Characterization of T cells and cytokines in the aqueous humour (AH) in patients with Fuchs' heterochromic cyclitis (FHC) and idiopathic anterior uveitis (IAU)

Abstract

FHC and IAU are two forms of anterior uveitis which are localized to the eyes with no evidence of systemic involvement. However, FHC has distinct clinical features and differs from IAU in that the inflammation is low grade, steroid non-responsive, and has a less aggressive clinical course. To try to dissect the mechanism for this difference the phenotypes of the cells in the AH and blood (PB) and the cytokines present in the AH in patients with FHC and IAU were compared. Three-colour flow cytometry was performed on the cells isolated from the AH and PB. Percentage of cells bearing the following markers were determined: CD3, CD4, CD8, CD4/CD25, CD8/CD25, CD19 and CD14. The cytokines IL-4, IL-10, IL-12 and interferon-gamma (IFN-γ) were assayed by ELISA. In both groups T cell numbers were higher in the AH than PB, although the distribution of T cell subsets in PB was similar. In the AH, CD8⁺ T cell numbers were higher in FHC than in IAU (P = 0·003), whilst CD4⁺ numbers were higher in IAU than FHC (P = 0·01). AH cytokine profiles were different in the two groups: IFN-γ levels were higher and IL-12 levels lower in the FHC group than IAU (P = 0·02), whilst IL-10 levels tended to be higher in the FHC group (P = 0·5). We suggest that different local mechanisms governing the balance of T cell/cytokine-mediated inflammation in the anterior segment may underlie clinical differences such as chronicity and response to steroids in these disorders.

INTRODUCTION

Anterior uveitis is the most common form of uveitis, accounting for 75% of all uveitis cases with an annual incidence of eight cases per 100 000 population [1–3], and consists of inflammation predominantly in the anterior chamber and the ciliary body. It can be part of some well defined clinical syndromes, e.g. sarcoidosis or Behçet's disease, or a localized ocular disorder as in IAU and FHC. Even though FHC and IAU are both localized ocular diseases, they are strikingly different in their manifestations. FHC is characterized by a white eye with a chronic, low-grade anterior uveitis with widely scattered small non-pigmented keratic precipitates, a variable degree of iris atrophy and depigmentation of the iris. In contrast, IAU presents with either acute or chronic inflammation in which keratic precipitates tend to be distributed mainly in the inferior half of the corneal endothelium. Keratic precipitates comprise macrophages or other leucocytes which are deposited over the posterior layer of the cornea, i.e. the endothelium. One other contrasting feature between FHC and IAU is the response to topical steroids. Topical steroids do not help reduce inflammation in FHC, whereas in IAU they have a significant effect. The explanation of this differing response to topical steroids has not been found, nor why FHC has such a benign chronic inflammatory course.

Many workers have investigated the histopathology of FHC and other types of chronic uveitis from iris biopsies taken at the time of cataract surgery, and despite the use of light microscopy [4–7] and electron microscopy [8–10], no difference in the pathology of FHC compared with other types of uveitis has been identified. The overall picture is that of a chronic inflammation with mainly lymphocytes and plasma cell infiltration in addition to some eosinophils and mast cells. An immunohistochemical analysis of the iris samples obtained by peripheral iridectomy at the time of cataract or glaucoma surgery in patients with FHC and other types of uveitis did not detect any specific immunohistological differences [11]. Hence it has been the general consensus that it is impossible to differentiate FHC from other types of chronic iridocyclitis on histopathologic and immunohistologic changes seen in the iris.

The differing clinical profile between FHC and other types of chronic uveitis, despite similar histological changes, suggests that there may be some differences in the functional characteristics of the infiltrating cells influencing the outcome of disease in these entities. The similar cellular types present structurally in these eyes with different clinical features may imply that the cells are functionally different. Several workers have investigated the lymphocyte subsets in the peripheral blood of FHC and other types of uveitis [12–15]. However, in uveitis occurring without systemic involvement, immunological changes occurring within the eye may not be reflected in the peripheral blood. The aim of this study was to compare the cellular phenotypes in AH and peripheral blood (PB) of patients with FHC and IAU and the cytokine profile present in the AH in these two clinically distinct types of uveitis. The cellular phenotypes were quantified simultaneously in the PB to enable any selective increase of any of the cellular phenotypes present in the AH to be interpreted.

