Expression of Somatostatin Subtype 1 Receptor in the Rabbit Retina

Abstract

purpose. To detect mRNAs for somatostatin (somatotropin release-inhibiting factor [SRIF]) receptor subtypes 1 to 5 (sst₁ through sst₅) in rabbit retinas by reverse transcription–polymerase chain reaction (RT-PCR) and to investigate the distribution of sst₁ by single- and double-label immunocytochemistry.

methods. Semiquantitative RT-PCR using sst-specific primers from mouse sequences was performed. sst₁ was localized using a polyclonal antiserum directed to human sst₁ in cryostat sections of retinas from either normal or optic nerve–transected animals. Immunolabeled cell sizes and densities were measured in wholemounted retinas using computer-assisted image analysis. Double-label immunofluorescence was performed using the sst₁ antiserum in conjunction with monoclonal antibodies directed to SRIF, tyrosine hydroxylase (TH), parvalbumin (PV), or γ-aminobutyric acid (GABA).

results. With RT-PCR it was found that all five sst mRNAs were expressed in the rabbit retina, with highest levels of sst₁ mRNA. sst₁ immunolabeling was localized to amacrine cells in the proximal inner nuclear layer (INL) of all retinal regions and to displaced amacrine cells in the ganglion cell layer (GCL) of the ventral retina. Some large sst₁-immunoreactive (IR) somata were also present in the GCL. They were not observed after optic nerve transection. Double-label immunofluorescence showed sst₁ expression by all TH-IR amacrine cells and by other amacrine cells that were neither PV-IR nor GABA-IR. In addition, sst₁ was expressed by all SRIF-containing displaced amacrine cells.

conclusions. All five sst mRNAs are expressed in the rabbit retina. The localization of sst₁ suggests that it may mediate SRIF actions onto amacrine (including dopaminergic) and sparse ganglion cells. sst₁ expression in SRIF-IR cells suggests that this receptor may also act as an autoreceptor.

Somatotropin release-inhibiting factor (SRIF) is a neurotransmitter and a neuromodulator in the central nervous system. ^{1 2} Five different ssts have been cloned and designated sst₁ through sst₅. ³ Although there is a high degree of sequence and structural homology among different ssts, they differ in their pharmacologic and functional properties. ⁴ For instance, sst₁ and sst₂ differ in their affinity to specific SRIF agonists and in their modes of transmembrane signaling. ⁴ Both sst₁ and sst₂ are coupled to inhibition of adenylate cyclase (AC), but sst₂, and not sst₁, internalizes or desensitizes after exposure to the agonist. ⁴ These diversities may underlie different functional roles of the two receptors. In particular, although both sst₁ and sst₂ are involved in regulation of growth hormone secretion, ⁵ sst₁ may also act as an autoreceptor and inhibit SRIF release. ⁶ In addition, activation of sst₁ increases nerve cell responses to glutamate (GLU), whereas activation of sst₂ results in a decrease of GLU sensitivity. ⁷ Moreover, sst₂, but not sst₁, has been reported to regulate Ca²⁺ influx through voltage-gated Ca²⁺ channels. ⁸  

SRIF cell populations have been reported in a variety of vertebrate retinas. ^{9 10 11 12 13} In the rabbit, SRIF is expressed by sparse displaced amacrine cells in the ganglion cell layer (GCL) of the ventral retina. In spite of low cellular density of SRIF somata, SRIF processes extensively arborize in the inner plexiform layer (IPL) of all retinal regions, which suggests that SRIF may influence several cell types by acting at multiple levels of the retinal circuitry. ^{12 14 15 16}  

SRIF influences on retinal function are poorly understood. In the rabbit retina, SRIF has been shown to influence the spontaneous firing of retinal ganglion cells (GCs) and induce modifications in their receptive fields. ¹⁷ In the rat retina, SRIF has been suggested to modulate γ-aminobutyric acid (GABA)-ergic transmission through phosphorylation of GABA_A receptors. ¹⁸ In the avian retina, SRIF may participate in a dark–light switch operating through inhibition of dopamine (DA) release. ¹⁹ Both in avian and in rat retinas, SRIF appears to be positively coupled to AC, which is surprising, because ssts are generally thought to be negatively coupled to AC. ^{18 20}  

