Introduction

Sympathetic ophthalmia (SO) is defined as a bilateral granulomatous panuveitis following penetrating ocular trauma and intraocular surgery to one eye.1 The injured eye is referred to as the exciting eye and the fellow eye as the sympathizing eye. Although the pathophysiology of the disease is not clearly understood, an autoimmune process incited by the initial penetrating injury or by surgery appears to play a significant role. The typical histopathology is characterized by a diffuse non-necrotizing granulomatous infiltration seen throughout the uveal tract. The inflammatory cell infiltrate is composed primarily of numerous lymphocytes with nests of non-necrotizing granulomas that consist mainly of epithelioid and multinucleated giant cells. At the level of the retinal pigment epithelium, multiple Dalen–Fuchs nodules are observed that are composed of inflammatory cells, mainly monocytes, macrophages, and epithelioid cells.2, 3, 4

The mechanisms governing the recruitment of leucocytes into the eye in SO are not fully understood. Chemokines and matrix metalloproteinases (MMPs), in particular gelatinase B (MMP-9), play key roles in the migration of leucocytes to sites of inflammation.5, 6 Chemokines are a superfamily of 5- to 8-kDa secreted proteins that direct the recruitment of leucocytes to sites of inflammation. These chemokines are grouped into the CXC, CC, C, and CX3C subfamilies on the basis of the arrangement of the conserved cysteine residues.5 Monocyte chemotactic protein-1 (CCL2/MCP-1), the most potent CC chemokine, is a chemoattractant and activator for monocytes, T lymphocytes, basophils, and natural killer cells.7 Stromal cell-derived factor-1 (CXCL12/SDF-1), constitutively expressed in a broad range of tissues, was initially cloned from mouse bone marrow stromal cells and as CXC chemokine originally described as pre-B-cell growth-stimulating factor.8 It is a chemoattractant for monocytes, naive and memory T lymphocytes, and B lymphocytes.9, 10, 11 SDF-1 is also a costimulatory factor for CD4+ T-lymphocyte activation.12 In view of the presence of an important mononuclear cell infiltrate in SO, MCP-1 and SDF-1 are candidate chemokines for investigation. In addition, pharmacological SDF-1 blockage was effective in animal models of autoimmune diseases. AMD3100, a potent and specific antagonist of the SDF-1 chemokine receptor CXCR4, inhibited autoimmune joint inflammation in interferon-gamma receptor-deficient mice.13

The MMPs represent a family of Zn2+-containing endopeptidases that are recognized as key enzymes both for normal extracellular matrix turnover and for the exaggerated extracellular matrix breakdown associated with pathologic conditions including tumour cell invasion and metastasis, angiogenesis, inflammatory reactions, wound healing, and scar formation.14, 15 Gelatinase B cleaves denatured collagens (gelatins), collagen types IV, V, VII, and X, elastin, and fibronectin.14 Because of its unique and broad substrate spectrum, its involvement in leucocyte migration, and its role in other chronic inflammatory and autoimmune diseases,6, 15 we hypothesized that excessive production of gelatinase B may play a role in the pathogenesis of SO.

In the present study, we investigated the subtypes of mononuclear cells present in the inflammatory infiltrate and the expression of gelatinase B and the chemokines MCP-1 and SDF-1 in five enucleated exciting eyes with the classic clinical presentation and typical histopathological findings of SO.

Materials and methods

Five formalin-fixed, paraffin-embedded enucleated eyes with the classic clinical presentation and typical histopathologic findings of SO were obtained from the registry of Ophthalmic Pathology, Department of Ophthalmology, University of Ghent, Belgium. All the ocular tissues were of the exciting eyes. Serial sections of 5 μm thickness were utilized for the study and stained with haematoxylin and eosin for routine analysis and studied by immunohistochemical techniques.