PATIENTS AND METHODS

Patients

Patients with anterior uveitis presenting to the clinic were characterized on clinical grounds, with additional investigations to exclude systemic disease as required. The two study groups of patients were identified, i.e. 10 FHC and 18 IAU patients. IAU was defined as active anterior uveitis with no evidence of systemic disease clinically or after appropriate investigations. None of these patients was receiving systemic steroids. All patients with FHC had classic clinical features and were not on any topical therapy, and those with IAU were either on topical steroids or not on any treatment.

AH sampling procedure

Ethical approval was obtained and informed consent given by the patient for both paracentesis and venesection. Anterior chamber paracentesis was carried out using a 30 G needle on a 1-ml insulin syringe. AH (100–200 μl) was aspirated and was immediately aliquoted into three Eppendorf tubes containing EDTA to avoid cell clumping. Peripheral blood samples (5 ml) were taken from these patients immediately after paracentesis. PB samples were also taken from 12 healthy volunteers.

Immunofluorescent staining and flow cytometry

The AH samples were centrifuged for 5 min at 300 g at 4°C. Supernatants were aspirated and kept at −70°C for subsequent ELISAs. The cells were aliquoted into three separate tubes and washed twice with PBS and then resuspended in a final volume of 15–20 μl PBS. Triple staining was performed in the three tubes using only two different combinations in addition to a negative control tube due to the small number of cells in the samples: CD3/CD14/CD19 and CD4/CD8/CD25 with directly conjugated labelled MoAbs, and one tube with isotype-matched control antibodies (Table 1). In brief, cells were incubated with MoAbs for 45–60 min in the dark at 4°C. Cells were then washed twice with PBS and resuspended in 15–20 μl PBS, after which 500 μl FacsFix (Becton Dickinson, Oxford, UK) were added. Cells were then stored in the dark at 4°C for a minimum period of 60 min before data acquisition. In addition, three-colour flow cytometry was performed on the AH of two patients with FHC to ascertain the percentage of CD3⁻/CD8⁺/CD16⁺ cells (natural killer (NK) cells) in the AH of these patients.

Table 1.

Monoclonal antibodies* used for three-colour flow cytometry

In parallel, 100 μl of anticoagulated blood were added to 4 ml of lysis buffer (FACS lysis solution from Becton Dickinson Immunocytometry Systems). This was then allowed to stand for 10–20 min at room temperature to facilitate complete lysis of the erythrocytes. The washing procedure and staining were performed as described above for the AH.

Three-colour immunofluorescence was analysed using the FACScan flow cytometer (Becton Dickinson) equipped with a 15-mW argon laser, and filter settings for FITC (530 nm), PE (585 nm), and peridin chlorophyll protein (PerCP) emitting in the deep red (> 650 nm) were used. At least 2000 cells in the AH and a minimum of 5000 cells in the blood were analysed using Lysis II software. Only live cells were gated for cell size by forward scatter and granularity by side scatter, and a significant number of dead cells was not seen.

ELISA

Cytokines in the AH of only nine patients with FHC and IAU were quantified using sandwich ELISA techniques (R&D Systems Europe, Ltd., Abingdon, UK) due to an insufficient amount in the rest of the patients.

The concentrations of the various cytokines detected were in pg/ml, with the following minimum detection levels as determined by the manufacturers: IL-4 3·0 pg/ml, IL-10 1·5 pg/ml, IL-12 3·0 pg/ml and interferon-gamma (IFN-γ) 3·0 pg/ml.

Data presentation and statistical analysis

Data are presented as mean (s.d.); median and 90% confidence interval (CI 90). The non-parametric Mann–Whitney U-test was performed comparing the cellular phenotypes between FHC and IAU. The χ² test was performed to compare the proportion of patients with ‘high’ and ‘low’ levels of the various cytokines between the two clinical entities, after stratification of the data using the 75th centile of the pooled data from the two groups of patients as the cut-off point. P < 0·05 was considered significant in both statistical tests. Pearson correlation coefficient was used to ascertain the strength of association between the level of cytokines and the cellular phenotypes in the AH in both patient groups. Minitab 10·5 for Windows software was used for statistical analysis.

RESULTS

Cellular phenotypes (Table 2)

Table 2.