To get a deeper insight into SRIF functions, data on the retinal localization of specific ssts are needed. Of the two sst₂ isoforms, sst_2A has been immunohistochemically localized in rabbit ^{21 22} and in rat ^{23 24} retinas. In rabbits, it is expressed mainly by rod bipolar cells and by sparse amacrine cells. These amacrines have been reported to have no tyrosine hydroxylase (TH)-immunoreactivity (IR) ²¹ or to partially express ²² TH-IR. In the rat retina, sst_2A has been localized to amacrine cells, including TH-IR amacrine cells, to rod and cone bipolar cells, and to horizontal cells. ^{23 24} sst₁ expression has been investigated in rat retinas, where it can be observed in SRIF-expressing and displaced amacrine cells, as well as in rare GCs. ²³ Information on the localization of sst₁ in rabbits is needed to complete our understanding of sst_2A and sst₁ expression in rabbit and rat retinas and to inquire about possible species-specific functional roles of sst_2A and sst₁ in the retinas of rats and rabbits. 

In the present investigation, we first used reverse transcription–polymerase chain reaction (RT-PCR) to determine the relative levels of the five different sst mRNAs in the rabbit retina and found that sst₁ mRNA is the most abundantly expressed. Subsequently, both single- and double-label immunohistochemistry was performed to investigate the cellular expression pattern of sst₁ in rabbit retinas. 

Methods

Animals and Tissue Preparation

New Zealand albino rabbits were used in compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Rabbits were deeply anesthetized with 30% chloral hydrate in sterile saline (intraperitoneally, 1.0 ml/kg; Sigma, St. Louis, MO). For PCR analysis, retinas were dissected in RNase-free conditions and stored at− 80°C. For immunohistochemical experiments, both retinal sections and whole retinas were used. Retinas were immersion fixed in 4% paraformaldehyde in 0.1 M phosphate buffer (PB; pH 7.4) for 2 hours. They were cryoprotected with 25% sucrose in 0.1 M PB, and sections were cut either perpendicular or parallel to the vitreal surface at 12 to 15 μm with a cryostat, mounted onto gelatin-coated slides, and stored at −20°C. Alternatively, whole retinas were frozen and thawed, treated with 2.3% sodium metaperiodate in distilled water and subsequently with 1% sodium borohydride in 0.25 M Tris buffer before immunocytochemical staining. 

Optic Nerve Transection

Rabbits were deeply anesthetized with a mixture of medetomidine (intraperitoneally, 0.5 ml/kg; Domitor; Farmos, Turku, Finland) and Ketamine hydrochloride (1 ml/kg; Inoketam; Virbac, Carros, France). The optic nerve was transected after a procedure described by Rickman et al. ¹² Surgery was performed under sterile conditions, and National Institutes of Health guidelines were observed. Briefly, after local infiltration of procaine (Novocain; Angelini, Rome, Italy) in the lateral canthus of anesthetized rabbits, the eye was pulled forward, and the optic nerve was exposed. It was then transected approximately 2 mm behind the sclera. After 120-day survival, retinas were fixed, and the completeness of the optic nerve transection was assessed. 

Semiquantitative RT-PCR Analysis

Total RNAs were extracted from six rabbit retinas in guanidine hydrochloride by using a kit (Simple Nucleic Acid Preparation; Invitrogen, Leek, The Netherlands). Cyclophilin B mRNA was used as an internal standard. ^{25 26} First-strand cDNA for PCR was generated from 1 μg of total RNA. Reverse transcription was performed according to Viollet et al. ²⁷ One tenth of the RT product was amplified in a total volume of 50 μl using 1.5 U Taq polymerase. ²⁷ Each sst mRNA was coamplified with cyclophilin B mRNA. Amplification was performed in an automatic thermocycler (Hybaid, Teddington, UK) beginning with a denaturation step at 94°C for 30 seconds, followed by 26 cycles of 94°C for 15 seconds, 60°C for 30 seconds, and 72°C for 15 seconds. The reaction was terminated with a 10-minute extension at 72°C. 

Nucleotide sequences of rabbit sst genes are not known. Therefore, for amplification, sense primers were chosen among mouse sequences. The sequence of the primers are given in Table 1 . The reverse primer sequence common to all ssts (COM 2) was chosen in the very conserved seventh transmembrane domain. ²⁷ Cyclophilin B primers were chosen according to the mouse sequence. ³² COM 2 and cyclophilin reverse primers were 5′-end labeled with [³²P] adenosine triphosphate (NEN, Boston, MA) by using T4 polynucleotide kinase (Boehringer–Mannheim, Mannheim, Germany). 

For each amplification, two types of controls were performed: an RT-PCR mixture with no reverse transcriptase to control genomic DNA contamination and a PCR mixture with no cDNA template, added to check for possible external contamination. 