Immunohistochemical staining

After deparaffinization, endogenous peroxidase was abolished with 2% hydrogen peroxide in methanol for 20 min, and nonspecific background staining was blocked by incubating the sections for 5 min in normal swine serum. For CD3, immunoglobulin (Ig) A, IgG, and MCP-1 detection, antigen retrieval was performed by boiling the sections in 10 mM Tris-EDTA buffer (pH 9) for 30 min. For HLA-DR, CD20, SDF-1, and gelatinase B detection, antigen retrieval was performed by boiling the sections in 10 mM citrate buffer (pH 6) for 30 min. For CD68 detection, the sections underwent trypsinization for 10 min at 37°C using a mixture of 0.1% trypsin (Sigma-Aldrich, Bornem, Belgium) and 0.1% CaCl2 at pH 7.8. Subsequently, the sections were incubated with the monoclonal and polyclonal antibodies listed in Table 1. Optimal working concentration and incubation time for the antibodies were determined earlier in pilot experiments. For gelatinase B immunohistochemistry, slides were incubated for 30 min with gelatinase B-specific REGA-2D9 monoclonal antibody (1:100). The mouse monoclonal antibody REGA-2D9 was raised against natural human neutrophil gelatinase B. This implies that the antigen preparation was devoid of gelatinase A. The REGA-2D9 is an IgG1 subtype with a dissociation constant (Kd) value of 9.5 × 10−10 M, which implies high specificity.16 For CD3, CD68, HLA-DR, CD20, IgA, IgG, MCP-1, and SDF-1 immunohistochemistry, the sections were incubated for 30 min with goat anti-rabbit or anti-mouse immunoglobulins conjugated to peroxidase-labelled dextran polymer (EnVision+; Dako, Carpinteria, CA, USA). For gelatinase B immunohistochemistry, the sections were incubated for 30 min with the biotinylated secondary antibody and reacted with the avidin–biotinylated peroxidase complex (Dako). The reaction product was visualized by incubation for 10 min in 0.05 M acetate buffer at pH 4.9, containing 0.05% 3-amino-9-ethylcarbazole (Sigma-Aldrich) and 0.01% hydrogen peroxide, resulting in bright-red immunoreactive sites. The slides were faintly counterstained with Harris haematoxylin. Finally, the sections were rinsed with distilled water and coverslipped with glycerol. Omission or substitution of the primary antibody with an irrelevant antibody of the same species and staining with chromogen alone were used as negative controls. Control antibodies included anti-carcinoembryonic antigen (Dako), which is an IgG1 Kappa, and anti-cytokeratin 10/13 (Dako), which is an IgG2a Kappa. Sections from patients with Crohn's disease were used as positive controls.

Table 1 Monoclonal and polyclonal antibodies used in the study

Results

Table 2 provides the data on each of the five cases concerning the age of the patient at the time of injury, sex, inciting event, time interval from injury to onset of symptoms in the sympathizing eye, and time interval from onset of inflammation in the sympathizing eye to enucleation of the exciting eye. None of the patients was on systemic immunosuppressive therapy before enucleation of the exciting eye.

Table 2 Subject characteristics

In all cases, an important inflammatory infiltrate composed of mononuclear cells was present. The cellular infiltrate showed a diffuse distribution involving the whole thickness of the uveal tract but was more intense on the scleral side of the choroid. It was composed primarily of numerous lymphocytes with few plasma cells and in addition many large, epithelioid cells containing dark brown pigment granules. Among the lymphocytic infiltrate, there were many large granulomas consisting of epithelioid cells with or without multinucleated giant cells (Figure 1). Dalen–Fuchs nodules were observed in three cases.

Figure 1
figure 1

Photomicrographs showing extensive inflammatory infiltration in the choroid composed predominantly of lymphocytes (a), and a choroidal granuloma composed predominantly of epithelioid cells and multinucleated giant cells (b) (haematoxylin and eosin, original magnification, × 40).

