Isolation of Photoreceptors in the Cultured Full-Thickness Fetal Rat Retina

Abstract

purpose. To create a retina consisting mainly of photoreceptors for future use as donor tissue in retinal transplantation.

methods. Fetal full-thickness neuroretinas from Sprague-Dawley embryonic day (E) 17 or E20 rats were placed in culture for 7 or 14 days. Explants and age-matched control retinas were examined by light microscopy and with a panel of immunohistochemical markers labeling all seven of the major retinal cell types.

results. E17 and E20 control retinas displayed vimentin-labeled Müller cells, NF160-labeled ganglion cells, and synaptic vesicles labeled with synaptophysin. The remaining cell types were found in control specimens of postnatal age 2 days and older. After 7 or 14 days in culture, all explants were significantly thinner than their aged-matched controls and displayed multiple rows of cells organized in a single layer. Within this layer, they contained rhodopsin-labeled rod photoreceptors, presynaptic vesicles, and vertically arranged Müller cells. Transducin-labeled cone photoreceptors were found in all but the youngest explants. Scattered PKC-labeled rod bipolar cells and calbindin-labeled horizontal cells were found in the inner part of most explants, whereas β-III-tubulin–labeled ganglion cells and parvalbumin-labeled amacrine cells were seen only sporadically. No NF160-labeled ganglion cells were found.

conclusions. Fetal full-thickness rat retina in vitro develops into a retina consisting of predominantly synapse-containing cone and rod photoreceptors embedded in a scaffold of well-organized Müller cells. These explant retina characteristics are well adapted for use as donor tissue in future retinal transplantation experiments.

Retinal transplantation experiments aimed at replacing diseased photoreceptors affected by degenerative disease have been ongoing for more than two decades. ¹ During this time, a multitude of protocols has been explored making use of donor tissue from various sources in different forms. Our laboratory has focused on neuroretinal full-thickness transplants that, in several normal and degenerative animal models, have been shown to develop morphologically normal photoreceptors that survive for extended periods without immune suppression. ^{2 3 4} However, transmission of visual information from the transplant to the host central nervous system has not been attained because of lack of integration of graft-derived neurons with the host retina. One of the main reasons for the poor integration is the development and persistence of inner retinal layers in the full-thickness graft that physically hamper neuronal contact between transplanted photoreceptors and host bipolar cells. ^{2 3 4 5}  

One way to enhance the chance for neuronal integration is to separate photoreceptors from the inner retinal layers before transplantation. Removal of inner retinal layers has been explored by enzymatic disruption and vibratome or laser ablation. ^{6 7 8} Unfortunately, isolated photoreceptor transplants derived by mechanical or chemical disruption techniques have not yet been shown to achieve robust organization, survival, or integration after transplantation, possibly because of pretransplantation cell death and Müller cell trauma. ^{6 9 10 11 12}  

Another way to isolate photoreceptors without mechanical or chemical disruption of cell-to-cell contacts may be to manipulate the developing retina. In the mammalian retina, the different classes of cells are generated from a pool of multipotent progenitor cells, from which the retinal subtypes develop in a precise chronological order. ^{13 14} Theoretically, the relative amount of photoreceptors in the retina can be enhanced by influencing the commitment of the multipotent progenitor cell toward a photoreceptor fate or by eliminating the cells committed to inner retinal neuronal subtypes. Manipulation of cell differentiation in the retina has been shown in culture systems in which factors influencing the developing tissue can be controlled. ^{15 16 17 18} The rat retina has a comparatively extended postnatal period of retinal cell-type generation, and the postnatal rat retina in vitro is one of the most studied models of retinal development. ^{14 19 20} In contrast, studies of fetal rat retinal development in vitro are scarce. ²¹ Given that retinal cell generation and differentiation in the rat begin on embryonic day (E) 10, the gestational period lasts 22 days, and 4 of 7 classes of retinal cells are generated by birth, manipulation of retinal development during the fetal period may also be possible. In the present experiment, therefore, we explored fetal rat retinas maintained under standard culturing conditions, with special attention paid to inner and outer retinal development. We used a panel of immunohistochemical markers to study the development and survival of the seven major retinal subclasses with the ultimate goal of producing a retina consisting of well-organized photoreceptors for the purpose of future retinal transplantation. 

