Immunohistochemical study of pig retinal development

Animal models

All animal procedures were performed in compliance with the UK Animals (Scientific procedures) Act 1986. Mixed sex white Landrance pigs were obtained from Agri-Food and Biosciences Institute, Northern Ireland (Large Park, Hillsborough, Co. Down, Northern Ireland, UK) where animals were commercially bred on slats and fed with granulated pig food. The pigs were anaesthetised with intra-muscular azoperone (15 mg/kg) and ketamine (20 mg/kg) and euthanized with intravenous or intra-cardiac euthanasia-grade pentobarbitone (100 mg/kg).

Eyes were collected from embryos in the ninth gestational week (GW9), 1 week old piglets (1W), and 6 months (6M) old animals. Three pigs from each age group were analyzed.

Postnatal day (P) 5 in the mouse was used as control for antibody functionality and comparison of developmental stage. Wild type C57BL6 mice were bred in 12 h dark/light regime with free access to food and water. P5 mice were euthanized by intraperitoneal injection with pentobarbitone. The eyes from three P5 mice from different litters were analyzed. Comparison with adult mouse retina was performed using the data from published literature which is accordingly cited in the results section.

Preparation of retina for histology and immunohistochemistry

Eyes were enucleated from euthanized pigs, and the cornea and the lens were removed. The eyecups were fixed in 4% paraformaldehyde in PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 2mM KH₂PO₄ pH 7.4) for 1–4 h at room temperature. After fixation the specimens were cryoprotected in 10% sucrose for 6 h followed by 30% sucrose overnight. The eyecups were embedded in optimal cutting temperature compound (OCT; Sakura, Kobe, Japan) and snap frozen in an isopentane bath on dry ice. Transverse 16 µm cryosections were cut, mounted onto Superfrost Plus glass slides (Fisher Scientific, Loughborough, UK) and stored at −80 °C until use.

Slides were thawed at room temperature and post-fixed in 4% formaldehyde (Sigma-Aldrich, Poole, UK) in PBS for 20 min at room temperature. After rinsing in PBS, sections were blocked for 1 h in 10% normal goat serum, 0.3% Triton X-100, 0.01% NaN₃ in PBS, at room temperature. Slides were rinsed in PBS and incubated for 24 h at 4 °C with primary antibody diluted in 10% normal goat serum, 0.3% Triton X-100, 0.01% NaN₃ in PBS. The source information and the dilutions for all primary antibodies used are listed in the Table 1. Following removal of the primary antibody, slides were washed 6 times for 5 min in PBS and incubated for 1 h at room temperature in 1:500 fluorescent-conjugated secondary antibody (Alexa Fluor⁴⁸⁸ or Alexa Fluor⁵⁶⁸ goat anti-mouse or goat anti-rabbit) in PBS. Slides were then washed 3 times for 5 min in PBS at room temperature and cell nuclei were counterstained with 10 µg/ml propidium iodide (Sigma-Aldrich, Poole, UK) in PBS containing 10 µg/ml RNase A (Sigma-Aldrich) for 30 min at room temperature, or with 5 μM DAPI (Invitrogen, Paisley, UK) for 10 min. Slides were mounted in fluorescent mounting medium (Dako, Ely, UK). Negative immunohistochemistry (IHC) controls were performed in parallel by omission of primary antibody. Immunoreactive cells were visualized and images recorded with an inverted confocal microscope (Nikon, Model Eclipse TE 2000-U, Tokyo, Japan) using the Nikon Confocal Microscope EZ-C1 software. Representative images were taken from the midperipheral region of the retina.

Progenitor cells in the developing pig retina

Antibodies to intermediate filament protein nestin and the homeodomain transcription factor Pax6 were used to identify retinal progenitor cells in the developing pig retina (Figure 1). Nestin immunoreactivity was detected throughout GW9 pig retina, reflecting the undifferentiated state of the majority of cells at this stage (Figure 1A). The staining pattern was similar to that found in the P5 mouse retina (Figure 2A), where the strongest immunoreactivity localized to fibers in the ganglion cell layer (GCL). In the postnatal pig retina, nestin immunolabeling was seen in the cell fibers in the GCL and inner plexiform layer (IPL; Figure 1B) probably representing the extensions of Müller glia cells as has been confirmed in the 8W rat retina [25].