Immunohistochemical studies revealed no staining in the negative control slides and when the chromogen alone was applied. A summary of the results is given in Table 3. In all cases, the lymphocytic infiltrate of the uveal tract was predominantly composed of CD20+ B lymphocytes, whereas CD3+ T lymphocytes were fewer (Figure 2). The majority of plasma cells were IgA+. CD68+ monocytes/macrophages were scattered in the diffuse inflammatory infiltrate. Epithelioid cells and multinucleated giant cells within granulomas and cells in Dalen–Fuchs nodules were CD68+ (Figure 3). Most of the inflammatory mononuclear cells were HLA-DR+. In addition, epithelioid cells and multinucleated giant cells within granulomas were HLA-DR+.

Table 3 Summary of immunohistochemical data
Figure 2
figure 2

Immunohistochemical staining for T-cell marker (CD3) (a) and B-cell marker (CD20) (b) showing predominance of CD20+ B cells in the choroidal inflammatory infiltrate (original magnification, × 40).

Figure 3
figure 3

Immunohistochemical staining for monocyte/macrophage marker (CD68) showing immunoreactivity within choroidal granuloma (arrowheads) and in scattered inflammatory mononuclear cells (a), and in inflammatory mononuclear cells within Dalen–Fuchs nodule (arrow) (b) (original magnification, × 40).

In all cases, cytoplasmic immunoreactivity for MCP-1, SDF-1, and gelatinase B was noted predominantly in epithelioid cells and multinucleated giant cells within granulomas but also in the scattered epithelioid cells present in the diffuse inflammatory infiltrate. Cytoplasmic immunoreactivity for MCP-1 was also noted in retinal pigment epithelial cells (Figure 4), whereas a weak SDF-1 immunoreactivity was noted on endothelial cells of the choriocapillaries.

Figure 4
figure 4

Immunohistochemical staining for MCP-1 showing immunoreactivity within choroidal granuloma (arrowheads) and in retinal pigment epithelial cells (arrow) (a), SDF-1 showing immunoreactivity in scattered inflammatory mononuclear cells (arrows) (b), and gelatinase B showing immunoreactivity in cells within choroidal granuloma (c) (original magnification, × 40).

Discussion

In the present study, we demonstrated the following points: (1) B cells predominated in the uveal infiltrates in all SO cases; (2) cells from the monocyte/macrophage lineage, particularly the epithelioid cells and multinucleated giant cells within the choroidal granulomas, were the major cell types expressing gelatinase B and the chemokines MCP-1 and SDF-1; and (3) retinal pigment epithelial cells expressed MCP-1.

Immunohistochemical studies have described the choroidal infiltrates of SO to be predominated by activated T cells, and the proportion of B cells varies from 10 to 40%.2 Chan et al2 have reported a predominance of T helper cells in the early stage of disease and T suppressor cells at a later stage of the disease, indicating that the cellular kinetics in SO changes over time. Our results, however, show predominance of B cells in the uveal inflammatory infiltrate. Our findings are in agreement with previous reports that demonstrated predominance of B cells in the choroidal inflammatory infiltrate in SO.4, 17 In an immunohistochemical study of 29 exciting eyes, Shah et al4 demonstrated that predominance of B cells in four cases of their series was significantly correlated with longer duration of the disease. They suggested that, although SO is a T-cell-mediated disease, the predominance of B cells in some cases may represent the end stage of the disease process, or seems a secondary pathological process. In their study, the duration of disease referred to inflammation in the sympathizing eye was less than 3 months in 16 cases, 3–9 months in three cases, and more than 9 months in 10 cases. In the present study, it was not possible to determine from the medical records the duration of inflammation in the exciting eyes before the onset of symptoms in the sympathizing eyes. However, the time interval from onset of symptoms in the sympathizing eyes and enucleation of the exciting eyes was 15 days or less. In addition, the predominance of B cells in the choroidal inflammatory infiltrate was reported in one case of progressive subretinal fibrosis, a variant of SO,18 and in one case of multifocal choroiditis.19 Interestingly in experimental autoimmune uveitis, B-cell numbers increased progressively to become the predominant cell type in end-state lesions.20