Materials and Methods

Culture Procedure

Time pregnant Sprague-Dawley rats were obtained from three different sources (Scanbur, Sweden; Tactonic, Denmark; and Harlan Netherlands B.V., The Netherlands). The pregnant rats were killed with CO₂. After caesarean section, embryos from stage E17 (n = 14) or E20 (n = 14) were collected and put in ice-cold CO₂-independent medium (Gibco, Paisley, UK). Both eyes were removed by enucleation and immediately immersed in ice-cold CO₂-independent medium. Neuroretinas were carefully dissected free from the retinal pigment epithelium (RPE) and hyaloid vascular system with fine forceps. The optic nerve was thereafter cut with microscissors, and the neuroretinas were explanted on culture plate inserts (Millicell-HA 0.45-μm; Millipore, Billerica, ME) with the photoreceptor layer toward the membrane. All neuroretinas were put in culture within 180 minutes. The explants were cultured in 2 mL Dulbecco’s modified Eagle’s medium (DMEM)/F12 medium–l-glutamine (Gibco) supplemented with 10% fetal calf serum. 

A cocktail containing 2 mM l-glutamine, 100 U/mL penicillin, and 100 ng/mL streptomycin (Sigma-Aldrich, St Louis, MO) was added, and the retinas were maintained at 37°C with 95% humidity and 5% CO₂. After the first 3 days, half the culture medium was exchanged; this procedure was then repeated every second day. Specimens were kept under culture conditions for 7 or 14 days. 

All proceedings and animal treatment were in accordance with the guidelines and requirements of the Government Committee on Animal Experimentation at Lund University and with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Histology

For histologic examination, the explants were fixed for 1 hour in 4% formalin, pH 7.3, in 0.1 M Sørensen phosphate buffer (PB). After fixation, the specimens were washed with 0.1 M Sørensen PB and were washed again using the same solution containing sucrose of increasing concentrations (5%–25%). Specimens were sectioned at 12 μm on a cryostat, and every tenth slide was stained with hematoxylin and eosin according to standard procedures. 

For immunohistochemical staining, sections were incubated at room temperature with phosphate-buffered saline (PBS) containing 0.25% Triton X-100 and 1% bovine serum albumin for 30 minutes. This was followed by incubation of the slides overnight with primary antibodies (Table 1) . After incubation, the slides were rinsed in PBS and incubated with Texas red-conjugated secondary antibodies (Jackson ImmunoResearch, West Grove, PA) for 50 minutes in room temperature, rinsed again, and mounted in custom-made mounting media (Vectashield; Vector Laboratories, Burlingame, CA) containing DAPI (4′,6-diaminidin-2-phenylindoldihydrochloride). 

For rhodopsin-vimentin-DAPI triple labeling, sections were first incubated with the rhodopsin antibody for 18 to 20 hours. They were then rinsed in PBS-Triton and incubated with anti-mouse Texas red. After another thorough rinsing in PBS-Triton, vimentin immunolabeling was performed in a similar manner but with anti-mouse fluorescein isothiocyanate (FITC; Sigma, St. Louis, MO) as a secondary antibody. Because the rhodopsin and vimentin antibodies were made in the same species (mouse), this secondary antibody now recognized all mouse antigen in the tissue; consequently, the FITC fluorescence was present at antirhodopsin- and antivimentin-positive sites. To differentiate vimentin and rhodopsin labeling, separate images of the tissue activity of the fluorophores, together with a third image of the DAPI labeling derived from the mounting media, were superimposed (Photoshop; Adobe, San Jose, CA). In this composite image, rhodopsin-positive sites were yellow (green + red), vimentin-positive sites were green, and the remaining cells were blue from DAPI, which labeled all retinal neuronal nuclei. 

For rhodopsin and synaptophysin double labeling, the tissue was incubated with rhodopsin and synaptophysin antibodies for 18 to 20 hours and rinsed in PBS-Triton. Then the tissue was incubated for 45 minutes in darkness with a mixture of the secondary antibodies conjugated with two fluorophores, anti-rabbit FITC (Southern Biotechnology, Birmingham, AL), and anti-mouse Texas red. The dilution of each secondary antibody was 1:200. 