In GW9 pig retina, Pax6 strongly labeled most cells in the GCL, cells in the inner portion of the neuroblast layer (NBL) and some cells in the prospective outer plexiform layer (OPL) (Figure 1C). The rounded cells in the inner portion of the NBL represented putative differentiating amacrine cells. Labeled cells with elongated shape and weaker staining found throughout the outer NBL appeared to be migrating progenitors. This observation was corroborated by immunolabeling for Ki67, a marker for proliferating cells, which revealed many mitotic cells throughout the NBL, excluding the innermost portion where amacrine cells differentiate (Figure 1D). Rounded Pax6-labeled cells in the developing OPL (Figure 1C) are suggestive of differentiating horizontal cells [23]. GW9 pig retina contained more migrating progenitor cells compared to P5 mouse retina in which Pax6 labeling mainly colocalized with the prospective position of ganglion, amacrine, and horizontal cells (Figure 2B). Pax6 immunoreactivity was found in ganglion, amacrine, and some horizontal cells in the 1W pig retina (Figure 1E).

Neurons of the ganglion cell layer

The antibodies to POU domain transcription factor Brn3a and intermediate neurofilament (NF)-160 were chosen to characterize ganglion cells in the developing, postnatal and mature pig retina (Figure 3). The activating protein α (AP2α) transcription factor, an exclusive marker for amacrine cells, was used to visualize the population of displaced amacrine cells within the GCL. Essentially, the same labeling of the GCL was observed in 1W and 6M pig retinas, therefore only images of 1W pig retinas are shown.

Brn3a immunoreactivity was found exclusively in the GCL in GW9 pig retina (Figure 3A). The observed staining pattern, with densely labeled cells, closely resembled that found in P5 mouse retina (Figure 2C). Brn3a-immunopositive cells were evenly spaced in the GCL of 1W pig retina (Figure 3B).

In GW9 pig retina, NF-160 staining was observed in the GCL, IPL, and in some differentiating amacrine and horizontal cells (Figure 3C). A similar pattern was found in the P5 mouse (Figure 2D), with the exception of strong staining in the OPL, suggestive of more advanced histogenesis. In the 1W pig retina, robust immunoreactivity was seen in the GCL and neurofiber layer, with fine labeling of processes in the IPL and OPL (Figure 3D).

A weak AP2α-positive signal was found in sparse cells in the GCL of GW9 pig retina (Figure 3E). In P5 mouse retina AP2α immunoreactivity was significantly stronger, labeling a large proportion of cells in the GCL and prospective inner nuclear layer (INL) (Figure 2E). Similarly to the staining pattern in adult mouse [26], AP2α in the 1W (Figure 3F) and 6M (not shown) pig retina labeled cells densely distributed in the GCL and the inner portion of the INL, suggestive of displaced amacrine and amacrine cells respectively.

Neurons of the inner nuclear layer

To characterize developing neurons in the INL, namely amacrine and horizontal cells, we used antibodies against calretinin, calbindin, and the 65 kDa isoform of glutamic acid decarboxylase (GAD65; Figure 4). Since these markers label cell processes, they also gave a good indication of the development of plexiform layers. The LIM homeodomain transcription factor Islet1 and protein kinase C α isoform (PKCα) were primarily used to visualize bipolar cells (Figure 5).

In the developing pig retina at GW9, calretinin immunoreactivity was found in a subpopulation of cells in the GCL and rare cells in the inner NBL, suggestive of amacrine phenotypes (Figure 4A). Sparse cells in the developing OPL were also weakly labeled. Calretinin immunoreactivity in correspondence to ganglion and amacrine cells was significantly stronger in P5 mouse retina (Figure 2F) than in GW9 pig retina. Weaker labeling was found in mouse OPL. 1W and 6M retina (Figure 4B,C) displayed a calretinin immunostaining pattern similar to that of adult mouse (8–10W), with labeled ganglion, amacrine, and horizontal cells and the distinction of the three strata in the IPL [27].