Functional antigen-driven B-cell follicle-like structures with germinal centres and numerous plasma cells are found in the target tissues of autoimmune diseases such as the salivary glands of Sjögren's syndrome, rheumatoid synovial membranes, and thyroid glands from patients with Hashimoto's thyroiditis and Grave's disease.21, 22 B cells have many potential key roles in the pathogenesis of autoimmune diseases. It has recently been proposed that B lymphocytes may play a more critical role in the induction of immunological activities as an antigen-presenting cell population, and that they are essential for the initial development and/or activation of autoreactive T cells in rheumatoid arthritis.23 In addition, they can secrete chemokines and proinflammatory cytokines, produce autoantibodies, and activate T cells.24 The notion that B cells might be critical to the development of rheumatoid arthritis led to the extension of the use of selective B-cell depletion with anti-CD20 therapy to this condition and a recent double blind controlled trial has shown encouraging results.24 In the present study, the majority of plasma cells were IgA+, suggesting a pathogenic role for IgA in SO. The formation of IgA-containing immune complexes can activate the alternative complement pathway. Furthermore, there is increasing evidence that many leucocytes have receptors for IgA and binding of IgA to these receptors triggers a variety of effector functions.25 Similarly, the isotype of rheumatoid factor produced by B-cell lines developed from B-cell populations infiltrating the synovial tissue of patients with advanced rheumatoid arthritis was mostly IgA.26 In addition, several studies showed raised serum levels of IgA, IgA rheumatoid factor, and IgA-containing circulating immune complexes in rheumatoid arthritis and that IgA rheumatoid factors are produced locally in salivary glands and in synovial tissue.27, 28 Furthermore, rheumatoid arthritis patients with a predominantly increased IgA rheumatoid factor had a more erosive disease and a high frequency of associated sicca syndrome.29 The prevalence of IgA rheumatoid factor in juvenile rheumatoid arthritis was also found to be significantly higher in the polyarticular subset than the pauciarticular and the systemic onset group.30

SDF-1 is a chemokine that plays an important role in B-cell development, trafficking, and activation.8, 9, 10, 11 Several studies suggested that SDF-1 may have a crucial role in the development of autoimmune disease. Matin et al31 demonstrated that the chemokine SDF-1 may be an essential chemoattractant in trafficking and migration of mature autoreactive B cells from bone marrow to the periphery or inflammatory sites in the development of autoimmune diabetes. Furthermore, they showed that anti-SDF-1 antibody was effective in inhibiting autoimmune diabetes in non-obese diabetic mice.31 Additionally, the administration of anti-SDF-1 antibody prevented the development of autoantibodies, nephritis, and death in a murine model of lupus.32 The expression of SDF-1, which is capable of inducing lymphocyte migration to lymphoid organs,33 is increased in chronically inflamed tissues such as rheumatoid arthritis synovium,34, 35, 36 and the thyroid gland from patients with autoimmune thyroid disease.22 On the basis of a wide expression of SDF-1 in rheumatoid arthritis synovium, several studies demonstrated that SDF-1 and its receptor CXCR4 are involved in CD4+ and CD8+ T-cell34, 35 and CD68+ monocyte/macrophage36 migration to and retention in rheumatoid arthritis synovium. The expression of SDF-1 by monocytes/macrophages and endothelial cells in SO agrees with immunolocalization studies in rheumatoid arthritis synovium,36 and suggests a pathogenic role of SDF-1 in the recruitment of leucocytes including B lymphocytes into the eye. Our results are in agreement with a recent report that demonstrated early upregulation of SDF-1 mRNA during the evolution of experimental autoimmune anterior uveitis.37

