Localization of Acetylcholine-Related Molecules in the Retina: Implication of the Communication from Photoreceptor to Retinal Pigment Epithelium

It has been long speculated that specific signals are transmitted from photoreceptors to the retinal pigment epithelium (RPE). However, such signals have not been identified. In this study, we examined the retinal expression and localization of acetylcholine-related molecules as putative candidates for these signals. Previous reports revealed that α7 nicotinic acetylcholine receptors (nAChRs) are present in the microvilli of RPE cells that envelope the tips of photoreceptor outer segments (OS). Secreted mammalian leukocyte antigen 6/urokinase-type plasminogen activator receptor-related protein-1 (SLURP-1) is a positive allosteric modulator of the α7 nAChR. Therefore, we first focused on the expression of SLURP-1. SLURP-1 mRNA was expressed in the outer nuclear layer, which is comprised of photoreceptor cell bodies. SLURP-1 immunoreactivity co-localized with rhodopsin and S-opsin in photoreceptor OS, while choline acetyltransferase (ChAT) and high affinity choline transporter (CHT-1) were also expressed in photoreceptor OS. Immunoelectron microscopy identified that the majority of SLURP-1 was localized to the plasma membranes of photoreceptor OS. These results provide evidence that SLURP-1 is synthesized in photoreceptor cell bodies and transported to photoreceptor OS, where SLURP-1 may also be secreted. Our findings suggest that photoreceptor OS communicate via neurotransmitters such as ACh and SLURP-1, while RPE cells might receive these signals through α7 nAChRs in their microvilli.


Introduction
Nicotinic acetylcholine receptors (nAChRs) are members of the family of neurotransmitter-gated ion channels. They are pentameric receptors comprised of a combination of protein subunits (alpha2-10 and beta2-4) which have been described in the mammalian nervous system [1]- [3]. Electrophysiological studies revealed that the action of acetylcholine (ACh) is mediated in part via nAChRs [4].
Recently, an understanding of the non-neuronal cholinergic system has begun to emerge. ACh is present in various cells, including epithelial, endothelial, and immune cells. Interestingly, nAChRs have been shown to play important roles in such nonexcitable cells [5]. Although nAChRs act as signal transducers in neurons [6], they have different roles in nonexcitable cells where they contribute to homeostasis [7]- [9]. Maneu et al. [10] investigated the expression of a7 nAChR in retinal pigment epithelium (RPE) cells, and reported that immunoreactivity of a7 nAChR was observed in RPE cells, especially in the apical surface of RPE cells adjacent to photoreceptor outer segments (OS). In contrast, photoreceptor OS showed no a7 nAChR immunoreactivity. This pattern suggests that ACh might be present in photoreceptor OS and act on RPE cells, although the physiological roles are unknown.
Secreted mammalian leukocyte antigen 6/urokinase-type plasminogen activator receptor-related protein-1 (SLURP-1) is a recently identified secreted protein; it consists of an 8843 Da nonglycosylated polypeptide with a distinct pattern of disulfide bonding among 10 cysteine residues that gives it a three-finger folded structure similar to that of snake or frog neurotoxin [11]. Mastrangeli et al. [12] reported that SLURP-1 is expressed in various cells including human skin, exocervix, gums, stomach and esophagus. In mice, expression is evident in skin, whole lung, trachea, esophagus, stomach and whole eye. Although the function of SLURP-1 has yet to be determined, studies are beginning to uncover its role. Chimienti et al. [13] demonstrated that SLURP-1 acts as an allosteric agonist to enhance ACh-evoked macroscopic currents in Xenopus oocytes expressing recombinant human nAChRs. Notably, Fischer et al. [14] discovered mutations in the gene encoding SLURP-1 in patients with Mal de Meleda, a rare autosomal recessive skin disorder characterized by transgressive palmoplantar keratoderma. Grando [7] reported that SLURP-1 acts as an epidermal modulator that is essential for keratinocyte homeostasis. Moreover, several other studies also reported that SLURP-1 acts as a positive modulator of a7 nAChR signaling [15]- [18].
The distribution of SLURP-1 in the eye has yet to be described until now. Distribution of SLURP-1 and ACh around the apical surface of RPE cells, consistent with the known pattern of a7 nAChR immunoreactivity, might indicate that SLURP-1 acts as a modulator of a7 nAChRs at the interface between the retina and RPE.
In this study, we used in situ hybridization, immunohistochemistry, and immunoelectron microscopy to identify the expression of SLURP-1 mRNA and protein in the murine retina. We also investigated the expression of choline acetyltransferase (ChAT) and high affinity choline transporter (CHT1), markers of AChsynthesizing cells [19]- [21], in order to assess the likelihood that SLURP-1 acts together with ACh and a7 nAChR at the interface between the retina and RPE.