Percentage expression of cellular phenotypes in the AH and peripheral blood (PB) of patients with FHC and IAU

The AH and PB of 10 FHC (age range 18–69 years, mean 30·7 years) and 18 IAU (age range 21–67 years, mean 45 years) patients were analysed. In addition, the PB of 12 healthy volunteers was analysed as a comparison. There was no significant difference in the cellular phenotypes of the PB between patients and control patients.

There was a predominance of T cells (CD3) compared with B cells (CD19) in the AH of all the patients. The percentage expression of CD3 was significantly higher in the AH compared with PB in both FHC and IAU (P = 0·042 and P = 0·015, respectively). CD4 expression in the AH of the IAU patients (55·7 ± 24·2% (mean ± s.d.); 60·3% (45·5–66·0%) median (CI 90)) was significantly higher than in the AH of FHC patients (29·4 ± 18·0%; 28·4% (18·2–14·5%) (P = 0·01)). In addition, there was a selective increase in CD4 expression in the AH (55·7 ± 24·2%; 60·3% (45·5–66·0%)) compared with PB (29·5 ± 15·8%; 33·5% (18·2–39·8%), P = 0·005) only in IAU and not in FHC. Similarly, the levels of double-positive CD4/CD25 were only significantly raised in the AH (5·8 ± 3·7%; 5·9% (4·2–7·3%)) compared with PB (0·9 ± 1·3%; 0·2% (0·3–1·4%) P = 0·0001) in IAU but not in FHC.

Interestingly, the percentages of CD8⁺ T cells in the AH were significantly higher in FHC (46·8 ± 16·6%; 47% (37·7–57·4%)) compared with IAU (21·9 ± 14·3%; 19% (8·6–28·3%) P = 0·003). In addition, CD8 expression in the AH (46·8 ± 16·6%; 47% (37·7–57·4%)) was significantly higher than in PB (24·5 ± 11·2%; 27·3% (15·8–33·1%), P = 0·005) in FHC, suggesting a selective increase in CD8⁺ T cells in the AH in this patient group. In contrast, this was not observed in IAU. The percentage of CD3⁻/CD8⁺/CD16⁺ in the AH of the two FHC patients was 0·25 ± 0·07%; 0·25% (0–0·6%).

The expression of CD19 (a marker for B cells) was very significantly lower in the AH compared with PB in both FHC (P = 0·003) and IAU (P = 0·000 01) patients. The expression of CD14 (a marker for monocyte/macrophage) was significantly higher in the AH in IAU (9·0 ± 5·8%; 8·8% (3·7–14·2%)) compared with FHC (2·2 ± 3·0%; 1·1% (0·4–4·0%) P = 0·015). In addition, CD14 expression in the AH in FHC was significantly lower compared with PB (5·4 ± 5%; 5·1% (2·6–8·2%) P = 0·02)).

Cytokines

The levels of the following cytokines were measured in pg/ml in the AH of both patient groups: IL-4, IL-10, IL-12 and IFN-γ (Table 3). The cytokine results were stratified to high and low value using the 75th centile as a reference point (Table 4). A χ² test was performed comparing the proportion of patients with ‘high’ level of the cytokines between FHC and IAU. The proportion of patients with ‘high’ level of IFN-γ was higher in FHC compared with IAU (P = 0·02). However, the proportion of patients with ‘high’ level of IL-12 was significantly higher in IAU compared with FHC (P = 0·02). There was no significant difference in the proportion of patients with ‘high’ level of IL-10 in the AH between FHC and IAU. IL-4 was not detected in either patient group.

Table 3.

Concentration of cytokines in the AH of patients with FHC and IAU

Table 4.

Levels of cytokines in the AH stratified into ‘high’ and ‘low’ groups

There was a positive correlation between CD8 expression and the level of IFN-γ in the AH of patients with FHC with a correlation coefficient, r = 0·68.

DISCUSSION

The results of this study demonstrate significant differences in cellular phenotypes and cytokines between the AH of patients with FHC and IAU. This is in complete contrast to the findings in iris biopsies, which showed no specific histological changes [4–11]. The importance of T cells in the immunopathogenesis of anterior uveitis [12–15] is further confirmed by their predominance in the AH in both patient groups.

