Intracellular T lymphocyte cytokine profiles in the aqueous humour of patients with uveitis and correlation with clinical phenotype

Abstract

This study aimed to investigate whether T cells in aqueous humour are different in different types of uveitis and correlate with clinical phenotype. Patients with clinically different types of uveitis, but all displaying active anterior uveitis, were phenotyped and samples of aqueous humour (AH) and peripheral blood (PB) collected. Cells from AH and PB were separated by centrifugation and by density gradient centrifugation (to obtain mononuclear cells PBMC), respectively. Cells were activated with PMA and ionomycin in the presence of Brefeldin A, stained for surface markers and intracellular cytokines, and analysed by flow cytometry. The cytokine profile was correlated with the clinical phenotype. Increased percentages of interleukin (IL)-10⁺-, but not interferon (IFN)-γ⁺ T lymphocytes were found in AH compared with PB in patients with acute anterior uveitis (AAU), FHC or chronic panuveitis (PU). There was a trend towards elevated levels of IL-10⁺ T cells in AH from patients with FHC compared with AH from acute uveitis and panuveitis patients. Increased levels of IL-10⁺ T cells in AH compared with PB were also found in samples from patients with isolated uveitis, but not those with associated systemic disease. Levels of cytokine-positive T cells were not associated with the use of topical steroids or to the severity of the anterior uveitis. While type I cytokine-producing T lymphocytes are present in AH during AU, the presence of increased proportions of IL-10⁺ T lymphocytes in AH from patients with uveitis may be indicative of an anti-inflammatory mechanism that may influence the type and course of ocular inflammation in these patients.

Introduction

Uveitis is an umbrella term for a range of disease processes characterized by inflammation inside the eye. There is a variety of types but they are divided broadly into those of non-infectious  and  those  of  infectious  aetiology.  Of  those  that are of non-infectious aetiology, the majority are immune-mediated and may or may not be associated with a systemic auto-immune disease. The disease is characterized clinically as to whether the inflammation is confined to the front of the eye [anterior uveitis (AU)] or occurs throughout the eye [panuveitis (PU)]. The inflammation may be acute, defined as lasting less than 3 months, which is predominantly anterior uveitis in type or chronic, lasting more than 3 months, which is the most common type of posterior uveitis. Anterior uveitis is most commonly acute in nature but can be chronic, and there are different types of chronic anterior uveitis defined by clinical criteria. The inflammatory response in AU can vary in severity so that in some recognized clinical phenotypes it is mild (e.g. Fuchs’ heterochromic cyclitis (FHC) [1] whereas in others, such as that associated with HLA B27 and its associated diseases [2], it can be very severe. In most patients AU is treated successfully with topical corticosteroids, but in FHC there is no response to topical steroids, suggesting that there is something different about the cells in this disorder. In PU the anterior uveitis may require topical steroid treatment, but the whole eye may also need treatment with periocular or systemic steroids to control the posterior inflammatory process to prevent visual loss. Visual loss is more common in posterior uveitis than anterior uveitis.

The influx of leucocytes into the anterior chamber in AU is characterized by a predominance of T lymphocytes [3,4]. Previous studies have shown that in isolated idiopathic AU (i.e. no evidence of an associated systemic disease process) there is a higher proportion of CD4⁺ to CD8⁺ T lymphocytes, whereas in FHC CD8⁺ cells predominate [5]. Studies have shown that proportions of T lymphocytes in AH are higher than in peripheral blood [5,6], indicating selective entry of T cells into the anterior chamber. Furthermore, activated CD4⁺ T lymphocytes are increased in AH from patients with idiopathic AU compared with patients with additional systemic disease [4,7]. In patients with intermediate uveitis, T cells from the vitreous in the posterior segment of the eye were shown to be very similar in type to those seen in AU, suggesting that similar cells infiltrate the anterior and posterior segments in intermediate uveitis and that cells found in the AH reflect those in the vitreous [8].