Materials and Methods
All animal procedures were performed in accordance with the ARVO Statement for the use of Animals in Ophthalmic and Vision Research and the National Institutes of Health Guidance for Care and Use of Laboratory animals. The protocol was approved by the Gunma University Ethics Committee. We used more than 4 animals for each experiment to conclude the results.

Animals and Preparation of Eye Sections
Male C57BL/6 mice (Japan SLC, Hamamatsu, Shizuoka, Japan) 8-10 weeks of age were used for this study. Animals were deeply anesthetized with ether, then fixed by transcardiac perfusion with 0.1 M phosphate buffer (PB) containing 4% paraformaldehyde (pH 7.4, 25 ml, 300 ml/hr) for immunofluorescence microscopy and 4% paraformaldehyde/0.1% glutaraldehyde/0.1M PB for immunoelectron microscopy. For immunofluorescence microscopy, the whole eye was immediately enucleated, embedded in OCT compound (Sakura Finetek Japan Co, Ltd., Tokyo, Japan), and quickly frozen. Cryostat sections were then cut at a thickness of 16 mm with a Leica CM3050 cryostat and thaw-mounted on MAS-coated glass slides (Matsunami Glass, Osaka, Japan). For immunoelectron microscopy, the whole eye was cryoprotected with 30% sucrose/0.1M PB, then cut at 30 mm on a freezing microtome (SM2000R; Leica).
In Situ Hybridization of SLURP-1 mRNA in Retina

Primary antibodies
A summary of the primary antibodies used in this study is found in Table 1. We used rabbit polyclonal anti-SLURP-1 antibody

Immunohistochemistry
The eyes were collected either at around noon or at around midnight. The eye sections were preincubated in blocking buffer (PBS containing 2% bovine serum albumin and 10% normal donkey serum) with 0.4% Triton X-100 for 30 min at room temperature. The sections were then incubated with primary antibodies for 1 hr at room temperature. After they were washed in two changes of PBS, the sections were incubated with Alexa488-,  rhodamine-, or Alexa568-conjugated secondary antibodies for 1 hr at room temperature. TO-PRO-3 was used as a counterstain. The sections were then washed in two changes of PBS, mounted with Vectashield Mounting Medium (Vector Laboratories, Burlingame, CA), and visualized by confocal laser scanning microscopy. Adobe Photoshop was subsequently used to assemble images into the final figure format.

Organotypic tissue culture of rat retina
Organotypic tissue culture of rat retina was performed to examine SLURP-1 expression following artificial retinal detachment. Rats were used in this phase of the study as the viability of cultured rat retina is better than murine retina and for the convenience of isolating retina from the rat eye. Pieces of rat retina (about 1 cm 2 ) were placed ganglion cell-side-up on a 0.4 mm Millicell tissue culture insert (Millipore). Approximately 25 ml of Ames' medium (Sigma-Aldrich, St. Louis, MO) containing 1% horse serum, 1% N2 supplement, 100 U/ml penicillin, 100 U/ml streptomycin, and 0.3 mg/ml L-glutamine (Invitrogen, Grand Island, NY) was added to the dish such that the retina was in contact with the medium via the Millicell filter on the photoreceptor side, and the ganglion cell side was exposed to the incubator atmosphere (5% CO 2 , 35uC, humidified) [22], [23].

