Immune Cells in the Porcine Retina: Distribution, Characterization and Morphological Features

Abstract

purpose. To investigate the presence of immunocompetent cells in the porcine retina and to compare the findings with those obtained earlier in human retinas.

methods. Retinal wholemounts or cryostat sections from outbred Dutch Landrace pigs were analyzed for the presence of microglia (CD45), macrophages-monocytes (SW3, CD163, 2A10, CD14), major histocompatibility complex (MHC) class II–positive cells (MCA1335), granulocytes (MCA1219), B lymphocytes (IgM), and T lymphocytes (CD6, CD4, CD8), by using specific monoclonal antibodies followed by immunohistochemical staining.

results. A uniform distribution of CD45-positive microglial cells was observed throughout the porcine retina (mean number, 289 ± 16 cells/mm²). The microglia were observed along blood vessels and within the tissue between the inner limiting membrane and the inner nuclear layer. MHC class II–positive cells were mainly observed along the large- and middle-sized retinal blood vessels. Double-staining experiments showed that 54% of the microglial cells lining the larger retinal vessels were MHC class II positive. Macrophages were only incidentally observed along the larger retinal blood vessels. No T lymphocytes, B lymphocytes, monocytes, or granulocytes were observed within the retinal tissue.

conclusions. The porcine retina contains a rich network of microglial cells. Approximately half of the microglial cells lining the larger retinal vessels express MHC class II. The normal porcine retina is devoid of lymphocytes, monocytes, and granulocytes. The distribution of immunocompetent cells in the porcine retina largely resembles that observed in the human retina.

Intraocular inflammation is an important cause of blindness throughout the world. Advances in our understanding of the pathogenesis and treatment of uveitis are largely due to the development of animal models. In some of these models, retinal antigens are used to induce ocular inflammation (experimental autoimmune uveoretinitis; EAU) in a variety of animals including rats, mice, and monkeys. ¹ Although the studies of EAU in different species have greatly increased our understanding of the mechanisms involved in the pathogenesis of human uveitis, differences are also apparent. There are many human posterior uveitis entities, and EAU may only represent a few of these disorders. Many human uveitis entities differ from EAU in their clinical presentation and in the duration and recurrence of the ocular inflammation. ^{2 3} These differences could be due to quantitative or qualitative variations between immunocompetent cells in the retinas of the various species used. 

Retinal tissue macrophages, also known as microglia, are thought to play an important role in inflammatory processes in the retina. Earlier studies have shown that there are numerous major histocompatibility complex (MHC) class II–positive microglial cells in the human retina, ^{4 5} whereas no, or very few, MHC class II–positive cells have been reported in the retina of rats. ^{6 7} Whether and to what extent central nervous system (CNS) glial cells can function as antigen-presenting cells has received a great deal of attention over the past few years, but is still controversial. Interferon-γ treatment can markedly augment MHC class II expression in rat retinal microglia, but as yet no evidence is available showing that these cells can actually function as antigen-presenting cells. ⁷ Depending on the experimental system used, CNS microglia have been shown to be capable of antigen processing and restimulation of T cells, whereas astrocytes are not considered to be efficient antigen-presenting cells. ⁸  

One of the limitations in these studies is the source of microglial cells. CNS microglial cells have been obtained from newborn rodent or adult human brains and expanded in vitro. To analyze the antigen-presenting potential of retinal microglia, these cells could be obtained either from many small rodent eyes or from eyes of a larger animal. The pig has been suggested as an appropriate animal for human biomedical research. ⁹ Of all nonprimate mammals, the pig is most similar to humans. It is omnivorous, is a daytime animal, and has a physiology, anatomy, and body size very similar to that of humans. 

The dimensions of the porcine eye are similar to the human eye, and, of the domestic and laboratory species, the anatomic and physiological features of the pig most closely resemble that of humans. ^{10 11} The pig orbit is open and continuous with the temporal fossa. Unlike humans, the pig contains both a Harderian and nictitans gland, the latter being emptied through a nictitating membrane. The drainage of the anterior chamber occurs through a scleral venous plexus, and there is no structure such as a single Schlemm canal. The retina of the pig resembles the human retina, in that is has an extensive vascular tree fed by four large retinal arteries. Although there is no true fovea, the porcine retina has a central area that is devoid of larger blood vessels. ¹²  

As yet, it is not known whether the porcine retina has features in common with the human retina regarding the presence of immune cells. The present study was designed to investigate the presence of immunocompetent cells in the porcine retina using immunohistochemical analysis of retinal flatmounts. Numerous MHC class II–positive microglial cells were detected along the larger vessels in the porcine retina, similar to results obtained earlier in the human retina. 

