P-immunoreactive neurons in the primate retina. J. Comp. Neurol. 393, 439^456. Dick, E., Miller, R.F., 1981. Peptides in£uence retinal ganglion cells. Neurosci. Lett. 26, 131^135. Djamgoz, M.B., Wagner, H.J., 1992. Localization and function of dopamine in the adult vertebrate retina. Neurochem. Int. 20, 139^191. Eberhardt, L.L., 1967. Some developments in ‘distance sampling’. Biometrics 23, 207^216. Euler, T., Wa«ssle, H., 1995. Immunocytochemical identi¢cation of cone bipolar cells in the rat retina. J. Comp. Neurol. 361, 461^478. Famiglietti, E.V., 1981. Functional architecture of cone bipolar cells in mammalian retina. Vis. Res. 21, 1559^1563. Famiglietti, E.V., Kolb, H., 1976. Structural basis for ON- and OFF-center responses in retinal ganglion cells. Science 194, 193^195. Famiglietti, E.V., Sharpe, S.J., 1995. Regional topography of rod and immunocytochemically characterized ‘blue’ and ‘green’ cone photoreceptors in rabbit retina. Vis. Neurosci. 12, 1151^1175. Fontanesi, G., Gargini, C., Bagnoli, P., 2000. Postnatal development of somatostatin 2A (sst2A) receptors expression in the rabbit retina. Dev. Brain Res. 123, 67^80. Glickman, R.D., Adolph, A.R., Dowling, J.E., 1982. Inner plexiform circuits in the carp retina: e¡ects of cholinergic agonists, GABA, and substance P on the ganglion cells. Brain Res. 234, 81^99. Goehler, L.E., Sternini, C., Brecha, N.C., 1988. Calcitonin gene-related peptide immunoreactivity in the biliary pathway and liver of the guinea pig: distribution and colocalization with substance P. Cell Tissue Res. 253, 145^150. Greferath, U., Gru«nert, U., Wa«ssle, H., 1990. Rod bipolar cells in the mammalian retina show protein kinase C-like immunoreactivity. J. Comp. Neurol. 301, 433^442. Hampson, E.C.G.M., Vaney, D.I., Weiler, R., 1992. Dopaminergic modulation of gap junction permeability between amacrine cells in mammalian retina. J. Neurosci. 12, 4911^4922. Hartveit, E., 1997. Functional organization of cone bipolar cells in the rat retina. J. Neurophysiol. 4, 1716^1730. Helboe, L., Moller, M., 1999. Immunohistochemical localization of somatostatin receptor subtypes sst1 and sst2 in the rat retina. Invest. Ophthalmol. Vis. Sci. 40, 2376^2382. Ishimoto, I., Millar, T., Chubb, I.W., Morgan, I.G., 1986. Somatostatin-immunoreactive amacrine cells of chicken retina : retinal mosaic, ultrastructural features, and light-driven variations in peptide metabolism. Neuroscience 17, 1217^1233. NK1 receptors in rabbit retina 1321 Jensen, R.J., 1989. Mechanism and site of action of a dopamine D1 antagonist in the rabbit retina. Vis. Neurosci. 3, 573^585. Jeon, C.J., Masland, R.H., 1995. A population of wide-¢eld bipolar cells in the rabbit’s retina. J. Comp. Neurol. 360, 403^412. Johnson, J., Wong, H., Walsh, J.H., Brecha, N.C., 1998. Expression of the somatostatin subtype 2A receptor in the rabbit retina. J. Comp. Neurol. 393, 93^101. Johnson, J., Wu, V., Wong, H., Walsh, J.H., Brecha, N.C., 1999. Somatostatin receptor subtype 2A expression in the rat retina. Neuroscience 94, 675^683. Jonsson, G., Hallman, H., 1982a. Substance P counteracts neurotoxin damage on norepinephrine neurons in rat brain during ontogeny. Science 125, 75^77. Jonsson, G., Hallman, H., 1982b. Substance P modi¢es the 6-hydroxydopamine induced alteration of postnatal development of central noradrenaline neurons. Neuroscience 7, 2909^2918. Kolb, H., Cuenca, N., Dekorver, L., 1991. Postembedding immunocytochemistry for GABA and glycine reveals the synaptic relationships of the dopaminergic amacrine cell of the cat retina. J. Comp. Neurol. 310, 267^284. Kolb, H., Cuenca, N., Wang, H.H., Dekorver, L., 1990. The synaptic organization of the dopaminergic amacrine cell in the cat retina. J. Neurocytol. 19, 343^366. Kolb, H., Fernandez, E., Ammermuller, J., Cuenca, N., 1995. Substance P: a neurotransmitter of amacrine and ganglion cells in the vertebrate retina. Histol. Histopathol. 10, 947^968. Kondoh, A., Houtani, T., Ueyama, T., Baba, K., Ikeda, M., Yamagishi, K., Miki, H., Uyama, M., Nakanishi, S., Sugimoto, T., 1996. In situ hybridization analysis of substance P receptor in the rat retina. Exp. Brain Res. 112, 181^186. Lambrecht, H.-G., Koch, K.-W., 1992. Recoverin, a novel calcium binding protein from vertebrate photoreceptors. Biochim. Biophys. Acta 1160, 63^66. Li, H.B., So, K.F., Cheuk, W., 1999. Substance P-immunoreactive neurons in hamster retinas. Vis. Neurosci. 