PATIENTS AND METHODS

Patients

Eleven consecutive Saudi patients with active VKC seen at the outpatient clinic of King Abdulaziz University Hospital were included in the study. The patients were seven males and four females. The mean age was 12.6 (SD 2.3) years (range 8–15 years). The symptoms mentioned by all the patients were itching, redness, photophobia, and tearing. Each patient underwent complete ophthalmic examination, and the corneal and conjunctival changes were noted and recorded. All patients had the limbal form of the disease characterised by broad gelatinous infiltrates of the limbus. Limbal conjunctival biopsy specimens were obtained from each patient. None of the patients was on topical therapy before obtaining the biopsy. In addition, eight limbal conjunctival biopsy specimens were obtained from patients undergoing strabismus surgery without obvious inflammation and served as controls. None of the controls had a history or signs of VKC. The controls were from the same age group, and were five males and three females. All the controls were Saudi, and were operated in King Abdulaziz University Hospital. This study was approved by the Research Center, College of Medicine, King Saud University, and the patients admitted to the study gave their informed consent.

Immunohistochemical staining

The conjunctival biopsy specimens were immediately snap frozen in Tissue-Tek optimum cutting temperature (OCT) compound (Miles Laboratories, IN, USA) and maintained at −80°C until use. For immunohistochemistry, 5 μm serially cut cryostat sections were dried overnight at room temperature, fixed in absolute acetone for 10 minutes, and then treated with 2% hydrogen peroxide in methanol for 3 minutes to block endogenous peroxidase activity. After rinsing three times in phosphate buffered saline (PBS) at pH 7.2 for 15 minutes, the slides were incubated for 30 minutes with the monoclonal and polyclonal antibodies listed in Table 1. Optimal concentrations of all antibodies used were determined in pilot experiments. After a wash with PBS, the sections were incubated for 30 minutes with Envision+, Peroxidase, Rabbit, or EnVision+, Peroxidase, Mouse (Dako, CA, USA). These are goat anti-rabbit or anti-mouse immunoglobulins conjugated to peroxidase labelled dextran polymer. The products react with rabbit immunoglobulins or with mouse immunoglobulins of all classes and minimally with human immunoglobulins, thus allowing better visualisation. The slides were washed again with PBS and the reaction product was visualised by incubation for 10 minutes in 0.05M acetate buffer at pH 4.9, containing 0.05% 3-amino-9-ethylcarbazole (Sigma-Aldrich, Bornem, Belgium) 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 was used as a negative control. Positive controls were previously positive staining tissues. As positive controls, we used samples from other tissues, including small intestine from patients with Crohn’s disease, synovia from patients with rheumatoid arthritis, liver, and lymph nodes, showing a positive staining reaction with the same set of antibodies as that used in the present study.

Double immunohistochemistry

To examine the phenotype of inflammatory cells expressing chemokines, cryostat sections were studied by sequential double immunohistochemistry. Co-localisation studies were performed in four VKC specimens. After rinsing the slides with PBS, they were incubated for 30 minutes with the monoclonal antibody to determine the cellular phenotype (CD68 (KP1); 1:1000, monocytes/macrophages, Dakopatts A/S, Copenhagen, Denmark) and rinsed again in PBS. Subsequently, the sections were incubated for 30 minutes with Envision+, Peroxidase, Mouse (Dako, CA, USA) and washed again with PBS. Then, the reaction product was visualised by incubation for 10 minutes in 0.05M acetate buffer at pH 4.9, containing 0.05% 3-amino-9-ethylcarbazole and 0.01% hydrogen peroxide, resulting in red immunoreactive staining. Afterwards the sections were rinsed with PBS, washed with distilled water and incubated for 30 minutes with the chemokine antibodies. After a wash with PBS, the sections were incubated for 30 minutes with a rabbit anti-mouse alkaline phosphatase labelled antibody (Sigma-Aldrich). The blue reaction product was developed using fast blue BB salt (4-benzoylamino-2.5-diethoxybenzene-diazonium chloride) (Sigma-Aldrich) for 5 minutes.

Quantitation

Cells were counted in five representative fields that were chosen on the basis of the presence of an adequate number of inflammatory cells. We ignored fields in which no positively stained cells were present. We used an eye piece calibrated grid with 40× magnification. With this magnification and calibration, we counted the cells present in an area of 0.33 × 0.22 mm.