Materials and Methods

Thirty-six eyes were obtained from normal female outbred Dutch Landrace pigs (between 2 and 3 months of age). These pigs were used to prepare kidney cell cultures for the diagnosis of the foot and mouth disease virus infection in our institute. All animals were treated according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Pigs were killed by intravenous pentobarbital injection. Pigs were treated with heparin before death and perfused with cold phosphate-buffered saline (PBS, pH 7.4) through the carotid artery to remove blood from the capillary beds of the eyes. In a number of cases, eyes were used for making 8-μm-thick cryostat sections. The spleen was removed from one animal to provide a positive control for the various monoclonal antibodies used in this study. For preparation of the retinal wholemounts, the eyes were dissected behind the ciliary body into an anterior and posterior part. The lens and vitreous were removed, and the posterior eye cup was cut into quadrants. The surface of the retina was brushed gently with a small paintbrush to remove remaining vitreous. The retina was subsequently separated from the underlying retinal pigment epithelium and choroid. The obtained retinas were fixed in either cold 100% ethanol or 4% paraformaldehyde for 5 minutes and then placed in PBS in a 24-well tissue culture dish. 

Single Immunohistochemical Staining of Retinal Wholemounts and Cryostat Sections

The tissue pieces were incubated in 0.06% hydrogen peroxidase for 10 minutes to eliminate endogenous peroxidase activity. Increased permeabilization of the tissue was achieved by incubation in PBS with 1% bovine serum albumin (BSA) and 0.1% Tween-20 for 20 minutes. Incubation with the primary antibody was performed at room temperature for 1 hour or at 4°C overnight. The secondary and tertiary steps were performed at room temperature in an incubation of 30 minutes each. Shorter incubation times were sufficient for these latter steps, probably because of a higher avidity of these latter polyclonal antibodies in comparison with the monoclonals used as a primary reagent. Primary monoclonal antibodies used in the study were mouse anti-porcine CD45 leukocyte common antigen (MCA1222 and all porcine leukocytes; Serotec, Oxford, UK), mouse anti-porcine MHC class II antigen (MCA1335; Serotec; described by Hammerberg and Schurig ¹³ ), mouse anti-human HLA DQ class II antigen (MCA379G; Serotec), mouse anti-porcine SWC3 (clone 74.22.15; common myelocytic marker for macrophages, monocytes, and granulocytes; described by Lunney ¹⁴ ), mouse anti-porcine 2A10 (CD163 subpopulation of macrophages; described by Sanchez et al. ¹⁵ ), mouse anti-porcine CD163 (subpopulation of monocytes-macrophages CVI Swine NL 517.2; ID-Lelystad-BV, Lelystad, The Netherlands ¹⁶ ), mouse anti-porcine CD14 (MCA1218 and monocytes; Serotec), mouse anti-porcine granulocytes (MCA1219, B-cells, and granulocytes; Serotec), mouse anti-porcine CD6 (MCA 1221 and pan T cell; Serotec), mouse anti-porcine CD4 (CD4⁺ T cells; described by Pescovitz et al. ¹⁷ ), mouse anti-porcine CD8 (CD8⁺ T cells; described by Jonjic and Koszinowski ¹⁸ ), mouse anti-porcine IgM (B lymphocytes; CVI-SwIgM-28.4, catalog no.7500950; ID-Lelystad-BV ¹⁹ ), mouse anti-human CD86 (clone BU63; Dako, Glostrup, Denmark). All antibodies were diluted in PBS containing 1% BSA. After incubation with the primary antibody and washing with PBS (three times, 5 minutes each), a biotinylated rabbit anti-mouse immunoglobulin (Dako) was applied for 30 minutes at room temperature, followed by incubation with peroxidase-conjugated streptavidin (Dako) for 30 minutes. Horseradish peroxidase activity was developed by using 3,3-diaminobenzidine tetrahydrochloride (Sigma, St. Louis, MO) or 3-amino-9-ethylcarbazole (AEC; Sigma) as the substrate. Retinal tissue pieces incubated with buffer without the primary antibody were used as a control to ensure the absence of nonspecific staining. Cryostat sections of pig spleen served as a positive control. 