A 10-μl sample of the PCR reaction was electrophoresed on a 8% polyacrylamide gel (Bio-Rad, Hercules, CA). After migration, the gel was dried and exposed to film (X-O-MAT; Eastman Kodak, Rochester, NY). Bands corresponding to the amplified products (characterized on the basis of their molecular weights) were cut and counted in aβ -scintillation counter (LKB, Wallac, Finland). 

The values, expressed as amount of recovered radioactivity, are relative to the cyclophilin B mRNA level (SRIF receptor mRNA/cyclophilin B mRNA). A semilogarithmic plot of recovered radioactivity versus cycle number showed an exponential increase in the PCR amplification between cycles 23 and 29 followed by a plateau. Curves of cyclophilin B and SRIF receptor mRNA amplification were parallel. All the conditions were chosen in the linear part of the coamplification. The cDNA from mouse pituitary, which express sst_{1 to 5}, was used as a positive control (Petrucci et al., unpublished result, 1999). 

Immunohistochemistry

A rabbit polyclonal antibody directed to the carboxyl-terminal sequence of human sst₁ ³³ was used (kindly provided by Lone Helboe, University of Copenhagen, Denmark). It has been used to study sst₁ distribution in the rat hypothalamus ⁶ and retina. ²³ In rat retinas, Western blot analysis showed no cross-reactivity of sst₁ antiserum with other ssts. ²³  

The dilutions of primary and of secondary antibodies reported herein were established in pilot experiments or were in accordance with previous studies. ^{12 34 35}  

Cryostat sections were incubated with the sst₁ antiserum (1:5000) in 0.1 M PB containing 1% Triton X-100 overnight at 4°C. Sections were washed in 0.1 M PB and incubated with indocarbocyanine (Cy3)-conjugated sheep anti-rabbit IgG (1:100; Sigma) and the slides coverslipped with mounting medium (Vectashield; Vector, Burlingame, CA). Alternatively, after incubation with the primary antibody, sections were incubated in biotinylated goat anti-rabbit IgG (1:50; Vector) and subsequently in an avidin-biotin-peroxidase mixture (ABC; Vectastain ABC Kit, Vector), both for 2 hours at room temperature. Sections were then treated with 0.05% 3,3′-diaminobenzidine tetrahydrochloride (DAB) and 0.03% hydrogen peroxide (H₂O₂), dehydrated, and coverslipped with Permount (BDH; Poole, Dorset, UK). Wholemount preparations were incubated with the sst₁ antiserum (1:2000) in 0.25 M Tris buffer containing 10% normal goat serum and 1% Triton X-100 for 3 to 4 days at 4°C. After incubation in biotinylated goat anti-rabbit IgG (1:50; Vector) for 2 days at 4°C, retinas were treated with ABC (2 days at 4°C). DAB and H₂O₂ treatment followed. Retinas were then mounted GCL up, dehydrated, and coverslipped with Permount. 

Specificity of the immune reaction was assessed by using either the preimmune serum or the primary antibody preadsorbed with 10⁻⁵ M sst₁ synthetic peptide overnight at 4°C in place of the primary antibody. Unspecific staining was observed in photoreceptor outer segments in cells that resembled microglial cells located in the IPL and in the GCL and in rare somata in the distal inner nuclear layer (INL) and in the outer plexiform layer (OPL). 

In double-labeling experiments, the sst₁ antiserum was used in conjunction with a rat monoclonal antibody directed to SRIF (1:50; Chemicon, Temecula, CA), or with mouse monoclonal antibodies directed to TH (1:100; Boehringer–Mannheim), parvalbumin (PV; 1:1000; Sigma), or GABA (1:200; Sigma). Cryostat sections were washed in 0.1 M PB and incubated in 0.1 M PB with 0.1% Triton X-100 containing both the sst₁ antiserum (1:5000) and one of the four specified antibodies overnight at 4°C. In addition, double-labeling experiments were performed using the TH monoclonal antibody in conjunction with a guinea pig polyclonal antiserum directed to GABA (NT108, 1:750, Eugene Tech, Allendale, NJ). After incubation with the primary antibodies, sections were incubated in the presence of the appropriate affinity-purified secondary IgGs conjugated with either fluorescein isothiocyanate (FITC; 1:50, Vector) or Cy3 (1:100, Sigma) for 2 hours at room temperature. Sections were then washed in 0.1 M PB and coverslipped with mounting medium (Vectashield; Vector). To eliminate the possibility of cross-reaction between primary and secondary antibodies in double-labeling experiments, control sections were made by omitting either of the primary antibodies. Control experiments were also performed to ensure that the primary antibodies did not cross-react when mixed together and that the secondary antibodies reacted only with the appropriate antigen-antibody complex. ³⁶ Finally, the preadsorbed sst₁ antiserum (see above) was used in conjunction with normal mouse serum in place of SRIF, TH, PV, or GABA monoclonal antibodies. 