Comparison of AH and PB revealed a markedly high level of CD4⁺ T cells in the AH of IAU, implying a predominant role for these cells in the immunopathogenesis. The importance of these cells in the immunopathogenesis of other ocular and non-ocular autoimmune diseases has been reported [16–20]. The percentage of CD4⁺ T cells in the AH was significantly higher in IAU compared with FHC. In addition, the percentage of activated CD4⁺ T cells was significantly higher in the AH compared with PB in IAU, but not in FHC. Hence, their predominance in the AH of IAU patients compared with FHC patients may explain the more aggressive clinical course and worse outcome of this disease in comparison with FHC.

The selective elevation of CD8⁺ T cells in FHC is interesting. In this study it has not been possible to confirm whether these CD8⁺ T cells have a cytotoxic or suppresser function, because FACS analysis defines phenotype and not the function of these cells. Apart from the well known subdivision of CD4⁺ helper T cells into Th1 and Th2 based on their cytokine production, CD8⁺ T cells have also been shown to be subdivided on the basis of cytokine expression, many producing a spectrum of cytokines including Th1- and Th2-like cytokines [21–25]. It is possible that the CD8⁺ T cells which are predominant in the AH of FHC patients may be producing both IFN-γ and IL-10, which are detected at a higher level in the AH of FHC compared with IAU, even though it was only statistically significant in the former. CD8⁺ T cells have been shown to produce both of these cytokines in a study by Hoiden & Moller [26]. The possibility of the CD8⁺ T cells being NK cells (CD3⁻/CD8⁺/CD16⁺) has been excluded by flow cytometry of AH from other FHC patients, in whom only a very small percentage of these cells has been found.

The lack of response of the ocular inflammation in FHC to topical steroids is an intriguing feature of this condition. Systemic steroids have long been known to alter immune responses by affecting cellular traffic and function, and have been shown to cause a reduction in the percentage of CD8⁺ T cells in the blood and cerebrospinal fluid [27–29]. None of the patients in the study were on systemic steroids, but some of the IAU patients were on topical steroids. Topical steroids are not known to affect the lymphocyte subset percentages, unlike the systemic steroids which are known to have such an effect. The predominance of CD8⁺ T cells in the AH of FHC patients may imply the possibility of a viral aetiology. However, using polymerase chain reaction (PCR) with specific oligonucleotide primers, Epstein–Barr, cytomegalovirus, herpessimplex and varicella-zoster virus were not detected in the AH of 20 FHC patients [30]. This of course does not rule out a viral aetiology, as it could be associated with other viruses.

The immune response is affected by steroids via its suppressive effect on a variety of cell types, including activated macrophages, interference with antigen-presenting cell (APC) function and a reduction in the expression of MHC antigen. The macrophage, being an important APC, was found to be significantly lower in the AH of FHC compared with IAU, as evidenced by the lower CD14 expression. Both the macrophages and CD4⁺ T cells were significantly lower in the AH of FHC in contrast to IAU, where there was a selective increase of CD4⁺ T cells in the AH compared with PB. These findings might explain the steroid non-responsiveness of FHC, as the helper T cells and macrophages which are the predominant target cells of the corticosteroids appear to have less important roles compared with IAU.

It is well established that T lymphocytes mediate their immune function through the secretion of soluble mediators, namely cytokines, which are known to be pleiotropic in their function. Mossmann and associates have defined Th1 and Th2 subpopulations of mouse CD4⁺ T cells based on the array of cytokines produced by the individual T cell clones [31, 32]. The cytokines produced by Th1-like cells are IFN-γ, IL-2 and tumour necrosis factor-beta (TNF-β), whereas Th2 cells are known to produce IL-3, IL-4, IL-5 and IL-10 in the mouse. However, in humans IL-10 is produced by Th1, Th2, Th0 as well as mononuclear phagocytic cells, cytotoxic T cells, B lymphocytes and mast cells [33–38].

IL-10 was initially known as the cytokine synthesis inhibitory factor produced by mouse Th2 cells that could inhibit production of a number of cytokines, especially IFN-γ by Th1 cells responding to antigen in the presence of APC [39]. Its ability to block activation of cytokine synthesis and several accessory cell functions of macrophage renders this cytokine a potent suppresser of the effector functions of macrophages, T cells and NK cells. As a result of these actions, IL-10 has a down-regulatory effect on cell-mediated immunity, hence inflammation. This ‘protective cytokine’ was detected at a ‘high’ level in higher numbers of patients in FHC compared with IAU, but this was not statistically significant.