Neuroretinas from E17, E20, postnatal (P) day 2, P5, P9, and P12 Sprague-Dawley rats (n = 4 in each group) were used to compare cultured specimens with the corresponding in vivo counterparts. Sections from adult Sprague-Dawley rat retinas were used as positive controls. Negative controls were obtained by performing the complete labeling procedure without primary antibody on retinal sections from adult Sprague-Dawley rats. Photographs were obtained with a digital camera system (Olympus, Tokyo, Japan). When comparing immunolabeled sections, specimens were always derived from the same labeling batch, and were photographed in one session using a fixed exposure time, aperture, and ISO (equivalent) speed. Photographs were adjusted digitally for brightness and contrast. 

Results

Control Retinas

E17 control retinas displayed a neuroblastic cell layer (NBL) consisting of multiple rows of undifferentiated cells with a marginal zone in its innermost part (Fig. 1A) . In E20 specimens, an NBL and a ganglion cell layer with several rows of cells were found (Fig. 1B) . Immunohistochemistry (Table 1)revealed NF160-labeled ganglion cells with fibers in the nerve fiber layer in E17 and E20 specimens (Figs. 1C 1D) . Weakly labeled, synaptophysin-positive processes were found in both groups in the outer part of the NBL and in the E20 specimens in the thin inner plexiform layer (Figs. 1E 1F) . Vimentin-labeled Müller cell fibers were found throughout the retina in both groups (Figs. 1G 1H) . NF160, synaptophysin, vimentin, and the remaining antibodies directed against cone transducin, rhodopsin, calbindin, PKC, β-III-tubulin, and parvalbumin labeled their respective cells in P2, P5, P9, P12, and adult specimens (Table 1) . 

E17 Explants

In most cases, the E17 specimens kept 7 days in vitro (DIV; n = 14) were double folded. The part adjacent to the membrane displayed a laminated morphology without rosettes, whereas the remaining part contained multiple rosettes. The inner retinal layer differentiation seen in corresponding P2 controls was not found, and the explanted retinas in hematoxylin and eosin slides displayed an NBL with a comparatively loose arrangement of cells (Figs. 2A 2B) . Strongly labeled, rhodopsin-positive rod photoreceptors were found in the entire explant except for the innermost part (Fig. 2C) . These cells often extended processes ending at the innermost margin of the explant. A few scattered cells within the NBL were found to be calbindin positive (Fig. 2D) . Synaptophysin-labeled processes were present in almost the entire NBL, but no organization in plexiform layers could be seen (Fig. 2E) . Vimentin-labeled vertically arranged Müller cell fibers were present throughout the explants (Fig. 2F) . No labeling for cone transducin, parvalbumin, PKC, NF160, or β-III-tubulin was found (data not shown). 

E17 specimens kept 14 DIV (n = 14) in most cases were double folded and displayed a laminated morphology without rosettes in the part adjacent to the membrane. They were significantly thinner than their corresponding P9 controls and consisted of an outer nuclear layer (ONL) without any other subdivision in retinal layers (Figs. 3A 3B) . Well-labeled cone transducin-positive cells were seen throughout the specimens (Fig. 3C) . Morphologically, these cells were similar to cone photoreceptors: elongated cell body, inner segment, and axon with terminals in the innermost part. Rhodopsin-labeled rods were seen throughout the explants, with intense labeling in a narrow band at the outermost part of the specimens corresponding to the inner segments (Fig. 3D) . PKC-labeled bipolar cells were abundant in the innermost part of the explants (Fig. 3E) . These cells did not display the normal vertical arrangement but were randomly arranged horizontally. Scattered parvalbumin- and calbindin-labeled cells were seen in the inner part of the explants (Figs. 3F 3G) . As in 7DIV counterparts, synaptophysin-labeled processes without plexiform layering were seen (Fig. 3H) . Vimentin-labeled Müller cells were seen throughout the explants, with intense labeling in the vitread end-feet region (Fig. 3I) . No ganglion cells labeled with NF160 or β-III-tubulin were found (data not shown). 