Immunofluorescent materials were observed with both conventional fluorescence and confocal microscopy. 

Figure Preparation

Bright-field images were acquired at 300 dots per inch (DPI) using a digital imaging system (DC 100; Leica, Bensheim, Germany). Both bright-field images and electronic images from the confocal microscope were processed by computer (PhotoShop, ver. 5.0; Adobe, Mountain View, CA). 

Quantitative Analysis of Wholemount Preparations

Mean soma diameter and density (expressed as cells per square millimeter) of sst₁-IR cells were measured in three wholemounts originating from different animals to account for the variations in the number and the density of immunostained cells. ³⁷ Measurements of sst₁-IR cells were performed in 10,000-μm² fields at different sample locations regularly spaced throughout the retina. Thirty to 40 locations were analyzed in each retina. Details of the procedure for quantitative analysis have been published previously. ^{34 35 38} Quantitative data were obtained using a computer-assisted image analysis system that included a microscope (Axioplan; Carl Zeiss, Oberkochen, Germany) equipped with a color CCD video camera (JVC TK 1280E; SDS, Cambridge, MA), interfaced with a computer-assisted image analyzer. The software package for quantitative image analysis (Optimas 6.1; Media Cybernetics, Silver Spring, MD) included a routine for automatic counts of immunolabeled profiles and morphometric analysis. In the measurements, no correction for shrinkage was applied, because retinas were attached to the slides before dehydration. ³⁹ Estimations of absolute numbers of immunolabeled cells (cells per retina) were obtained by multiplying the mean cell density times the area of the retina. In the case of sst₁-IR putative displaced amacrine cells, the cell density was multiplied times the area of the ventral retina, because these cells were observed only in the ventral retina, as will be described later. Values of mean diameter, cell density, and absolute cell number are expressed as means ± SEM of the measurements in the three retinas analyzed. 

Results

RT-PCR Analysis

RT-PCR analysis using primers based on mouse sequences detected sst mRNAs in rabbit retinas. Positive controls (mouse pituitary) confirmed the specificity of the results. As shown in Figure 1A , RT-PCR on both mouse pituitary and rabbit retina samples yielded amplified products at 66 bp (sst₁), 78 bp (sst₂), 87 bp (sst₃), 69 bp (sst₄), and 363 bp (sst₅). As shown in Figure 1B , sst₁ mRNA appeared to be highly expressed in the rabbit retina, whereas moderate to low levels of sst₂, sst₃, and sst₄ mRNAs were observed. Finally, sst₅ mRNA was only slightly above the detection level. 

sst₁ Immunostaining Patterns

sst₁ immunolabeling was mostly observed to outline the plasma membrane of somata in the INL adjacent to the IPL (Figs. 2A , 2B) . sst₁-IR cells are also localized to the GCL. They were characterized by either small or large soma size (Fig. 2B) . The small-sized immunolabeled somata in the GCL were observed only in ventral retinal regions. The large-sized immunostained cells were sparsely distributed throughout the retina. In retinal sections treated with the avidin-biotin-peroxidase technique, sst₁-IR fibers appeared to be scattered in the IPL, and we could not associate them with specific IPL laminae. However, with confocal microscopy these fibers appeared to be confined to laminae 1 and 5 (according to Cajal ⁴⁰ ) of the IPL, with immunolabeling often observed also in lamina 3. These fibers originated from cell bodies located in the inner INL (Fig. 3A ) and in the GCL (Fig. 3B) . These observations are consistent with the expression of sst₁ in amacrines, displaced amacrines and, possibly, GCs. 

Both in wholemount preparations and in horizontal sections through the INL, two different types of amacrine cells could be detected, based on their soma size, in all retinal regions (Fig. 4A ). The first type was represented by cell bodies with mean diameter of 12.6 ± 2.3 μm and with sparse distribution throughout the retina (mean density 15.52 ± 3.82 cells/mm²). An estimate of their absolute number is 6200 ± 1200 cells/retina. These immunolabeled somata gave rise to two to three thick primary processes that arborized in varicose collaterals. The second type of sst₁-IR amacrine cell was characterized by oval to round soma shapes, by apparent absence of immunostained processes, and by mean soma diameters of 8.9 ± 2.8 μm. These cells were densely distributed in all retinal regions with a mean density of 112.24 ± 24.68 cells/mm² and a mean absolute number of 56,700 ± 800 cells/retina. 