16, 475^481. Maggio, J.E., 1988. Tachykinins. Annu. Rev. Neurosci. 11, 13^28. Massey, S.C., Mills, S.L., 1996. A calbindin-immunoreactive cone bipolar cell type in the rabbit retina. J. Comp. Neurol. 366, 15^33. McGillem, G.S., Dacheux, R.F., 2001. Rabbit cone bipolar cells: correlation of their morphologies with whole-cell recordings. Vis. Neurosci. 18, 675^685. Merighi, A., Raviola, E., Dacheux, R.F., 1996. Connections of two types of £at cone bipolars in the rabbit retina. J. Comp. Neurol. 371, 164^178. Mills, S.L., Massey, S.C., 1992. Morphology of bipolar cells labeled by DAPI in the rabbit retina. J. Comp. Neurol. 321, 133^149. Nelson, R., Famiglietti, E.V., Kolb, H., 1978. Intracellular staining reveals di¡erent levels of strati¢cation for on-center and o¡-center ganglion cells in the cat retina. J. Neurophysiol. 4, 427^483. Nelson, R., Kolb, H., 1983. Synaptic patterns and response properties of bipolar and ganglion cells in the cat retina. Vis. Res. 23, 1183^1195. Nilsson, J., von Euler, A.M., Dalsgaard, C.J., 1985. Stimulation of connective tissue cell growth by substance P and substance K. Nature 315, 61^ 63. Onoda, N., Fujita, S.C., 1987. A monoclonal antibody speci¢c for a subpopulation of retinal bipolar cells in the frog and other vertebrates. Brain Res. 416, 359^363. Otsuka, M., Yoshioka, K., 1993. Neurotransmitter functions of mammalian tachykinins. Physiol. Rev. 73, 229^308. Oyamada, H., Takatsuji, K., Senba, E., Manthy, P.W., Tohyama, M., 1999. Postnatal development of NK1, NK2 and NK3 neurokinin receptors expression in the rat retina. Dev. Brain Res. 117, 59^70. Peichl, L., Wa«ssle, H., 1981. Morphological identi¢cation of ON- and OFF-centre brisk transient (Y) cells in the cat retina. Proc. R. Soc. London B Biol. Sci. 212, 139^153. Pernow, B., 1983. Substance P. Pharmacol. Rev. 35, 85^141. Pourcho, R.G., 1982. Dopaminergic amacrine cells in the cat retina. Brain Res. 252, 101^109. Rickman, D.W., Blanks, J.C., Brecha, N.C., 1996. Somatostatin-immunoreactive neurons in the adult rabbit retina. J. Comp. Neurol. 365, 491^ 503. Rohrer, H., Acheson, A.L., Thibault, J., Thoenen, H., 1986. Developmental potential of quail dorsal root ganglion cells analyzed in vitro and in vivo. J. Neurosci. 6, 2616^2624. Shughrue, P.J., Lane, M.V., Merchenthaler, I., 1996. In situ hybridization analysis of the distribution of neurokinin-3 mRNA in the rat central nervous system. J. Comp. Neurol. 372, 395^414. Takatsuji, K., Miguel-Hidalgo, J.J., Tohyama, M., 1991. Substance P-immunoreactive innervation from the retina to the suprachiasmatic nucleus in the rat. Brain Res. 568, 223^229. Tsang, D., 1986. E¡ect of substance P on dopamine release in rat retina. In: Kon, O.L., Chung, M.C.M., Hwang, P.L.H., Leong, S.F., Loke, K.H., Thiayagarajah, P., Wong, P.H.T. (Eds.), Contemporary Themes in Biochemistry. Cambridge University Press, Cambridge, pp. 588^589. Versaux-Botteri, C., Martin-Martinelli, E., Nguyen-Legros, J., Ge¡ard, M., Vigny, A., Denoroy, L., 1986. Regional specialization of the rat retina : catecholamine-containing amacrine cell characterization and distribution. J. Comp. Neurol. 243, 422^433. Vigna, S.R., Bowden, J.J., McDonald, D.M., Fisher, J., Okamoto, A., McVey, D.C., Payan, D.J., Bunnett, N.W., 1994. Characterization of antibodies to the rat substance P (NK-1) receptor and to a chimeric substance P receptor expressed in mammalian cells. J. Neurosci. 14, 834^ 845. Voigt, T., Wa«ssle, H., 1987. Dopaminergic innervation of AII amacrine cells in mammalian retina. J. Neurosci. 7, 4115^4128. Wa«ssle, H., Boycott, B.B., 1991. Functional architecture of the mammalian retina. Physiol. Rev. 71, 447^480. Wa«ssle, H., Riemann, H.J., 1978. The mosaic of nerve cells in the mammalian retina. Proc. R. Soc. London B Biol. Sci. 200, 441^461. Yang, D.S., Li, Z.K., Morgan, I.G., 1997. A light-driven rhythm in neurotensin-like immunoreactivity in the chicken retina. Aust. New Zealand J. Ophthalmol. 1 (Suppl), S67^D69. Young, H.M., Vaney, D.I., 1991. Rod-signal interneurons in the rabbit retina. 1. Rod bipolar cells. J. Comp. Neurol. 310, 139^153. Zalutsky, R.A., Miller, R.F., 1990a. The physiology of substance P in the rabbit retina. J. Neurosci. 10, 394^402. Zalutsky, R.A., Miller, R.F., 1990b. The physiology of somatostatin in the rabbit retina. J. Neurosci. 10, 383^393. (Accepted 14 June 2002)