E20 Explants

Most E20 explants kept 7 DIV (n = 14) were flat and displayed a laminated structure without rosettes. They were significantly thinner than their corresponding P5 controls and displayed an ONL without no other distinguishable layers (Figs. 4A 4B) . Weakly labeled cone transducin-positive cells were seen throughout the specimens, and rhodopsin-labeled rod photoreceptors were found in all but the innermost part of the explant (Figs. 4C 4D) . Occasional PKC-labeled rod bipolar cells, β-III-tubulin–labeled ganglion cells, and calbindin-positive horizontal cells were seen at the innermost margin of the explants (Figs. 4E 4F 4G) . Strongly labeled synaptophysin-positive cells and vertically arranged vimentin-labeled Müller cells were seen throughout the explants (Figs. 4H 4I) . No labeling for parvalbumin or NF160 was found (data not shown). 

Most E20 specimens kept 14 DIV (n = 14) were double folded. The part adjacent to the membrane displayed a laminated structure without rosettes but had a more degenerated appearance than E20 7 DIV and E17 explants. The retinas were significantly thinner than their corresponding P12 controls and in hematoxylin and eosin slides displayed an ONL with a rudimentary outer plexiform layer and a few singular cell bodies in the inner nuclear layer (Figs. 5A 5B) . Transducin-labeled cone photoreceptors with a varying degree of labeling intensity were present throughout the ONL (Fig. 5C) . Rhodopsin-labeled rod photoreceptors were seen throughout the specimens, with strong labeling intensity in the inner segments (Fig. 5D) . Few scattered PKC- and calbindin-labeled cells were seen at the innermost margin of the explants (Figs. 5E 5F) . Synaptophysin-labeled cells were seen throughout the explants without any apparent organization (Fig. 5G) . Vimentin-labeled Müller cell fibers were seen throughout the explants, but these were fewer than E20 7DIV and E17 explants and were more weakly labeled except in the vitread end-feet region (Fig. 5H) . No labeling for parvalbumin, NF160, or β-III-tubulin was found (data not shown). 

To determine the origin of synaptic vesicles within the explants, a double-labeling experiment including antibodies against rhodopsin and synaptophysin was performed (Fig. 6In all explant groups, the two antibodies were colocalized to a large extent. 

Explant Structure Analysis

To further explore the overall structure of the explants, triple labeling including rhodopsin, vimentin, and DAPI was performed (Fig. 7) . In the adult control (Fig. 7A) , rhodopsin labeling was located primarily to the outer segments of the rod photoreceptors, with minimal labeling in the ONL and the outer plexiform layer. Vimentin-labeled Müller cells displayed the normal vertical arrangement, with strongest intensity in the innermost part of the retina. DAPI-labeled neuronal nuclei were seen in all three nucleated layers. In E17 7DIV explants, rhodopsin-labeled rod photoreceptors were observed in the outer half of the specimens, and vimentin labeling of Müller cells was observed primarily in the inner half (Fig. 7B) . DAPI-labeled cells were seen throughout the explant. In E17 14DIV explants and in both groups of E20 specimens, rhodopsin-labeled cells were present in each whole specimen except for in the innermost part, where vimentin-labeled Müller cell fibers were observed (Figs. 7C 7D 7E) . DAPI-labeled cells were seen throughout the explants, coinciding with rhodopsin labeling in all but the innermost part of the specimens. 

Discussion

In this study, we examined developmental characteristics of the fetal full-thickness rat neuroretina in vitro. Our findings indicated that several of the immunohistochemical markers of the seven principal retinal cell types were expressed in vitro but that inner retinal development was severely compromised by the culturing procedure. Cultured E17 and E20 explants developed into a retina consisting of rod and cone photoreceptors, synaptic vesicles, and well-organized Müller cells, whereas inner retinal neurons were almost absent. The literature regarding fetal explanted rat retina is scarce. ²¹ To our knowledge, the present study represents the first in which all major retinal cell types have been explored under standard culturing conditions. 