Quantitative analysis was performed as follows. Positive cells in three separate fields of a retinal wholemount were counted using a calibrated eye piece graticule under a microscope with a 20× objective lens, and the mean number of these cells per retina was calculated. Retinas from 20 pigs were examined, and the mean ± SD was calculated. 

Double Immunohistochemical Staining of Retinal Wholemounts

The double staining for the CD45 marker and the MHC class II antigen was performed by incubating retinal wholemounts with a primary monoclonal mouse anti-human class II IgG2a at 4°C overnight. Previous experience in our laboratory has shown that this antibody cross-reacts with porcine class II, which is in accordance with earlier reports showing extensive cross-reactions between a variety of anti-human MHC class II monoclonal antibodies and porcine class II molecules. ²⁰ Double staining could not be performed with the antiporcine MHC class II antibody, because it had the same isotype as the anti-CD45 antibody. Tissue pieces were washed (three times in PBS, 5 minutes each) before incubation with alkaline phosphatase-conjugated goat anti-mouse IgG2a (Southern Biotechnology Associates, Birmingham, AL) for 1 hour at room temperature. Tissues were subsequently incubated with mouse anti-pig CD45 IgG1 (Serotec) at room temperature for 1 hour, rinsed in PBS and incubated with peroxidase-conjugated goat anti-mouse IgG1 (Southern Biotechnology Associates) for 30 minutes at room temperature. The alkaline phosphatase chromogen was developed first, using fast Blue BB (Sigma) for 20 minutes at 37°C. Tissues were washed thoroughly, and AEC was used as a substrate for the development of peroxidase activity. The retinal wholemounts were placed on gelatin-coated glass slides with the inner retinal side up and finally embedded in an aqueous mounting medium. The red chromogen contrasted well with the blue color, and double-labeled cells appeared purple. 

Results

Analysis of microglial cells in the porcine retina was performed by using a monoclonal antibody directed against the pan leukocyte marker CD45. This marker has been used earlier by others to analyze microglial cells in the human retina. ⁴ Numerous CD45⁺ cells with a dendriform appearance were observed throughout the porcine retina (Fig. 1A) . Cells were observed adjacent to the retinal blood vessels but also within the tissue. No qualitative difference in the density of these cells was noted when different parts of the retina were compared (posterior versus peripheral). 

Almost all CD45⁺ cells observed in the retina had a dendriform morphology, although round cells were noted occasionally within small vessels in retinas obtained from eyes that had not been perfused (data not shown). In perfused eyes, only the dendriform cells were noted in the retina. Calculation of the number of CD45⁺ cells in the porcine retina showed a mean (±SD) of 289 ± 16 cells/mm². 

MHC class II⁺ cells were mainly observed along the retinal blood vessels, especially the large- and middle-sized vessels (Fig. 1B) . Very few MHC class II–positive cells were noted within the retina. The MHC class II⁺ cells also showed a dendriform appearance and had the same morphologic features as the CD45⁺ microglial cells. To further substantiate this observation, double staining was performed for CD45 and MHC class II. Quantitative analysis showed that 53.8% of the microglial cells were MHC class II positive, and no single-labeled MHC class II–positive cells were observed (Figs. 1C 1D) . 

SWC3⁺ cells (staining a common myelomonocytic antigen) were detected along the retinal blood vessels, although staining with this antibody was not very strong. These cells also had dendriform morphology. CD163⁺ cells (macrophage subpopulation) were occasionally observed in association with the larger retinal blood vessels. The CD163⁺ retinal macrophages had a very characteristic appearance with a strongly staining taillike dendrite (Fig. 1E) . Analysis of choroid and iris flatmounts for CD163-positive macrophages, revealed a dense network of these cells (data not shown). No CD14 positive cells were observed in the porcine retina, whereas a large number of these cells were readily identified in iris or choroid (Chen et al., manuscript in preparation). Staining of retinal wholemounts for granulocytes did not reveal any positive cells. 