In the GCL, sst₁-IR cells were likely to be both displaced amacrines and GCs (Fig. 4B) . With the assumption that immunolabeled somata with soma sizes similar to those of amacrine cells in the INL were displaced amacrines, whereas those with larger somata were GCs, we performed a separate analysis for these two distinct cell groups in the GCL. Putative sst₁-IR GCs had mean soma diameters of 18.3 ± 4.0 μm and a sparse distribution throughout the retina with a mean density of 2.71 ± 0.90 cells/mm² and a mean absolute number of 1000 ± 110 cells/retina. Bundles of sst₁-IR processes were observed in the GC axon layer. The dendritic arbors of putative sst₁-IR GCs were poorly immunolabeled, and only the proximal portion of primary processes could be detected (Fig. 4B) . After optic nerve transection, sst₁ immunolabeling in the GCL was restricted to small sized sst₁-IR somata, and sst₁-IR putative GCs were no longer observed. 

Putative sst₁-IR displaced amacrines had a mean soma diameter of 14.3 ± 3.8 μm, were characterized by ovoid or multipolar soma shapes, and were almost exclusively observed in ventral retinal regions. They originated thick primary processes that arborized into a meshwork of fine varicose fibers in laminae 1 and 5 of the IPL (Figs. 5A , 5B) distributed in all retinal regions. Their mean absolute number amounted to 1700 ± 120 cells/retina. Their densities, measured at three retinal eccentricities in the ventral retina, are shown in Figure 5C . These cells were sparsely distributed along the visual streak (4.57 ± 1.40 cells/mm²) and in midperipheral retina (5.14 ± 1.35 cells/mm²), whereas their highest density was observed at the ventral retinal edge (12.14 ± 2.91 cells/mm²). 

Double-Labeling Experiments

Both the quantitative and the morphologic features of the first population of sst₁-IR amacrine cells were in the range of those previously reported for the population of TH-IR amacrine cells in adult rabbit retinas. ^{34 41 42 43} In addition, the distribution of sst₁-IR processes in laminae 1, 3, and 5 of the IPL was similar to that of processes of TH-IR amacrine cells. Therefore, double-labeling experiments were performed using the sst₁ antiserum in conjunction with an antibody directed to TH. As shown in Figure 6 , TH-IR amacrines also expressed sst₁, both on their cell bodies and on their processes. All the TH-IR amacrine cells were also labeled with the sst₁ antiserum, and all the large-sized sst₁-IR amacrines were TH-IR. 

The size and the shapes of immunolabeled cell bodies belonging to the second population of sst₁-IR amacrine cells resembled those of AII amacrine cell somata, that can be identified with antibodies directed to PV. ³⁵ As shown in Figures 7A and 7C , double-label immunofluorescence using the sst₁ antiserum in conjunction with a PV antibody failed to show sst₁ expression in PV-IR amacrines. We also observed that sst₁-IR cells did not constitute a subset of GABAergic amacrine cells, because GABA-IR somata did not display sst₁ immunoreactivity (Figs. 7B 7D) . Because a colocalization of GABA and TH has been reported in rat retinas, ⁴⁴ we also tested the possible occurrence of GABA in TH-IR amacrine cells of rabbit retinas. In agreement with previous findings, ⁴⁵ our results confirm that, in rabbits, TH-IR amacrines did not contain detectable GABA immunoreactivity (not shown). 

The unique localization of displaced sst₁-IR amacrines to the GCL of the ventral retina, their sparse distribution, and the presence of sst₁-IR processes in laminae 1 and 5 of the IPL suggests these cells are SRIF-displaced amacrines. Double-labeling experiments with sst₁ antiserum in conjunction with an antibody directed to SRIF demonstrated coexpression of both SRIF and sst₁ immunoreactivities in the same population of displaced amacrine cells. All the sst₁-IR putative displaced amacrines also displayed SRIF immunoreactivity, and all the SRIF-IR somata were also labeled by sst₁ antibodies (Figs. 8A , 8B) . 

Discussion

The present results established the presence of all five sst mRNAs in the rabbit retina as well as expression and localization of sst₁ in populations of amacrines, displaced amacrines, and GCs. 