Statistical analysis

All data are presented as mean (SD). The data were analysed using one way analysis of variance (ANOVA), non-parametric one way ANOVA based on the Kruskal-Wallis test, correlation analysis, and ridge regression. Programs 4R, 7D, and 3S from the BMDP Statistical Package were used. The differences were considered significant if the p value was <0.05.

RESULTS

There was no staining in the negative control slides. In normal conjunctiva, weak CD26 immunoreactivity was observed on vascular endothelial cells in the upper substantia propria. There was no immunoreactivity for PARC, MDC, I-309, CCR4, CD25, CD62L, CD71, and CD30.

In VKC specimens, a heavy inflammatory infiltrate of CD3⁺ T lymphocytes was noted in the epithelium and in the substantia propria just beneath the epithelium in all specimens. Inflammatory mononuclear cells expressing granular cytoplasmic PARC, MDC, and I-309 were noted in the upper substantia propria at sites of heavy CD3⁺ T lymphocyte accumulation (Fig 1). PARC⁺ inflammatory cells were noted in all specimens, whereas MDC⁺ inflammatory cells were observed in nine specimens and I-309⁺ inflammatory cells were observed in six specimens. Both PARC antibodies that we used generated the same labelling pattern on inflammatory cells. Membranous immunoreactivity for CCR4 was noted on inflammatory mononuclear cells in the upper substantia propria in three specimens.

Double immunohistochemistry to confirm the phenotype of chemokine positive inflammatory cells showed that all the mononuclear cells expressing chemokines were CD68 positive monocytes/macrophages (Fig 2). Many of the PARC⁺ inflammatory cells had a dendritic morphology.

The numbers of inflammatory cells expressing PARC were higher than the numbers of inflammatory cells expressing MDC and I-309 (Table 2). The mean values of the three groups differed significantly (p = 0.0117, ANOVA), Furthermore, post-ANOVA pairwise comparisons based on the Bonferroni method showed that the numbers of inflammatory cells expressing PARC were significantly higher than the numbers of inflammatory cells expressing I-309 (p = 0.0032, t test).

Membranous immunoreactivity for CD25, CD26, CD62L, CD71, and CD30 was noted on mononuclear inflammatory cells in the epithelial and stromal inflammatory infiltrate (Fig 3). Inflammatory cells expressing CD25, CD26, CD62L, and CD30 were noted in all specimens, whereas CD71⁺ cells were observed in seven specimens. A marked upregulation of CD26 expression was noted on superficial and deep stromal vascular endothelial cells (Fig 4). CD30 expression was confirmed by staining with two anti-CD30 monoclonal antibodies. The activation markers and CCR4 were shown to be expressed on CD3⁺ T lymphocytes by serial sections (Fig 3).

The numbers of inflammatory cells expressing CD25 were higher than the numbers of inflammatory cells expressing CD26, CD62L, CD71, and CD30 (Table 2). The mean values of the five groups differed significantly (p <0.001, ANOVA done by non-parametric Kruskal-Wallis test). Furthermore, post-ANOVA pairwise comparisons showed that the numbers of inflammatory cells expressing CD25 were significantly higher than the numbers of inflammatory cells expressing CD71 (Z* = 4.24) and CD30 (Z = 4.44). The numbers of inflammatory cells expressing CD26 were significantly higher than the numbers of inflammatory cells expressing CD71 (Z = 3.59) and CD30 (Z = 3.79). (*The critical Z value was Z ≥2.81 for five groups at a 5% level of significance (Kruskal-Wallis test).)

Correlations between the numbers of CD3⁺ T lymphocytes and the numbers of inflammatory cells expressing chemokines

A significant positive correlation was observed between the numbers of CD3⁺ T lymphocytes and the numbers of inflammatory cells expressing PARC (r = 0.6709, p = 0.0238), MDC (r = 0.7453, p = 0.0085), and I-309 (r = 0.6389, p = 0.0343). Ridge regression analysis indicated that the numbers of inflammatory cells expressing PARC had the strongest correlation (79% explained variation) with the numbers of CD3⁺ T lymphocytes.

Correlations between the numbers of CD3⁺ T lymphocytes and the numbers of inflammatory cells expressing activation markers

A significant positive correlation was observed between the numbers of inflammatory cells expressing CD3⁺ T lymphocytes and the numbers of inflammatory cells expressing CD25 (r = 0.7789, p = 0.0047), CD62L (r = 0.6464, p = 0.0316), and CD71 (r = 0.6259, p = 0.0394). Ridge regression analysis indicated that the numbers of inflammatory cells expressing CD25 had the strongest correlation (46.1% explained variation) with the numbers of CD3⁺ T lymphocytes.