Only occasional IgM⁺ (B lymphocytes), CD6⁺, CD4⁺, or CD8⁺ cells (T lymphocytes) were noted within retinal vessels from nonperfused eyes. No positive cells were detectable in retinas after the eyes had been properly perfused. Staining of retinal wholemounts for IgM as a marker of B lymphocytes revealed a strong appearance of all retinal vessels (Fig. 1F) . IgM was associated with the vessels, and its presence decreased after prolonged perfusion of the eyes. B lymphocytes express IgM on the cell surface, and this marker is often used to stain tissue sections for B cells or in the flow cytometric analysis of B cells in mixtures of isolated cell populations. Because IgM is one of the immunoglobulin classes that is also abundantly present as a soluble protein in plasma, it is not surprising that the vascular endothelial cells reacted positively when the retinal wholemounts were tested with a monoclonal anti-IgM antibody. 

Analysis of the distribution of immunocompetent cells in retinal cryostat sections showed CD45⁺ cells and MHC class II⁺ cells along the blood vessels and in the retinal tissue between the inner limiting membrane and the inner nuclear layer. Unlike the results in the wholemounts the positive cells were not intact and appeared only as irregular or pleiomorphic brown dots. Staining of retinal cryostat sections for IgM showed positive staining of the retinal vessels. 

At present, no reagents are available to analyze the presence of costimulatory molecules in the pig. We attempted staining with a mouse monoclonal antibody directed against human CD86 (B70/B7-2), but analysis of porcine spleen cryosections did not produce a positive result, indicating the absence of cross-reactivity. 

Discussion

In this study, MHC class II–positive microglial cells were observed in association with the larger blood vessels of the porcine retina. At present, we can draw no conclusions as to the true three-dimensional arrangement of the retinal microglia, especially in relation to their exact association with vessels (intraluminal or perivascular). Immunoelectron or confocal microscopic examinations are needed to further clarify this issue. Our findings in the pig are similar to those reported earlier in the human retina. ^{4 5} The group of Penfold ⁴ have shown that virtually all microglial cells in the human retina express MHC class II antigens. In the porcine retina, approximately 54% of the vessel-associated microglial cells expressed MHC class II antigens. Analysis of MHC class II expression on rat retinal microglial cells showed that these cells do not regularly express this antigen, ^{21 22} but that expression is induced by treatment with interferon-γ. ⁷ This contrasts with the observation that human microglial MHC class II expression is downregulated after stimulation with interferon-γ and lipopolysaccharide (LPS) through an IL-10–dependent mechanism. ²³  

It is not clear why not all porcine retinal microglia expressed MHC class II, as was reported in the human retina. It may be a true species-dependent phenomenon or, alternatively, may be due to differences in the sensitivity of the methods used. Another possible explanation is a difference in the length of time between death and processing of the retinal tissue, which is much shorter for porcine than for human eyes. Analysis of the effect of postmortem time on microglial MHC expression may resolve this question. 

The exact role of MHC class II antigen expression on retinal microglia is not yet clear. Furthermore, it is not yet known whether these cells express costimulatory molecules needed for appropriate antigen presentation. In vitro studies using cultured rat retinal microglia showed that these cells constitutively express B7-2 and that stimulation with interferon-γ leads to the expression of both MHC class II and B7-1, indicating that these cells may play a role in local antigen presentation. ²⁴ Functional tests of the antigen-presenting capacity of retinal microglial cells have not yet been reported. Analysis of costimulatory molecules in the porcine retina is not yet possible, because of the absence of proper reagents. Expression of MHC class II antigens on porcine CNS or retinal microglia has not yet been reported. However, the expression of MHC class II antigens on microglia in the CNS of other species, including humans, has now been well established. ^{8 25} Recent observations strongly suggest that microglial MHC expression is controlled by neurons through a CD200 receptor ligand interaction. ²⁶ The CD200 receptor is restricted to cells of the myeloid lineage, and interaction with tissues expressing the CD200 ligand is considered to negatively regulate macrophage and microglial activation. The CD200 receptor has been shown to be absent in normal retinal microglia, in contrast to the expression of this receptor on macrophages present at sites of retinal inflammation. ²⁷ Comparison of cytokine and free radical production between porcine, murine, and human microglia show no differences between the source of the cells and their release of TNF-α or interleukin-1. ²⁸ In contrast to mouse microglia, both human and porcine microglial cells are unable to generate nitric oxide after cytokine stimulation. 