Neurogenesis in the rat retina has been well described in vivo, whereas detailed information regarding the phenotypic differentiation characteristics of the various cell types is not as readily available. In the early phase of neurogenesis, ganglion cells, cone photoreceptors, and horizontal cells are born during a period lasting from E10 to E18. Amacrine cells are born from E12 to E20, whereas the late phase encompasses the birth of rod photoreceptors, Müller cells, and bipolar cells at E19 to P4. Phenotypic differentiation follows neurogenesis, with ganglion cells at one end of the spectrum displaying axons invading the optic nerve at E15 and rod photoreceptors at the other and do not show complete differentiation until several weeks after birth. ^{22 23} Discrepancies do exist; for instance, vimentin-positive Müller cells are already present at E14. ²⁴ In our culture system, most retinal cell types appeared and expressed specific markers indicating birth and differentiation largely in accordance with the in vivo timetable. Two exceptions included ganglion cells, which were absent, and cone photoreceptors, which expressed transducin later in vitro than in vivo. To our knowledge, prenatal expression of other immunohistochemical cone markers, such as cone opsin, has not been reported in vivo or in vitro, which is consistent with our own experience (data not shown). However, Liljekvist-Larsson et al. ²⁵ reported atypical cone opsin expression in postnatal cultures, supporting the notion that cone development in vitro may be delayed. 

Another more fundamental difference between the in vivo and in vitro situation is apparent when inner and outer retinal development is compared. The poorly developed inner retina in our fetal explants stood in contrast with previously published reports on postnatal rat retinal explants in which inner retinal layers are well organized and contain most of the inner retinal cells. ^{19 20} Postnatal rat retinal development, in vivo and in vitro, has been studied extensively because of a relatively long period of development after birth. One of the most prominent effects of the culturing procedure is rapid ganglion cell loss caused by optic nerve transection. ²⁶ In the postnatal rat retina, axotomy-induced loss of ganglion cells is followed by the death of other inner retinal cell types. ^{16 27} In the present study, we found almost no surviving ganglion cells in the cultured specimens, and it is not unlikely that our fetal culture system induced the same retrograde degeneration found in postnatal counterparts, which in turn may explain the poor development of the remaining inner retinal subtypes. 

In contrast to inner retinal neuronal development, cone and rod photoreceptors appear not only to follow their in vivo developmental timetable, they also survive well in the fetal culture system. In the normal rat, cone photoreceptor generation is completed by E16, but several differentiation characteristics coincide with rod photoreceptors that are born much later. ^{14 28} We found rhodopsin-labeled rods in all explants corresponding to in vivo expression from P2 and onward. These cells in 14 DIV specimens displayed an accumulation of rhodopsin in inner segments, but no outer segments developed, probably because of the absence of retinal pigment epithelium in the culture system. ^{20 29} Transducin-labeled cones were found in explants corresponding to P5 and on, with the strongest labeling and the strictest organization found in E20 7 DIV specimens. In spite of a severely reduced number of inner retinal neurons, synaptophysin labeling was seen in concert with rhodopsin and transducin labeling throughout the explants, indicating a massive presence of photoreceptor-derived synaptic vesicles. We also found that rod and cone photoreceptors could extend axons to the innermost margin of the explants, mimicking events during normal development when these axons can project far into the inner retina. ³⁰ These findings are of the utmost importance for future transplantation during which graft-host connections will be dependent on the ability of transplanted photoreceptors to sprout fibers into the host retina. The pattern of synaptophysin-labeled photoreceptor synapses diffusely distributed in the ONL without any apparent organization is also seen in the developing rat retina in which the outer plexiform layer is gradually organized from P5 to P24. ³¹ The lack of proper postsynaptic targets prevents such organization in the cultured retinal specimens, but in transplantation, such targets are available in the host retina and may help to reestablish proper neural circuitry in the new graft-host retina. 

Another cell type that has bearing on retinal transplantation is the Müller cell, which is essential for the formation of the proper three-dimensional cytoarchitecture of the retina and provides metabolic and structural support for growing neurites. ^{32 33} We found that Müller cells in all explant groups displayed the normal vertical arrangement, supporting the notion that, in spite of the almost absent inner retina, the remaining ONL remains well organized. In 14 DIV specimens, however, Müller cell fibers in the inner part of the explant became strongly vimentin positive. This finding, which may imply a gliotic process, suggests that for transplantation purposes, fetal retinas cultured for 14 days may not be ideal. Another finding that will be important to monitor before transplantation is the double folding of the explants, which is not desirable because it hampers proper graft-host neuronal connections. Double folding and gliosis of the explanted retina was not present in E20 7 DIV specimens, making this group of explants most suitable for transplantation. 