It is known that activation of CD4⁺ T lymphocytes induces these cells to produce the cytokines interferon (IFN)-γ, interleukin (IL)-2, IL-4 and IL-10 [9,10]. The pattern of cytokines produced by T cells has enabled classification of subsets based on their cytokine profile [10,11]. Whereas IFN-γ and IL-2 are Th1-type cytokines that have proinflammatory effects, IL-10 is thought to have a role in regulation of Th1 responses [12], and several studies have reported that IL-10 can be produced by Th1, Th2 and Th0-like human T cell clones [9,13,14]. Increased levels of the proinflammatory cytokines IFN-γ and IL-2 are found in AH from AU patients [15], suggesting that the ocular inflammation is a predominantly type I response. Previous work has shown that levels of cytokines in AH samples from patients with FHC differed from those from patients with intermediate uveitis [5], and that T cell lines from the vitreous of these types of patients produce more IFN-γ, IL-2 and IL-10 than T cell lines from PB [8]. This would suggest that the T cell and cytokine profiles found in ocular fluids may be characteristic of the type of presenting uveitis. Because activated T cells are a significant feature of all types of active uveitis, it is likely that the cytokine profile of these T cells has a role in determining the clinical phenotype rather than the levels of cytokines as measured by enzyme-linked immunosorbent assay (ELISA), which for many cytokines such as IL-10 is the result of production by many types of cells. By using intracellular flow cytometrical techniques, it is possible to detect the intracellular production of cytokines by specific cell types [16].

To determine whether T lymphocytes in AH produce different cytokine profiles in different types of uveitis, we examined intracellular production of two cytokines, IFN-γ and IL-10, by T cells from the AH of patients with AU. Being limited by the relatively small numbers of T lymphocytes available in AH, IFN-γ and IL-10 were chosen as their production by T cells represents a proinflammatory and regulatory response, respectively. We also compared cytokine profiles of AH T lymphocytes with T cells from paired peripheral blood samples. The study was designed to determine whether the cytokine profiles of individual T lymphocytes infiltrating the eye and present in the aqueous humour are correlated with the clinical phenotype of the uveitis. In addition, we aimed to determine whether the cytokines secreted by these T cells (as measured by intracellular positivity) differed in acute versus chronic disease, with severity of the anterior uveitis as judged by the number of cells seen in the AH and with the duration of topical treatment. A greater understanding of the role of T cells in determining disease phenotype and severity will allow new therapeutic regimes, rather than topical corticosteroids with their side effects, to be devised which can change these profiles to help disease resolution.

Materials and methods

Patients

Patients with active uveitis (defined as the presence of cells) were recruited to the study and phenotyped on clinical grounds into anterior uveitis (AU) and panuveitis (PU). Patients with AU were subdivided further into those with FHC and those with acute anterior uveitis (AAU group). Patients with chronic AU not of the FHC type were not recruited, as most of them did not have enough cells in the anterior chamber.

Samples of AH were taken from a total of 49 patients with active AU, 20 with acute anterior uveitis (AAU), six with Fuch's heterochromic cyclitis (FHC) and 23 with chronic panuveitis (PU).

Five of the patients in the AAU group and four in the PU group had associated systemic inflammation, three with Behçet's disease, three with sarcoidosis and three with ankylosing spondylitis. Eleven of the 20 patients with AAU, none of the FHC patients and 12 of the 23 patients with PU were receiving topical corticosteroids, with topical treatment unknown for three patients. At the time of sampling none of the patients in any group were receiving or had received in the last 3 months systemic steroids or other immunosuppressive treatment. The anterior chamber of each patient was examined by slit-lamp biomicroscopy and the cellular infiltration graded 1–4 according to a standard grading system [17]. The study was approved by Moorfields Eye Hospital Ethics committee, and after informed consent, volumes of 50–150 µl AH were obtained from patients by paracentesis with an insulin syringe through the limbus. Peripheral blood (PB) samples from patients were collected into heparinized tubes at the time of paracentesis. AH samples from normal healthy controls, or from previously inflamed eyes that were now quiet, were not included as no cells can be detected [7].