Results
We examined the retinal cell types which express SLURP-1 mRNA by in situ hybridization. Binding of the antisense probe revealed that SLURP-1 mRNA was expressed in the outer nuclear layer which is comprised of photoreceptor cell bodies ( Fig. 1 left  panel). In contrast to the antisense probe, no signal was detected with the sense probe ( Fig. 1 right panel), indicating that our antisense probe correctly recognized SLURP-1 mRNA. Next, we further examined the cellular distribution of SLURP-1 by immunohistochemistry. Interestingly, SLURP-1 was localized only within photoreceptor OS ( Fig. 2A and 2B). Histologically, cone OS are shorter than rod OS, the identities of which were confirmed by S-opsin and rhodopsin immunoreactivity, respectively. SLURP-1 in the OS co-localized with both rhodopsin and S-opsin ( Fig. 2A and 2B). Observation at higher magnification confirmed the positive signals within the photoreceptor OS; however, it also revealed that the cellular distribution of SLURP-1 was different from that of rhodopsin (Fig. 2C). We hypothesized that the difference in cellular distribution was based on the molecular properties of each protein as SLURP-1 is a secreted allosteric ligand of a7 nAChR whereas rhodopsin is a visual pigment. To further confirm this differential localization, we performed immunoelectron microscopic analysis utilizing an anti-SLURP-1 antibody. Consistent with immunofluorescence, preembedding immunogold electron microscopy for SLURP-1 heavily labeled the OS (Fig. 3A). The majority of metal particles in the OS were associated with the cell membranes (arrows in Fig. 3B). We also observed sparse labeling in intracellular sites of the OS and IS (arrowheads in Fig. 3B). When plotting the perpendicular distance of metal particles from the cell membranes (Fig. 3C), the OS, but not the IS, exhibited a peak distribution just inside the cell membrane. Moreover, some SLURP-1 molecules were detected outside of the plasma membranes of the OS. These results strongly indicate that SLURP-1 might be secreted from photoreceptor OS.
Several reports have confirmed that RPE expresses a7 nAChR [10], [24]. Thus, we speculated that acetylcholine might also be enriched in photoreceptor OS. To confirm this, we examined the immunoreactivity of ChAT (a synthetic enzyme of acetylcholine) and CHT-1 (a high affinity choline transporter required for synthesis of acetylcholine). Consistent with previous report [25], ChAT immunoreactivity was observed in photoreceptor OS in addition to the inner plexiform layer where it appeared as two stratified bands, the amacrine cell bodies in the inner nuclear layer and the ganglion cell layer (Fig. 4A). Consistent with our hypothesis, a similar expression pattern was observed for CHT-1 immunoreactivity (Fig. 4B) suggesting that photoreceptor OS contain acetylcholine.
This expression pattern raised the possibility that SLURP-1 might be secreted from photoreceptor OS, in agreement with previous studies on other cell types, as the majority of SLURP-1 was localized to the plasma membranes of photoreceptor OS, consistent with many reports that described SLURP-1 as a secreted molecule. To test this possibility, we designed a unique set of experiments to generate SLURP-1 protein diffusion utilizing organotypic tissue culture of the retina. We hypothesized that SLURP-1 might be highly secreted under organotypic tissue culture conditions, since the concentration of SLURP-1 would be very low and we continuously stir the medium (Fig. 5A). We speculated that this secretion could be detected as a reduction of SLURP-1 from photoreceptor OS without a reduction of SLURP-1 mRNA. As hypothesized, following preparation of organotypic tissue cultures, SLURP-1 immunoreactivity in photoreceptor OS was not detected while rhodopsin immunoreactivity was present and unchanged (Fig. 5B-D), indicating that the structure of the OS was maintained in vitro; however, SLURP-1 protein expression was eliminated. In contrast, SLURP-1 mRNA expression was maintained in the outer nuclear layer in vitro (Fig. 5D) as well as in vivo (Fig. 1), indicating that organotypic tissue culture did not affect SLURP-1 expression. Taken together, these results strongly suggest  that SLURP-1 might be secreted from photoreceptor OS as a signaling molecule.
We further tested whether SLURP-1 expression or localization in the retina was affected by dark adaptation. For this examination, we collected eye samples around midnight, whereas all of the above results were obtained from eyes collected around noon. SLURP-1 immunoreactivity in the OS did not change between day and night (Figs. 2 and 6). Next, we performed realtime RT-PCR of SLURP-1 with GAPDH as a control mRNA. The expression levels of SLURP-1 did not change between day and night (Figs. 6C and 6D). These results indicate that SLURP-1 expression does not change between day and night. To further elucidate the effect of dark adaption, we generated visually deprived eyes via surgical operation and examined SLURP-1 expression one week later. SLURP-1 expression and localization in the retina were not affected by the deprivation (Fig. S1). Collectively, our findings indicate that SLURP-1 expression is not altered via dark adaptation.