SRIF Receptor mRNAs

Semiquantitative RT-PCR showed the presence of all five sst mRNAs in the rabbit retina, indicating that rabbit sst mRNAs can be identified by RT-PCR methods using sst-specific primers based on mouse sequences. SRIF receptors have been cloned in human, rat, and mouse, and their sequences are highly conserved. ^{46 47} In particular, nucleotide sequence identity between human and rat ssts ranges from 97% for sst₁ to 81% for sst₅, ⁴⁶ and amino acid sequences display a greater than 90% identity between the same sst in different species. ¹ In addition, a recent RT-PCR analysis of sst mRNAs in guinea pig tissues has demonstrated close homology between the guinea pig and the human and rat sequences. ⁴⁸ In our experiments, the RT-PCR gels of rabbit retinas produced results identical with those of mouse pituitary, suggesting that rabbit sst mRNAs possess a sufficient homology with mouse sst mRNAs to be detected with RT-PCR. 

The presence of sst_2A mRNA has been recently demonstrated in the rat retina with RT-PCR. ²⁴ In addition, a previous RT-PCR study reported the presence of sst mRNAs in rat ocular tissues and identified sst₂ mRNA as the most abundantly expressed sst mRNA in the rat retina, followed by moderate levels of sst₁, sst₃, and sst₄ and by low levels of sst₅ mRNAs. ⁴⁹ In contrast, our observations of the rabbit retina indicated high levels of sst₁ mRNA, moderate levels of sst₂ mRNA, and low to very low levels of sst₃, sst₄, and sst₅ mRNAs. These differences may reflect different levels of expression of the various sst mRNAs in rat and rabbit retinas. However, the amount of sst mRNAs does not necessarily correlate with comparable levels of expressed proteins, because different messenger-to-protein ratios may result from differential posttranscriptional regulation of ssts. 

Localization of sst₁

The present study reports a detailed analysis of the distribution of sst₁ in the rabbit retina. The sst₁ antiserum used in this investigation has been characterized, ³³ and it has been used to localize sst₁ in the rat hypothalamus ⁶ and retina. ²³ It has also been used recently to confirm the decrease in sst₁ expression in the rat hypothalamus after sst₁ antisense treatment. ⁵ In our experiments, this antiserum specifically recognized sst₁ expressed by amacrines, displaced amacrines, and GCs in rabbit retinas. 

The observed sst₁ expression in TH-IR amacrines indicates that SRIF may participate in the control of dopaminergic cell functions (see below). TH-IR amacrine cells make extensive synaptic contacts with AII amacrine cells, ³⁵ thereby influencing the flow of visual information through the rod pathway. The presence of sst₁ in TH-IR amacrines may indicate an indirect influence of SRIF in the modulation of the rod pathway at the level of AII amacrine cells. Possible direct SRIF effects on AII amacrines would not be mediated by sst₁, because there was no absence of sst₁ expression in PV-IR amacrines, which include AII amacrine cells. ³⁵  

Regarding the numerous sst₁-expressing, non–TH-IR amacrines, quantitative studies of the rabbit retina confirmed that a large majority of all amacrine cells contain either glycine or GABA. ⁵⁰ Because sst₁-IR unidentified amacrines are not AII- or GABA-containing cells, they may constitute a subpopulation of glycinergic amacrine cells different from AII amacrines. Consistent with our findings, in the rat retina sst₁ is not expressed by glutamate decarboxylase-IR amacrines. ²³  

One discrete population of displaced amacrine cells in the rabbit retina is represented by SRIF displaced amacrines. ¹² These cells are characterized by the expression of sst₁ which is therefore likely to function as an autoreceptor. Similar observations have been recently reported in rat retinas. ²³ Further evidence for sst₁ as an autoreceptor is provided by its localization on SRIF neurons of the rat hypothalamus. ⁶  

sst₁-IR large-sized somata in the GCL are likely to be GCs, as indicated by the absence of sst₁-IR putative GCs after optic nerve transection. The observation of sst₁ expression in a small number of GCs is consistent with recent data on sst₁ expression in the rat retina. ²³ As shown by quantitative measurements, sst₁-IR GCs seem to comprise a very sparse group. However, the possibility exists that we have underestimated the population of sst₁-IR GCs, because the amount of immunolabeled fiber bundles in the GC axon layer would indicate a higher number of sst₁-IR GCs. 