Cell preparation and stimulation

Aqueous samples were aliquoted into microfuge tubes (Greiner Labortechnik, Glos, UK) coated with 50 m m ethyline diamine tetra-acetic acid (EDTA) (BDH, Leics, UK) to minimize cell clumping, then centrifuged for 5 min at 600 g and the supernatant removed and stored at − 80°C. The pelleted cells were washed twice with 200 µl of RPMI-1640 supplemented with glutamax-1 and 25 m m Hepes (Life Technologies, Paisley, UK), and with 50 µg/ml gentamycin (Sigma, Poole, UK). In those aqueous samples in which isolated  leucocytes  were  counted,  cell  numbers  ranged  from 0·8 to 8·0 × 10³. Aqueous cells were resuspended in 100 µl RPMI-1640 supplemented with 10% fetal calf serum (FCS) (Labtech International, E. Sussex, UK), in a single well of a round-bottomed 96-well plate (Life Technologies). Peripheral blood mononuclear cells (PBMC) were isolated by density gradient centrifugation on Histopaque (Sigma). After washing,  the  PBMC  were  adjusted  to  a  concentration  of 10⁵/ml, and 100 µl, containing 10⁴ cells, was added to a single well to obtain cell numbers similar to those in AH.

Aqueous cells and PBMC were stimulated for cytokine production with 50 ng/ml phorbol myristate acetate (PMA) (Sigma) and 1 µg/ml ionomycin (Sigma) for 17 h at 37°C in a 5% CO₂ atmosphere in the presence of 20 µg/ml Brefeldin A (BFA) (Sigma). This reagent disrupts intracellular protein transport [18] causing cytokines to accumulate in the Golgi apparatus, producing an enhanced cytokine signal that can be detected by flow cytometry. Unstimulated PBMC incubated with BFA were used as controls.

Intracellular cytokine staining and flow cytometry

After stimulation, AH and PBMC were transferred to microcentrifuge tubes. The wells were washed with RPMI-1640, which was added to the tubes. Leucocytes were pelleted by centrifugation for 5 min at 600 g, then resuspended in 50 µl RPMI-1640 and stained both for surface antigens and for intracellular cytokines. Briefly, the cell suspensions were labelled with fluorochrome-conjugated antibodies against cell surface antigens (CD3-PerCP, Becton Dickinson, Oxford, UK) for 15 min at room temperature in the dark. Following a wash with 200 µl FACSFlow (Becton Dickinson), the cell pellet was resuspended in 100 µl Cytofix/Cytoperm  reagent  (Becton  Dickinson)  and  incubated  for 20 min at 4°C. After washing with Perm/Wash buffer (Becton Dickinson) cells were pelleted, resuspended in 50 µl Perm/Wash buffer, then labelled for 30 min at 4°C with fluorochrome-conjugated anticytokine antibodies. Antihuman IFN-γ-FITC plus antihuman IL-10-PE were used in combination, and antihuman IL-2-PE was also used (fluorochrome-conjugated anti-IFN-γ, -IL-2 and -IL-10 antibodies were purchased from Becton Dickinson).

Following washing with Perm/Wash buffer, the cells were resuspended in 200 µl FACSFlow for flow cytometric analysis. Controls included cells labelled with isotype-matched irrelevant MoAbs, and non-activated PBMCs stained with anticytokine antibodies. To ensure specificity of the staining procedure, binding of each MoAb was blocked by preincubation with an excess of recombinant cytokine (IL-2, IL-10 and IFN-γ, Peprotech Ltd, UK). Cells from AH samples and PB were analysed using a flow cytometer (FACScan, Becton Dickinson) with CellQuest software (Becton Dickinson). Analysis gates were set on lymphocytes according to forward and side scatter, and on CD3⁺ cells.

Data presentation and statistical analysis

The results of flow cytometric analysis of cells from AH and PB are presented as the median and mean ± 1 standard deviation of the percentages of cytokine-positive CD3⁺ T lymphocytes. Data are plotted as the mean ± standard deviation of the percentages of cytokine-positive cells. Differences in the mean percentages of cytokine-positive cells or cytokine levels between patient groups were compared using a Mann–Whitney U-test for unpaired data or a Wilcoxon signed rank test for paired non-parametric data.

Results

The percentages of cytokine-positive T lymphocytes in AH of patients was determined by flow cytometry and compared with those in paired PB samples. Taking the results of all AH samples together, the percentage of IL-10⁺ T lymphocytes in AH (mean 4·7% ± 10·0, median 2·1%) was higher than that in PB (mean 1·5% ± 1·2, median 1·2%, P = 0·0001). Table 1 shows that this difference between AH and PB was found with samples from patients within either the AAU (P = 0·04), FHC (P = 0·03) or PU groups (P = 0·03). By contrast, there was no difference between the percentage of IFN-γ⁺ T lymphocytes in AH and PB either overall, or within any of the patient groups.