To conclude, in striving to optimize donor tissue for retinal transplantation, we found that the fetal full-thickness rat retina in vitro develops into a retina consisting predominantly of synapses containing cone and rod photoreceptors embedded in a scaffold of well-organized Müller cells. These characteristics are theoretically attractive, and future experiments will be directed toward practical application in a retinal transplantation paradigm. 

 Supported by the Faculty of Medicine, University of Lund; the Swedish Research Council; the Princess Margaretas Foundation for Blind Children; and the Torsten and Ragnar Söderberg Foundation.

 Submitted for publication June 5, 2008; revised July 5, August 27, and September 16, 2008; accepted December 22, 2008.

 Disclosure: F. Ghosh, None; K. Arnér, None; K. Engelsberg, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Specifications and Developmental Time of Expression of Immunohistochemical Markers in In Vivo Control Rats

Antigen	Antibody Name	Target	Species	Dilution	Source	Developmental Expression Time
β-III-tubulin	Anti-β-tubulin isotope III	Ganglion cells	Mouse monoclonal	1:100	Sigma, St Louis, MO	From PN2
Cone transducin	Anti-G protein Gyc subunit	Photoreceptor (cones)	Rabbit polyclonal	1:1000	Cytosignal, Irvine, CA	From PN2
Neurofilament 160 kDa (NF160)	Anti-neurofilament 160 clone NN18	Ganglion cells	Mouse monoclonal	1:500	Sigma, St Louis, MO	From E17
PKC	Phospho-PKC (pan)	Rod bipolar cells	Rabbit polyclonal	1:200	Cell Signaling, Beverly, MA	From PN2
Parvalbumin	Mouse anti-parvalbumin	All amacrine cells	Mouse monoclonal	1:1000	Sigma, St Louis, MO	From PN2
Calbindin	Anti-calbindin-D-28K	Horizontal cells	Mouse monoclonal	1:200	Sigma, St Louis, MO	From PN2
Synaptophysin	Rabbit anti-human synaptophysin-protein	Presynaptic vesicles	Rabbit polyclonal	1:100	DakoCytomation, Glostrup, Denmark	From E17
Rhodopsin	Rho4D2	Photoreceptor (rod)	Mouse monoclonal	1:100	Kind gift of Robert S. Molday, Vancouver, Canada	From PN2
Vimentin	Mouse anti-vimentin	Müller cells	Mouse monoclonal	1:500	Chemicon International, Temecula, CA	From E17

Figure 1.

 

E17 and E20 control retinas. VITR, vitreal (inner) aspect; SCL, scleral (outer) aspect. (A, B) Hematoxylin and eosin staining. (C–H) Immunohistochemistry. (A) The E17 neuroretina consists of an NBL composed of immature cells and a thin marginal zone (MZ) along the innermost border. (B) In the E20 neuroretina, the NBL constitutes the major part of the retina, and a ganglion cell layer (GCL) with multiple rows of cells is present. Ganglion cells labeled with the NF160 antibody can be seen in the inner part of the NBL in the E17 retina (C), and in the GCL in the E20 specimen (D). In the E17 (E) and E20 (F) retina, weakly labeled synaptophysin-positive neuronal processes are present in the outer part of the NBL. In the E20 specimen, a thin inner plexiform layer (IPL) is present. (G, H) Vimentin-labeled Müller cell fibers are seen in both specimens. Scale bar, 50 μm.

Figure 2.

 

(A) P2 control retina displays a NBL, an IPL, and a GCL. (B–F) E17 neuroretina 7 days in vitro. (A, B) Hematoxylin and eosin staining. (C–F), immunohistochemistry. The cultured retina is double folded and displays a NBL but no inner retinal layers (B). (C) Strongly labeled, rhodopsin-positive rod photoreceptors are seen in the entire explant except for at the innermost part. Arrows: Rhodopsin-labeled processes ending at the innermost margin of the explant. (D) A few scattered cells within the NBL are weakly labeled with the calbindin antibody. (E) Synaptophysin-labeled processes are present in almost the entire NBL, but no organization is seen in plexiform layers. (F) Vimentin-labeled vertically arranged Müller cell fibers were present throughout the cultured retina. (G) Negative control. Scale bar, 50 μm.

Figure 3.