Correlations between the numbers of inflammatory cells expressing chemokines and the numbers of inflammatory cells expressing activation markers

A significant positive correlation was observed between the numbers of inflammatory cells expressing PARC and the numbers of inflammatory cells expressing CD25 (r = 0.8782, p = 0.0004). In addition, there was a significant positive correlation between the numbers of inflammatory cells expressing MDC and the numbers of inflammatory cells expressing CD26 (r = 0.6354, p = 0.0357).

DISCUSSION

In vitro, PARC mRNA is expressed by monocyte derived dendritic cells, and activated monocytes.^12,15–19 PARC preferentially attracts naive resting T cells (CD45RA⁺).¹² However, Hieshima et al¹⁶ demonstrated that PARC is chemotactic for both activated (CD3⁺) T cells and non-activated (CD14⁻) lymphocytes. PARC binds to a yet unknown receptor expressed on naive T cells.^12,16 The specific expression of PARC by dendritic cells at the site of initiation of an immune response, combined with its chemotactic activity for naive T cells, suggests that PARC has an important role in the induction of immune responses.¹²

In the present study, we detected cytoplasmic expression of PARC by CD68 positive monocytes/macrophages. Many of these cells had a dendritic morphology. In agreement with our data, previous studies showed that PARC mRNA was restricted to CD68 positive macrophages in human atherosclerotic plaques.¹⁵ In addition, high PARC expression was noted in alveolar macrophages and dendritic cells in the germinal centres of regional lymph nodes.^12,16,19

MDC is constitutively produced by dendritic cells, macrophages, and thymic medullary epithelial cells, whereas monocytes, B lymphocytes, natural killer cells, and CD4⁺ T lymphocytes produce MDC only upon appropriate stimulation.^20–25 More recently, MDC was shown to be preferentially produced by activated Th2 cells.²⁶ MDC is chemotactic for activated T lymphocytes, IL-2 activated natural killer cells, and monocyte derived dendrtic cells.^20,21 Subsequently, MDC was shown to act selectively on chronically activated Th2 lymphocytes by interacting with the CCR4 receptor,¹³ which is preferentially expressed by this T cell effector subset.^27,28 More recently, MDC was also found to be able to induce human eosinophil chemotaxis in a CCR3 and CCR4 independent manner.²⁹

The conjunctiva from patients with VKC showed cytoplasmic expression of MDC, and I-309 by CD68 positive monocytes/macrophages. Previous studies have reported expression of MDC by dendritic cells in skin biopsy specimens from patients with atopic dermatitis, a Th2 oriented disorder.^25,26 In a mouse model of allergic inflammation in the lung, MDC is produced by alveolar macrophages.³⁰ In this animal model, Lloyd et al³¹ showed that the CCR4/MDC pathway has a dominant role in effector Th2 recruitment under conditions of chronic, repeated antigen stimulation. The expression of I-309 by monocytes/macrophages in the conjunctiva from patients with VKC is consistent with in vitro studies that demonstrated I-309 expression and secretion by appropriately stimulated human peripheral blood monocytes.³² Surprisingly, few T cells in VKC conjunctiva expressed CCR4. Our finding of poor CCR4 expression might be due to CCR4 internalisation after stimulation with its ligands.^33,34

The expression of PARC, and MDC by monocytes/macrophages in VKC conjunctiva might be explained by upregulation by inflammatory cytokines. Recently, several studies demonstrated that the Th2 cytokines, IL-4, and IL-13 induced PARC expression in macrophages¹⁹ and MDC expression in monocytes.^24,35

In this study, we found that the T cells infiltrating the conjunctiva of patients with VKC express several activation markers. CD25⁺ cells were the most numerous followed by CD26⁺ cells and CD62L⁺ cells. CD71⁺ cells and CD30⁺ cells were few. In addition, the numbers of CD25⁺ cells had the strongest correlation with the numbers of CD3⁺ T lymphocytes. T cells express CD25 upon activation³⁶ and CD25 expression is promoted by the Th2 cytokine, IL-4.³⁷ Stimulation of CD4⁺ T cells secreting Th2 type cytokines by dendritic cells elicits T cell proliferation, production of cytokines, and upregulates CD25 expression.³⁸ Recently, Suto et al³⁹ demonstrated that CD4⁺ CD25⁺ T cells modulate the Th1 and Th2 cell balance towards Th2 cells and thus upregulate Th2 cell mediated allergic inflammation in the airways. Our findings are in agreement with previous studies in other allergic disorders.^40–43