Table 1.

Percentages of IFN-γ⁺ and IL-10⁺ T lymphocytes in AH and PB from patients with AAU, FHC and PU.

	% T lymphocytes
	AH mean ± s.d. (median)	PB mean ± s.d. (median)
IL-10+
All patients	4·7 ± 10·0 (2·1)	1·5 ± 1·2 (1·2)
AAU	2·6 ± 2·7 (2·1)	1·6 ± 1·2 (1·3)
FHC	6·6 ± 4·6 (6·3)	1·1 ± 0·9 (1·1)
PU	6·0 ± 14·1 (1·7)	1·5 ± 1·3 (1·1)
IFN-γ+
All patients	24·4 ± 22·9 (18·0)	16·5 ± 10·5 (12·9)
AAU	23·7 ± 25·7 (13·2)	14·6 ± 10·5 (10·9)
FHC	28·6 ± 21·7 (32·1)	14·2 ± 9·0 (14·3)
PU	23·8 ± 21·5 (19·5)	18·6 ± 10·8 (15·0)

The percentages of IL-10⁺ T lymphocytes in AH from patients with FHC (mean 6·6%, median 6·3%) were found to be higher than those of AH from patients with AAU compared with mean 2·6%, median 2·1%) although this difference did not quite reach statistical significance (P = 0·051). The mean value of IL-10⁺ T cells in the AH of patients with PU (6·0%) was similar to that of samples from FHC patients (6·6%), a result likely to be due to one unusually high value of 69% in the PU group. However, the median values differed (FHC = 6·3%, PU = 1·7%), suggesting that AH samples from patients with FHC are likely to contain higher levels of IL-10⁺ T cells than those from either the majority of samples from AAU or PU patients. Such differences were not reflected in the levels of IFN-γ⁺ T cells, or in the percentages of cytokine-positive T cells in PB samples (data not shown). The variability in the frequency of IFN-γ⁺ T cells was higher than that of the IL-10⁺ T cells. This could be due to the fact that some patients had a higher baseline expression of IFN-γ⁺ T cells due to their disease being active and stimulation with PMA failing to augment fully or any further the number of cytokine-positive T cells. However, it was not possible to assess the variability in the levels of IFN-γ⁺ T cells in unstimulated AH cells due to the small numbers of T cells present in AH samples.

The pattern of an increased level of IL-10⁺ T lymphocytes in AH compared with PB was also found in the group of patients with isolated uveitis (AH mean 5·1% ± 10·9, median 2·4%, PB mean 1·6% ± 1·3, median 1·2%, P = 0·001), although not in the group of patients with uveitis and associated systemic disease (P = 0·15) (Fig. 1). There was a suggestion of an increased level of IFN-γ⁺ T lymphocytes in AH (mean 24·9% ± 22·6, median 18·3%) compared with PB (mean 16·0% ± 10·1, median 12·1%) of patients with isolated uveitis, but this difference did not reach significance (P = 0·08). When levels of cytokine-positive cells in AH or in PB from these two patient groups were compared, no differences were found. In most cases, the percentage of IFN-γ⁺/IL-10⁺ double-positive T cells was between <1–2%, with the exception of one AH sample from a patient with PU that contained an unusually high level (69%) of IL-10⁺ T cells and 14% of IFN-γ⁺/IL-10⁺ T cells.

Fig. 1.

AH and PB cells from patients with AU were stimulated for 17 h with PMA and ionomycin then stained with fluorescently labelled antibodies to IFN-γ and IL-10. The bars and error bars represent the mean percentage ± 1 standard deviation, respectively. The results are shown for IL-10⁺ T cells only and are for AH and PB from patients with isolated uveitis (solid bars, n = 35) or with anterior uveitis associated with systemic disease (diagonal striped bars, n = 8). *The level of IL-10⁺ T cells was higher in AH than in PB in patients with isolated uveitis (P = 0·001).

Samples of AH from patients who were receiving topical treatment at the time of sampling were compared with those who were off treatment (Fig. 2). No differences were found between the two patient groups in the percentages of either IL-10⁺ or IFN-γ⁺ T lymphocytes in AH (IL-10: P = 0·075, IFN-γ: P = 0·61). However, it was noted that most of the samples with the highest percentages of IL-10⁺ T lymphocytes were from patients not receiving treatment and that these samples came from each of the patient groups (two FHC samples, two PU and one AAU).