 

(A) P9 control compared with (B–I) E17 neuroretina 14 days in vitro. (A, B) Hematoxylin and eosin staining. (C–I) Immunohistochemistry. (A) P9 control retina displays most normal retinal layers. (B) Cultured neuroretina is significantly thinner than P9 controls and displays an ONL without any other distinguishable layers. (C) Well-labeled cone transducin–positive cones span the entire specimen with terminal ending in the innermost part. (D, arrows) Rhodopsin-labeled rods are seen throughout the explant, with strong labeling intensity in a narrow band at the outermost part of the specimens corresponding to inner segments. (E) A multitude of PKC-labeled rod bipolar cells with a random horizontal organization is seen in the inner part of the explant. (F, G) Scattered amacrine and horizontal cells in the inner part of the explants are labeled by parvalbumin and calbindin. (H) Synaptophysin-labeled processes are seen throughout the explant without any apparent organization in plexiform layers. (I) Vimentin-labeled Müller cell fibers are seen throughout the specimen, with strong labeling intensity in the vitread end-feet region (arrows). (J) Negative control. Scale bar, 50 μm.

Figure 4.

 

(A) P5 control retina displays most normal retinal layers. (B–I) E20 neuroretina 7 days in vitro. (A, B) Hematoxylin and eosin staining. (C–I) Immunohistochemistry. (B) Cultured neuroretina is significantly thinner and displays an ONL without any other distinguishable layers. (C) Weakly labeled cone transducin–positive cones are seen throughout the specimen, and (D) rhodopsin-labeled rods are seen in all but the innermost part of the explant. (E) Well-labeled rod bipolar cells (PKC), (F) ganglion cells (β-III-tubulin), and (G) horizontal cells (calbindin) are seen at the innermost margin of the explants. (H, I) Synaptophysin-positive processes and vertically arranged vimentin-positive Müller cells are seen throughout the explant. (J) Negative control. Scale bar, 50 μm.

Figure 5.

 

(A) P12 control retina displays most normal retinal layers. (B–H) E20 neuroretina 14 days in vitro. (A, B) Hematoxylin and eosin staining. (C–H) Immunohistochemistry. (B) Cultured neuroretina is significantly thinner than the P12 controls and displays an ONL without any other distinguishable layers except for a few scattered cells at the innermost margin (arrows). (C) Transducin-positive cones with a varying degree of labeling intensity are present throughout the ONL. (D) Strongly labeled rhodopsin-positive rods can be seen throughout the specimen with strong labeling intensity in the inner segments (arrows). (E, F) A few scattered, poorly labeled rod bipolar and horizontal cells positive for PKC and calbindin are seen at the innermost margin of the explant. (G) Synaptophysin-labeled processes are seen throughout the explant without any apparent organization in plexiform layers. (H) Vimentin-labeled Müller cell fibers are seen throughout the specimen, with strong labeling intensity in the vitread end-feet region (arrows). (I) Negative control. Scale bar, 50 μm.

Figure 6.

 

Rhodopsin (red) and synaptophysin (green) double labeling. In all explant groups, rhodopsin and synaptophysin labeling are colocalized to a great extent, which can be seen in the composite image (orange). Scale bars, 50 μm.

Figure 7.

 

Rhodopsin (yellow), vimentin (green), and DAPI (blue) triple labeling. (A) In the adult control, rhodopsin labeling is concentrated to the outer segments (OS) of the rod photoreceptors, with minimal labeling located in the ONL and the OPL. Vimentin-labeled Müller cells display the normal vertical arrangement, with strong intensity in the innermost part of the retina. DAPI-labeled neuronal nuclei are seen in the ONL, INL, and GCL. (B) In the E17 7DIV explant, rhodopsin-labeled rod photoreceptors are present in the outer half of the specimen, and vimentin labeling of Müller cells is mostly seen in the innermost half. DAPI-labeled cells are seen throughout the specimen. (C) In the E17 14DIV explant and (D, E) in both groups of E20 specimens, rhodopsin-labeled cells are present in the entire specimen except for in the innermost part, where vimentin-labeled Müller cell fibers are seen. DAPI labeling is seen throughout the explants, coinciding with rhodopsin labeling in all but the innermost part of the specimens. Scale bars, 50 μm.