A subpopulation of cells in the GCL and some cells in the developing OPL were strongly calbindin-immunopositive in GW9 pig retina (Figure 4D). Weak staining was found in sparse developing amacrine cells. Labeling was also seen at the outer edge of the NBL, suggesting immunoreactivity of some developing photoreceptors. Labeling of amacrine cells and of the IPL was more prominent in P5 mouse retina (Figure 2G) compared to GW9 pig, where the INL and IPL only occasionally displayed weak positive signal. The labeling pattern of horizontal, amacrine, and ganglion cells and their processes in the OPL and IPL in both 1W and 6M pig retinas (Figure 4E,F) closely resembled that described for the adult mouse retina. In the pig (1W and 6M) retina, calbindin also labeled sparse photoreceptors, which had a typical wide and tapered cone-like shape (Figure 4E, inset), and did not stain for rhodopsin in double labeling experiments (not shown). Cones appeared immature and mostly inner segments were visualized in 1W pig retina (Figure 4E, inset). At 6M, cones were intensely labeled for calbindin, with inner and outer segments clearly visible (Figure 4F).

In the GW9 pig retina, GAD65 immunostaining was found in amacrine and ganglion cell bodies, with stronger signal localized to the IPL (Figure 4G). Some GAD65-positive cells with immature labeled processes, suggestive of developing horizontal cells, were also highlighted. This staining pattern was similar to P5 mouse retina (Figure 2H), albeit less intense. In mouse, weak signal was also seen in the developing horizontal cells and OPL. In postnatal and adult pig retinas, strong labeling was found in processes layering the IPL, with a characteristic striated appearance (Figure 4H,I) as seen in adult mouse retina [27,28]. Some labeled horizontal cell bodies were visible in 6M pig retina (Figure 4I).

In the GW9 pig retina, moderate to intense Islet1 immunoreactivity was observed in the GCL and in some amacrine cells at the inner border of the NBL (Figure 5A). Strongly immunopositive cells were also localized to the distal part of the NBL in the prospective OPL. The position and round shape of these cells suggests their differentiation toward horizontal cells. Contrary to findings in pig at GW9, no horizontal cells were labeled in P5 mouse retina (Figure 2I). However, developing bipolar cells in the central portion of the INL were strongly Islet1-immunopositive. From the first postnatal week onwards (Figure 5B,C) Islet1-positive staining was found in the GCL, IPL, and INL of the pig retina. At 1W and 6M, Islet1-immunoreactive amacrine cells, as well as densely distributed cell bodies, were visible in the inner and outer portion of the INL, respectively (Figure 5B,C). Some cells in the outer portion of the INL had a spindle shape characteristic of bipolar cells, whereas others appeared rounded, suggestive of horizontal cell phenotypes. Double-labeling experiments with calbindin (Figure 5B, inset) confirmed the presence of Islet1-immunoreactive horizontal cells in 1W pig retina, a feature also described for human retina (12–85 years-old) [29].

Weak PKCα immunolabeling was detected in the GCL, IPL, inner NBL, and in the prospective OPL of GW9 pig retina (Figure 5D). P5 mouse retina displayed a similar labeling pattern (Figure 2J). However, the staining was more prominent compared to GW9 pig. In both 1W and 6M pig retina (Figure 5E,F) oval-shaped immunoreactive cells were found in the INL, with their cell bodies located close to the OPL and their processes extended toward both OPL and IPL, as is characteristic of rod bipolar cells. Axon processes in the IPL were clearly visualized in the 1W pig retina (Figure 5E), whereas staining in the 6M was mainly localized to the axon endings of rod bipolar cells, presented in the form of characteristic swellings or granulate labeling (Figure 5F). Some cell bodies adjacent to the vitreous and their processes were PKCα-immunoreactive in the 1W and 6M pig retinas. No PKCα-positive amacrine or cone photoreceptor cells were detected in the 1W or 6M pig retina, as reported for the mouse (8–10W) [27,30].

Photoreceptors

An antibody to calcium binding protein recoverin, which labels both rods and cones, was used as an early marker of photoreceptor differentiation (Figure 6). Rod photoreceptors were further characterized with an anti-rhodopsin antibody. Antibodies for mouse S-opsin and M-opsin used for this study did not cross-react with pig proteins (data not shown); therefore we used an antibody against Gα subunit of the cone transducin (GNAT2) to label cones in the pig retina.