CD26 is a cell surface protease constitutively expressed on a wide variety of epithelial, endothelial, and lymphoid cell types. It is present on resting T cells, and T cell activation is accompanied by enhanced expression of CD26. CD26 has been shown to have a role in T cell activation and T cell mediated immunity.^44–46 The conjunctiva from patients with VKC showed strong expression of CD26 on T cells and vascular endothelial cells. Similar expression was noted in the human bronchus.⁴⁷ In vitro studies demonstrated that CD4⁺ memory T cells with the capacity for transendothelial migration express CD26 and CD25 and that CD26 is one likely candidate for control of T cell transmigration.⁴⁸ It is suggested that CD26 on blood vessels might be involved in the processing of intravascular peptides, such as substance P and in the regulation of substance P induced plasma leakage.⁴⁷

Recently, several studies demonstrated that CD26 can process several chemokines such as eotaxin, RANTES, and MDC.^23,49,50 After cleavage of eotaxin by CD26, its chemotactic potency for blood eosinophils and its signalling capacity through the CCR3 receptor were reduced.⁴⁹ Processed RANTES is inactive in chemotaxis of purified monocytes and enhances chemotactic migration of T cells.⁵⁰ Truncated MDC does not interact with CCR4 and does not attract Th2 cells.²³ Recently, Wilheim et al⁵¹ demonstrated that CD26 expression is correlated with a Th1-like phenotype. Therefore, CD26 expressed mainly on Th1 cells may represent a negative feedback for allergic reactions mediated by CC chemokines.

Del Prete et al⁵² have shown that CD30, a member of the tumour necrosis factor/nerve growth factor receptor superfamily, is preferentially expressed by human T cell clones producing Th2 cytokines. In contrast, Hamann et al⁵³ demonstrated that CD30 can be expressed by Th1-like cells, contributing to interferon γ secretion. Furthermore, Alzona et al⁵⁴ found that CD30 defines a subset of activated T cells that can produce both interferon γ and IL-5 and exhibit potent activity for B cell immunoglobulin production. Evidence for in vivo involvement of CD30⁺ cells in allergic disease has appeared from immunohistochemical finding of CD30⁺ cells in acute atopic dermatitis biopsies. On the other hand, no significant CD30 expression could be found in subacute/chronic atopic dermatitis lesions or in any of the specimens of allergic contact dermatitis.⁵⁵ In the present study, CD30 expression by a few T cells in the conjunctiva of patients with VKC was confirmed by staining with two anti-CD30 monoclonal antibodies (Ber-H2, and Ki-1). Similarly, CD71, a marker expressed by activated proliferating T cells,⁵⁶ was expressed by a few T cells.

CD62L is a cell surface glycoprotein present on most circulating human lymphocytes. It is mainly involved in the adhesion of T cells to the high endothelial venules of peripheral lymph nodes and subsequent extravasation in peripheral lymph nodes. In addition, it is suggested that CD62L is also crucial for migration of lymphocytes to sites of inflammation. Lymphocytes shed CD62L upon activation, but can be re-expressed upon return to the resting state.⁵⁷ The conjunctiva from patients with VKC contained relatively few CD62L⁺ lymphocytes. Our observations are consistent with previous data in inflamed skin, appendix, and synovium,⁵⁸ and in the airways of a murine model of allergic asthma.⁵⁹ These findings are consistent with shedding of CD62L upon activation.⁵⁸ Interestingly, Kanegane et al⁶⁰ demonstrated that CD62L positive human memory CD4⁺ T cells produce mainly Th2 cytokines, whereas CD62L negative CD4⁺ T cells produce mainly Th1 cytokines.

In conclusion, PARC and MDC expressed by CD68 positive monocytes/macrophages may function to recruit naive and activated T lymphocytes in VKC. PARC could play a major part in the recruitment of naive T cells which might initiate the immune responses. PARC and MDC might further attract activated T cells. T lymphocytes infiltrating the conjunctiva of patients with VKC express several activation markers which might contribute to the pathogenesis of VKC.