Functional Implications

In the rabbit retina, exogenous SRIF enhances the signal-to-noise ratio and shifts the center-surround balance of GCs. ¹⁷ Although some GCs express sst₁, the observed SRIF effects on GC physiology are likely to result from complex functional interactions involving different ssts and different retinal cell types. For instance, SRIF has been reported to inhibit Ca²⁺ influx in rod bipolar cells of both goldfish ⁵¹ and rat ⁵² retinas. In addition, in the salamander retina, SRIF has been shown to reduce the Ca²⁺ current of rod photoreceptors but to increase that of cone photoreceptors. ⁵³ The sst_2A isoform is expressed by rod bipolar cells in rat ²⁴ and in rabbit ^{21 22} retinas and by rod and cone photoreceptors in the salamander retina. ⁵³ In addition, sst₂ has been reported to mediate the SRIF-induced inhibition of Ca²⁺ influx in cultured cells (for a review see Reference 8). These observations suggest that the reported SRIF-induced modulation of Ca²⁺ currents in bipolar cells and in rod and cone photoreceptors is mediated by the sst_2A isoform. Thus, sst_2A may mediate a possible SRIF control of glutamate (GLU) release in both photoreceptors and bipolar cells. 

SRIF effects mediated by sst₁ were confined to innermost retinal portions, and, as discussed, involved two different populations of amacrine cells (one of which is represented by TH-IR amacrines), the SRIF displaced amacrines, and some GCs. Regarding possible actions of SRIF onto dopaminergic cells, potent stimulatory effects of SRIF on DA release in the rat striatum have been reported ⁵⁴ ; however, this effect seems to be mediated by sst₂. ⁵⁵ SRIF control of dopaminergic amacrines is likely both in rat and in rabbit retinas, but rat and rabbit clearly differ in the specific sst that mediates control. Whereas in rat retinas dopaminergic amacrines express sst_2A but not sst₁, ²³ in rabbit retinas TH-IR amacrines seem not to express ²¹ or to express only partially ²² sst_2A, whereas they express sst₁. It is interesting to note that the effects of SRIF on rabbit GCs are similar to dark adaptation, ¹⁷ whereas DA, with an extracellular level that increases with increasing ambient light intensity, is involved in light adaptation. ⁵⁶ As previously suggested, ¹⁷ SRIF may be released in the dark and, therefore, it may act through sst₁ on rabbit dopaminergic amacrines as a dark signal, resulting in inhibition of DA release. This is consistent with observations of the chicken retina where amacrine cells coexpressing enkephalin, neurotensin, and SRIF are active in the dark and inhibit dopaminergic amacrines. ¹⁹ In contrast, in rat retinas, SRIF action on dopaminergic amacrines through sst_2A could stimulate DA release, as in the rat striatum. ⁵⁵ This hypothesis suggests different mechanisms for dark–light adaptation in rat and in rabbit retinas that may be related to the remarkable differences in diurnal activity patterns of rats and rabbits. Functional data on the effects of SRIF in the rat retina would shed some light on this issue. 

As a final consideration, we may observe that among rodent species there are differences in the molecular forms of SRIF that are expressed in the retina. Indeed, whereas in rat retinas only SRIF-14 is detected, ⁵⁷ rabbit retinas display a content of SRIF-28 that is approximately 10% of total SRIF ^{17 58} and, in mouse retinas, SRIF-binding sites display greater specificity for SRIF-28 than for SRIF-14. ⁵⁹ It is conceivable that differences in the organization of somatostatinergic systems are present in retinas of closely related species, and these differences may be reflected at various levels, including preferential expression of a particular molecular form of SRIF and species-specific patterns of expression of ssts. 

 Supported by Grant F06/PB/RS40% from the Italian Board of Education and Grant QLG3-1999-00908 from the European Community.

 Submitted for publication February 7, 2000; revised May 22, 2000; accepted May 31, 2000.

 Commercial relationships policy: N.

 

Table 1.

 

Primers Used for Amplification

Sense primers

sst1 (M1i) CAAGACGACGCCACCGTGAGCCA 28

sst2 (M2i) ATCAGCCCCACCCCAGCCCTGAA 28

sst3 (M3i) GTGTGCCCGCTGCCGGAGGAGCC 29

sst4 (M4i) ACCAGCCTCGATGCCACTGTCAA 30

sst5 (M5) GGGCTGGGGCACCTGCAACCTGA 27 31

Cyclophilin B CCATCGTGTCATCAAGGACTTCAT 32

Reverse primers

COM2 GGGGTTGGCACAGCTGTTG 27

Cyclophilin B TTGCCATCCAGCCAGGAGGTCT 32

Figure 1.