Discussion
We found the expression of SLURP-1 mRNA in photoreceptor cell bodies by in situ hybridization (Fig. 1), and determined by immunohistochemistry that SLURP-1 protein co-localized with rhodopsin and S-opsin in photoreceptor OS (Fig. 2). Immunoelectron microscopy revealed that SLURP-1 localized to the plasma  membrane of photoreceptor OS (Fig. 3). These results provide evidence that SLURP-1 is synthesized in photoreceptor cell bodies and transported to photoreceptor OS.
In previous reports, SLURP-1 was shown to act as a positive modulator of a7 nAChR signaling [7], [13]- [18]. In addition, nAChR was recognized to play important roles in several types of nonexcitable cells [5]. Some reports have confirmed that the RPE expresses a7 nAChR [10], [24]. In particular, Maneu V, et al. [10] reported that a7 nAChR is present in the apical surface of RPE cells, thought to be the microvilli of RPE cells which envelope the tip of photoreceptor OS. Histologically, cone OS are shorter than rod OS. Thus the elongated microvilli of RPE cells actually envelop the short cones, referred to as the ''cone sheath'' [26], [27]. In their results, the cone sheath also showed strong immunoreactivity for a7 nAChR. However, photoreceptor OS showed no immunoreactivity for a7 nAChR. These observations led us to consider the possibility that ACh from OS can affect the a7 nAChR in RPE cells, and that SLURP-1 from OS can modulate it. Therefore, we examined if OS contain ACh. It is well known that acetylcholine-synthesizing neurites in the retina are symmetrically distributed about the inner plexiform layer, one population of cholinergic amacrine cells has cell bodies in the inner nuclear layer and an equivalent population of displaced amacrine cells has cell bodies in the ganglion cell layer [28]- [32].
Here we found that ChAT-and CHT1-positive signals were detected also in photoreceptor OS (Fig. 4). Moreover, ChAT expression was identical to our previous report [25]. These results strongly indicate that ACh is localized to photoreceptor OS. In addition to SLURP-1, ACh acts as a neurotransmitter, suggesting that specific interactions might be formed between the OS and RPE cells through SLURP-1 and ACh signaling. It has been reported that RPE cells express muscarinic ACh receptors [33], [34], and the activation of them mediates phosphoinositide turnover [35]. These reports also support the existence of cellcell communications from photoreceptor OS to RPE cells. Presumably, ACh signals from the OS might be received by the RPE cells, and the muscarinic ACh pathway would promote phosphoinositide turnover.
Unfortunately, our study was insufficient to fully demonstrate that SLURP-1 is a secreted molecule from photoreceptor OS; however, our immunoelectron microscopic data strongly support this possibility (Fig. 3). We reasoned that the accumulation of SLURP-1 signals near plasma membranes was indicative of its release. Consistent with this idea, artificial organotypic tissue culture of the retina reduced SLURP-1 protein without any reduction in SLURP-1 mRNA (Fig. 5). Since our samples were obtained via artificial organotypic tissue cultures, the OS structures were not perfectly sustained. However, we previously reported that cultured retina contains OS although the OS structures are different from that observed in vivo [23]. Therefore, we believe that reductions of SLURP-1 protein did not occur as a result of a gross elimination of OS in the artificial organotypic tissue culture. We consider the inhibition of specific interactions between OS and RPE cells to be a toxic abnormal condition, therefore OS would need to release higher amounts of SLURP-1 to enhance signaling between OS and RPE cells. Alternatively, it is possible that organotypic tissue culture might cause the degradation of specific proteins such as SLURP-1, but not rhodopsin. The important conclusion to draw from our study, however, is that SLURP-1 protein levels were dynamically changed as a result of organotypic tissue culture, although we have yet to identify its physiological significance.