IL-12 was also detected in the AH of the patients. This non-T cell cytokine, also known as natural killer cell stimulatory factor (NKSF) or cytotoxic lymphocyte maturation factor (CLMF), has been found by many workers to be involved in the generation and maturation of Th1 cells [40–42]. In addition it has been found to facilitate the generation of cytotoxic T cells and lymphokine-activated killer cells [43]. The central role this cytokine plays in inflammation and immune response has been well recognized by inducing production of IFN-γ and TNF-β by resting and activated T and NK cells [44]. This cytokine is made by many cell types, but particularly by macrophages, and to a lesser extent by accessory cells and B cells. In this study there was a significantly higher proportion of FHC patients with ‘low’ levels of IL-12 compared with IAU patients. This is consistent with the significantly low percentage of macrophages, which is one of the main producers of this cytokine, in the AH of FHC. This low level of IL-12 may result in reduced cell-mediated immunity, which could be an additional factor contributing to the low-grade inflammation in FHC patients. To our knowledge this is the first study of the AH where IL-12 has been quantified, so comparison with other situations is not possible.

The relationship between IL-10 and IL-12 is interesting. It has been demonstrated by in vitro and in vivo studies that IL-10 down-regulates IL-12 and paradoxically IL-12 primes T cells for high production of IL-10 [45]. This provides a negative feedback mechanism for IL-12 production. In FHC the ratio of mean IL-10 to IL-12 was 21:1, compared with IAU where the ratio was 1:1. The combined lower level of IL-12 and higher level of IL-10 may contribute to reduced cell-mediated immunity and might explain the low-grade inflammation.

Another interesting finding in the cytokine assay is the significantly higher proportion of patients with FHC with ‘high’ levels of IFN-γ, which is the strongest activator of macrophages compared with IAU. In addition, there was a positive correlation between the percentage of CD8⁺ T cells and the detected level of IFN-γ in the AH of patients with FHC, which is consistent with the fact that the CD8⁺ T cell is one of the main producers of IFN-γ apart from NK and Th1 cells. As mentioned above, IL-10 is known to inhibit IFN-γ production by Th1 cells, resulting in its down-regulatory effect on inflammation. It is then possible that there is a balance between these ‘protective’ and ‘damaging’ cytokines, respectively, in the AH of these patients. This is well illustrated in the ratio of mean IL-10 to IFN-γ, which is 1:1 in FHC and 0·6:1 in IAU. This balance in FHC then results in a relatively benign microenvironment and could account for the benign clinical course and chronicity of this interesting ocular entity in contrast to IAU.

IL-4 was not detected in the AH samples of both patient groups. It is well known that in cytokine assays many factors can interfere with its detection. The assay may be complicated by the presence of inhibitors such as soluble cytokine receptors or receptor antagonists, or even by cytokine autoantibodies. However, the non-detection may mean that the cytokine is being consumed as soon as it is being produced. There is also the possibility that the cytokine may be degraded by proteases or other substances present in the AH. Alternatively, the level might be too low to be detected, which would suggest that this cytokine plays little if any role in the immunopathogenesis of uveitis. This also suggests that B cells have a negligible effect on the ocular inflammation, consistent with the low expression of CD19 in the AH in all the patients. One plausible reason for the non-detection of IL-4 in the AH of all the patients is that its production may be down-regulated by IFN-γ, which is known to inhibit the differentiation and effector functions of Th2 lymphocytes, leading to a dominant Th1 function which could account for the inflammation in both ocular conditions.

In conclusion, the cellular phenotypes and cytokine profile in the AH in FHC and IAU are different. FHC is a distinct entity from IAU both clinically and immunologically, in that CD8⁺ T cells appear to have an important role in its immunopathogenesis. The lower level of IL-12 and the balance between IL-10 and IFN-γ in the AH in FHC could account for the low-grade ocular inflammation in this intriguing entity.

This study has provided information as to the predominant cell types and the cytokine profiles in the AH of patients with uveitis which would be useful when planning certain immunotherapeutic modalities. The possibilities of using cytokines rather than steroids to down-regulate inflammation will be a major breakthrough, as the ill effects of steroids could be avoided.

Addendum

A copy of spreadsheet data analysis of the patients in their group comparing the AH cell phenotype, cytokine level, and nature of anterior eye disease is available to interested readers on request.