 

(A) RT-PCR products in homogenates from rabbit retinas (lanes a) and from mouse pituitary (lanes b, positive controls). Mouse subtype-specific primers yielded products with identical molecular weights, corresponding to sst₁ to sst₅, in rabbit retinas and in mouse pituitary (66 bp for sst₁, 78 bp for sst₂, 87 bp for sst₃, 69 bp for sst₄, and 363 bp for sst₅). Lane MW: Molecular weight markers. (B) Levels of SRIF receptor mRNAs in rabbit retinas as evaluated by RT-PCR semiquantitative analysis. Each histogram represents the mean ± SEM (n = 6) of SRIF receptor mRNA levels expressed as percentages of the cyclophilin B mRNA level.

Figure 2.

 

sst₁ immunostaining pattern in cryostat sections cut in a plane perpendicular to the vitreal surface and stained with the avidin-biotin peroxidase technique. (A, B) Most sst₁-IR somata were located in the INL adjacent to the IPL. In addition, some sst₁-IR cell bodies of either small or large size (B) were observed in the GCL. Dense immunolabeling is observed in the IPL. Light, unspecific immunolabeling is present in rare cell bodies in the distal INL and in the OPL. Scale bar, 20 μm.

Figure 3.

 

Confocal images (0.5 μm thick) through sections cut in a plane perpendicular to the vitreal surface showing sst₁ immunoreactivity as visualized with Cy3-conjugated secondary antibodies. Immunostained cell bodies were localized in the proximal INL (A) and in the GCL (B). Note the localization of sst₁ immunoreactivity to the plasma membrane of the cell bodies and to processes confined to laminae 1, 3, and 5 (A) or to laminae 1 and 5 of the IPL (B). Scale bar, 30 μm.

Figure 4.

 

Photomicrographs of retinal sections cut in a plane parallel to the vitreal surface and treated with the avidin-biotin-peroxidase technique. (A) shows large (arrow) and small (arrowhead) sst₁-IR somata in the INL. Processes of these cells are not visible because the section was cut through the INL with no IPL portions included. (B) Large and small immunolabeled somata in the GCL with immunolabeled proximal portions of primary processes and bundles of sst₁-IR fibers in the ganglion cell axon layer. Scale bar, 30 μm.

Figure 5.

 

Photomicrographs at two depths of a wholemounted retina treated with the avidin-biotin-peroxidase technique showing an immunolabeled displaced amacrine cell with processes that arborize in laminae 1 (A) and 5 (B) of the IPL and the cell body located in the GCL (B). Scale bar, 30 μm. (C) Densities (expressed as number of cells per square millimeter of retinal area) of sst₁-expressing displaced amacrine cells measured at different eccentricities of the ventral retina. Left: Three wholemounted retinas used for analysis. Open, shaded, and filled circles: Retinal locations in the visual streak, in the midperipheral retina and at the ventral retinal edge, respectively, where analysis was performed. Values in the graph are the means ± SEM of cell densities measured in the three retinas at corresponding eccentricities. The highest density of sst₁-containing displaced amacrine cells is observed at the ventral retinal edge.

Figure 6.

 

Confocal images showing optical sections (1 μm thick) through the soma (A, C) and through the proximal portion of process arborization (B, D) of an amacrine cell double labeled with antibodies directed to sst₁ and TH. sst₁ immunoreactivity was visualized with Cy3-conjugated secondary antibodies (A, B), whereas TH immunoreactivity was visualized with FITC-conjugated secondary antibodies (C, D). In the cell body sst₁ was on the plasma membrane, whereas TH was in the cytoplasm. Complete colocalization of sst₁ and TH immunoreactivities was observed both in the soma and in the processes. Scale bar, 25 μm.

Figure 7.

 

Confocal images (1 μm thick) through sections cut in a plane parallel to the vitreal surface and double labeled with sst₁ (A) and PV (C) antibodies or with sst₁ (B) and GABA (D) antibodies. sst₁ immunoreactivity was visualized with Cy3-conjugated secondary antibodies (A, B), whereas PV (C) and GABA (D) immunoreactivities were visualized with FITC-conjugated secondary antibodies. In these optical sections, sst₁ immunolabeling was visible in amacrine cell somata in the innermost INL and in varicose fibers located in the most distal portion of lamina 1 of the IPL (A, B). PV- and GABA-IR cells did not express sst₁ immunoreactivity. Scale bar, 25 μm.

Figure 8.

 

Confocal images showing sections cut in a plane parallel to the vitreal surface and double labeled with sst₁ (A) and SRIF (B) antibodies. In the ventral retina, all sst₁-IR displaced amacrines, as visualized with Cy3-conjugated secondary antibodies (A), also display SRIF immunoreactivity, as visualized with FITC-conjugated secondary antibodies (B), both in the cell body and in the processes. Scale bar, 20 μm.