If ACh and SLURP-1 are secreted from OS, how are these molecules released? It has been reported that photoreceptors contain two types of vesicles [36]. One type is a synaptic vesicle, which is delivered to ribbon synapses and released onto bipolar interneurons. The other type is an opsin-containing vesicle, which is delivered to discs in the OS. Consistent with previous report, we failed to observe synaptic vesicles in the OS in our immunoelectron microscopic analysis (Fig. 3). It has been reported that ACh is transported into synaptic vesicles by a vesicular ACh transporter (VAChT) [37]. Arvidsson et al. [38] reported that only amacrine cells contain VAChTs in the retina. In contrast, photoreceptors do not possess any VAChTs. Taken together, typical synaptic release of ACh and SLURP-1 is unlikely in the OS. Therefore, we should consider alternative hypothesis for the relevant secretion pattern. For example, reverse ACh transport (from intracellular to extracellular) might be involved in the mechanism.
Diverse molecular interactions between photoreceptor OS and RPE cells have been reported in relation to visual cycle, phagocytosis of OS by RPE, and neuroprotection, to name a few [39]. It has also been reported that dysfunctional interactions could lead to diseases such as retinitis pigmentosa or age-related macular degeneration. However, the mechanisms of molecular interactions between OS and RPE cells have not been fully elucidated. Our findings indicate ACh and SLURP-1 as potential new candidates for the signaling molecules bridging OS to RPE cells. This could represent a significant breakthrough in identifying the causes of retinal or RPE diseases. It has been reported that SLURP-1 promotes cell survival in periodontal ligament fibroblasts through the activation of anti-apoptotic signal phosphatidylinositol 3-kinase [40]. Thus, ACh and SLURP-1 from OS might play an important role in the maintenance of RPE by protecting RPE cells from apoptosis. Taken together, SLURP-1 might represent a novel signal messenger in the retina. Some case reports have described the development of ocular disorders such as macular yellow deposits or congenital cataract in Mal de Meleda patients [41], [42]. Therefore, we strongly believe that ACh signaling is important for the maintenance of eye function. Further study is required to elucidate the mechanistic relationship between photoreceptor OS and RPE. Figure S1 Expression of SLURP-1 in photoreceptor outer segments of deprived eye. SLURP-1 expression in photoreceptor outer segments (OS) did not change following seven days of deprivation. A. Confocal image showing SLURP-1 (green) and rhodopsin (red) immunoreactivity in mouse retina. SLURP-1 is localized in the photoreceptor OS and co-localizes with rhodopsin. B. Confocal image showing SLURP-1 (green) and S-opsin (red) immunoreactivity in mouse retina. SLURP-1 is localized in the photoreceptor OS and co-localizes with S-opsin. DNA was counterstained with TO-PRO-3 (blue). Scale bars: 20 mm (A, B); ONL, outer nuclear layer; INL, inner nuclear layer; RPE, retinal pigment epithelium; IS, inner segments. (TIF)