Granulomatous inflammation is a specific type of chronic inflammation that is characterized by monocyte recruitment and activation and the accumulation of macrophages and their derived cells, epithelioid cells, and multinucleated giant cells.38 Kreipe et al39 demonstrated that multinucleated giant cells are highly stimulated cells of the monocyte/macrophage lineage at a terminal stage of maturation. MCP-1 is an important mediator of monocyte-rich pathological processes. In vitro studies have shown that MCP-1 is a potent monocytic chemotactic factor that can also upregulate adhesion molecule expression and cytokine production by monocytes.7, 40 In the present study, MCP-1 immunoreactivity was predominantly associated with epithelioid cells and multinucleated giant cells present in granulomas. Our observations are consistent with previous reports showing that MCP-1 is produced in granulomatous inflammation and that MCP-1 reactivity is present primarily in macrophages, epithelioid cells, and multinucleated giant cells within granulomas.41, 42, 43, 44 He et al42 demonstrated that phagocytosis can serve as an important stimulus for MCP-1 production by macrophages. In addition, Flory et al41 demonstrated that administration of anti-MCP-1 antibody resulted in a dramatic decrease in the number and size of granulomas in a rat model of glucan-induced pulmonary granulomatous vasculitis.

In the present study, immunoreactivity for MCP-1 was also detected in retinal pigment epithelial cells. Our results are in agreement with previous in vitro studies that demonstrated active MCP-1 secretion by human retinal pigment epithelial cells stimulated with the inflammatory cytokines interleukin-1β or tumour necrosis factor-α.45, 46, 47 Furthermore, it was shown that MCP-1 secreted by retinal pigment epithelial cells was bioactive and that antibodies directed against MCP-1 strongly inhibited monocyte chemotaxis elicited by supernatants of retinal pigment epithelial cells.45 In general, the expression of MCP-1 in SO is consistent with previous reports, which demonstrated upregulation of MCP-1 protein and mRNA in the eyes of animals with experimental autoimmune uveitis37, 48, 49, 50 and in the aqueous humour from patients with acute anterior uveitis.51

In previous studies, we and others have described increased levels of gelatinase B in aqueous humour samples from patients with active uveitis that correlated significantly with the clinical disease activity.52, 53 Here we demonstrate that epithelioid cells and multinucleated giant cells in SO granulomas exhibited cytoplasmic immunoreactivity for gelatinase B. The intracellular localization of gelatinase B in cells from the monocyte/macrophage lineage can be interpreted as evidence of active synthesis of this enzyme because macrophages do not store this metalloproteinase.54 Similarly, previous studies demonstrated immunoreactivity for gelatinase B in epithelioid cells, multinucleated giant cells, and macrophages within granulomas in other human granulomatous disorders.55, 56, 57

Expression of gelatinase B has been associated with a wide variety of pathological processes, including autoimmune diseases such as multiple sclerosis,58 rheumatoid arthritis,59 and type I diabetes.60 The mechanism of action of gelatinase B in the pathogenesis of autoimmunity may be through the cleavage of substrate proteins into immunodominant epitopes. Further processing and presentation of the protein fragments on major histocompatibility complex class II molecules on the surface of antigen-presenting cells activates autoreactive T cells.58 In addition, gelatinase B is an important effector molecule for the migration of leucocytes.6 As one of the main components of endothelial basement membrane is type IV collagen, which is a substrate of gelatinase B,61 gelatinase B might increase the extravasation of leucocytes into the eye through the basement membrane zone underlying endothelial cell layer. Gelatinase B could also render basement membranes vulnerable and increase vascular permeability leading to disruption of the blood–ocular barrier in patients with SO. Therefore, selective inhibition of gelatinase B should be considered as a potential therapy in patients with uveitis. Recently, El-Shabrawi et al62 demonstrated that selective inhibition of gelatinase A and gelatinase B completely prevented the development of uveitis in almost half of the animals with experimental autoimmune uveitis.

In conclusion, the findings reported in this study provide evidence that immunoreactivity for gelatinase B and the chemokines MCP-1 and SDF-1 was predominantly associated with epithelioid cells and multinucleated giant cells within the choroidal granulomas in SO. These findings suggest a pathogenic role for gelatinase B and the chemokines MCP-1 and SDF-1 in the regulation of leucocyte recruitment in SO.