In the GW9 pig retina, strong recoverin immunoreactivity was detected in putative differentiating photoreceptors, localized to the apical surface of the NBL, and in sparse cells found deeper in the inner NBL (Figure 6A). Recoverin staining was seen throughout the developing ONL in P5 mouse retina (Figure 2K). However, labeling at the apical surface of the NBL was weaker than in GW9 pig. Some recoverin-positive ganglion cells were observed in mouse P5, but not in pi GW9g developing retina. In the 1W pig retina, recoverin immunoreactivity was seen throughout the ONL, including in the inner and outer segments (Figure 6B,C). Photoreceptor segments appeared longer at 6M than at 1W. Some recoverin-positive cells localized to the prospective INL in the GW9 pig retina could be regarded as developing bipolar cells (Figure 6A). However, no conspicuous bipolar cell labeling was observed in 1W or 6M pig retina (Figure 6B,C).

In GW9 pig retina, sparse elongated rhodopsin-immunoreactive cells were observed localized to the prospective ONL, suggestive of early differentiating rod photoreceptors (Figure 6D). Some rudimentary inner segments were also weakly labeled. No rhodopsin-positive cells were detected in the P5 mouse retina with the same antibody (data not shown). Rod outer segments were strongly labeled in both 1W and 6M pig retina, whereas inner segments and some cells bodies displayed weaker staining (Figure 6E,F). Rod outer segments appeared longer and more defined in 6M compared to 1W old animals, suggesting that outer segments continue to extend in length well into the postnatal period.

No labeling with GNAT2 antibody was observed in the GW9 pig retina. Similarly, mouse cones at P5 were not positive for this marker (not shown). Cone outer segments in both 1W and 6M pig retina were strongly immunoreactive for GNAT2 (Figure 6G,H). The size of the cone outer segments slightly increased from the 1W to the 6M pig retina.

Retinal glial cells

The antibodies to low affinity neurotrophin receptor P75 (P75^NTR) and glutamine synthetase (GS) were used to characterize Müller glia, whereas an anti-GFAP antibody was used to label astrocytes. In the GW9 pig retina, strong P75^NTR staining was present in the neurofiber layer (Figure 7A), similar to P5 mouse retina (Figure 2L). Labeled processes extended through the GCL and IPL and entered the inner portion of the NBL. Additional immunoreactivity was seen in the developing OPL, with the majority of stained cell bodies localized in the distal half of the prospective INL and some cell processes extending to the ONL. From the postnatal period onward the labeling pattern for P75^NTR in the pig retina remained unchanged, although it appeared that IPL displayed stronger staining in the 1W compared to the 6M retina (Figure 7B,C).

In the GW9 pig retina GS positive labeling was seen mainly in the GCL, where it likely represents processes of the developing Müller glial cells. Weaker staining was also observed in the IPL and the inner portion of the NBL (Figure 7D). Interestingly P75^NTR and GS staining pattern were almost identical in the P5 mouse retina (Figure 2L,M), whereas in the GW9 pig, P75^NTR staining was significantly stronger, suggesting temporal advantage in the expression of this marker. GS immunolabeling pattern in the 1W and 6M pig retina was very similar to that seen for P75^NTR, revealing strong staining in the GCL, IPL and INL, suggestive of Müller glia cells, as it has also been shown for adult (8–10W) rat retina (Figure 7E,F) [31,32]. As seen also for P75^NTR, somewhat stronger staining for GS was found in the IPL of 1W compared to the 6M pig retina.

In GW9 pig retina, GFAP immunoreactivity was observed adjacent to the vitreous in cells with thin and short processes, suggestive of immature developing astrocytes (Figure 7G). In the 1W retina, the majority of the staining was localized in the neurofiber layer and GCL, with rare processes protruding into the IPL (Figure 7H). Astrocytic processes in the adult retina became denser and longer, extending from the neurofiber layer through the GCL, IPL and INL, occasionally reaching the ONL (Figure 7I). PKCα colocalized with glial fibrillary acidic protein (GFAP) immunostaining in double labeling experiments, identifying PKCα-positive astrocytes (Figure 7J-N). Astrocytes in 1W pig retina were weakly labeled for PKCα, but in 6M pig retina these cells were strongly PKCα immunoreactive (Figure 7J-N).