Specialized macrophage subpopulations, expressing the CD163 antigen, were only occasionally observed in the porcine retina. In this respect, the pig resembles the rat, in which the tissue macrophage marker ED2 was also not found in retinal flatmounts, ²⁹ although low numbers were observed after flow cytometry of rat retinal tissue homogenates. ²¹ No cells carrying the CD14 marker were observed in the porcine retina, whereas these cells were readily observed in the iris and choroid (Chen et al., manuscript in preparation). In the human retina, a subpopulation of retinal microglia was identified by using the CD68 marker. ^{4 5} This marker is also expressed by human tissue macrophages, but has not yet been identified in the pig. 

The microglia are evenly distributed throughout the porcine retina, whereas in humans a marked increase has been observed in the density of these cells in the peripheral retina. ⁵ This difference is probably not caused by our studying only young pigs, because earlier studies analyzing the ontogeny of retinal microglia in humans have already noted high densities of these cells in the peripheral retina at as early as 10 weeks of gestation. ³⁰ Even distribution of microglia in the retina has been reported earlier in a number of mammals including cat, rabbit, owl monkey, galago, slow loris, tree shrew, ferret, and raccoon. ³¹ The mean density of microglial cells in the pig amounted to approximately 300 cells/mm², which is similar to the densities of human retinal microglia reported earlier by others. ⁴ Given the numbers of microglial cells in the porcine retina, it should be feasible to isolate these cells and perform functional antigen-presentation studies. 

T or B lymphocytes were not detected in the normal porcine retina. These results are in agreement with those observed earlier in the human retina. ^{4 5} Using IgM as a marker for B lymphocytes, a marked uniform staining of the retinal vessels was observed that was similar to the staining seen when a monoclonal antibody was used that is directed to vascular endothelial cells. These results indicate that IgM can be used as a marker to study retinal vasculature changes in models of diabetic retinopathy or after photodynamic therapy. 

The data presented add further support to the use of the pig as a model to study human ocular disease. The availability of pigs, the similarity in eye size to the human eye, the sharing of anatomic features, and the presence of many commercially available porcine reagents facilitate the use of this animal in eye research. 

 Supported by the collaboration project between the Education Ministry of China and Ministry of Education, Culture and Science of The Netherlands (coordinated by the Royal Academy of Sciences, The Netherlands); Grant 39925034 from the National Outstanding Youth Foundation of China (PY); a Trans-century Scientists and Outstanding Teacher Grant from the Education Ministry of China; Grant 99M048212, Guanddong Province Natural Science Key Project and International Collaboration Foundation; and the Edward and Marianne Blaauw Fund.

 Submitted for publication November 2, 2001; revised January 8, 2002; accepted January 16, 2002.

 Commercial relationships policy: N.

 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.

Figure 1.

 

Single and double immunohistochemical staining of wholemounted porcine retinas with various monoclonal antibodies. (A) Distribution and morphologic features of CD45⁺ microglial cells in the normal porcine retina. (B) MHC class II⁺ (anti-porcine monoclonal antibody) microglial cells lining a larger retinal blood vessel. The faint staining in the background was due to the expression of MHC antigens on the smaller retinal vessels. (C) Porcine retinal flatmount double stained for CD45 and MHC class II. Double-stained purple cells and single-stained red-brown CD45⁺ cells were observed along a large retinal blood vessel. In the retinal tissue adjacent to the blood vessel, a few single-stained microglial cells were observed. (D) Higher magnification of (C), in which a single-labeled microglial cell is in the middle of the panel, with a CD45⁺ MHC class II⁺ double-labeled cell to the right of it. (E) CD163⁺ cells lining a retinal blood vessel. Note the tail-like dendrite of these cells. (F) Marked staining of the porcine retinal vasculature for IgM.