Fig. 2.

AH cells were stimulated for 17 h with PMA and ionomycin then stained with fluorescently labelled antibodies to IFN-γ and IL-10. The percentages of IL-10⁺ T lymphocytes in individual AH samples are shown. Patients either had no topical treatment prior to aqueous tap, 0–14 days or >14 days topical treatment, and results are shown for patients in the following groups: ⋄ = AAU; ◊ = FHC; × = PU.

In a comparison of cytokine-positive cells in AH from patients with mild anterior uveitis (grade 2+) with those from patients with more severe anterior uveitis (grades 3+, 4+ or 5+), there were no differences in the percentages of either IL-10⁺ or IFN-γ⁺ T lymphocytes (data not shown). This would indicate that the levels of cytokine-positive cells in AH are not related to the clinically determined grade of severity of anterior uveitis.

Discussion

In this study we have shown that the entry of increased numbers of IL-10- but not IFN-γ-producing T cells in AH occurs in AAU, FHC and PU, with increased levels of IL-10⁺ lymphocytes in the AH of patients with either acute or chronic uveitis. This pattern was also found in patients with isolated uveitis, but not those with associated systemic disease. There was no association between the chronicity or severity of AU and the number of either IL-10- or IFN-γ-producing T cells in the AH.

Elevated  concentrations  of  IFN-γ in  AH  samples  have been reported in previous studies [15,19], and IFN-γ⁺ lymphocytes in the AH together with natural killer (NK) cells (which have been reported in small numbers in the AH of uveitis patients [6]) are likely to be a source of this cytokine. Therefore, we might have expected to find increased numbers of IFN-γ⁺ T cells in AH compared with PB. However, the proportions of IFN-γ⁺ T cells in both AH and PB in this study were not found to be significantly different, although a number of AH samples contained a relatively high proportion of IFN-γ⁺ T cells compared with PB, suggesting that in certain individuals there exists the possibility of a selective entry of Th1 or Tc1 cells into the anterior chamber.

There was no tendency for the AH from patients receiving topical treatment prior to aqueous sampling to contain the lowest numbers of IFN-γ⁺ T lymphocytes. This observation suggests that although the immunosuppressive activity of topical corticosteroids is effective in decreasing the production of IFN-γ and other cytokines by activated T cells [20], there is not necessarily a reduction in the infiltration of IFN-γ⁺ T cells. However, cytokine production by T lymphocytes in vivo is genetically determined so that high or low levels of cytokines are produced to a given stimulus depending on the polymorphism present [21,22] and it could also be influenced by other anti-inflammatory factors in AH such as transforming growth factor (TGF)-β[7,23].

In contrast with IFN-γ⁺ T cells, percentages of IL-10⁺ T cells in samples of AH from patients with active AU were much lower, although increased compared with PB. Whereas the proportions of IFN-γ⁺ T cells in AH were found to be similar to PB in samples from all patient groups, the investigation of IL-10⁺ T cells in AH produced different findings. It was found that compared with PB, the percentages of IL-10⁺ T cells were raised in the AH in all groups of patients, AAU, FHC and PU. The results suggest that although the presentation of AU may be either acute or chronic, in most cases there is still an increased level of IL-10⁺ T cells in AH compared with PB. The results also indicated that the percentages of IL-10⁺ T cells in AH samples from patients with FHC could be raised in comparison with AH from patients with AAU, which would correlate with the clinical picture of FHC presenting with a milder uveitis [1]. Most patients with FHC have less AH cells than a clinical grading of 2 + and cannot be recruited to the study, as such samples do not contain enough cells for staining. It is likely, therefore, that the AH samples that were studied came from patients with more active disease than the majority of patients who have much milder disease. If we were able to study the patients with milder disease, it is possible that the proportion of IL-10⁺ T cells may be considerably greater than those with more active disease. Our results suggest that the balance of IL-10⁺ and IFN-γ⁺ T cells may be different depending on the clinical presentation of anterior uveitis, and whether the uveitis affects other parts of the eye in addition to just the anterior segment. We also found that there were increased levels of IL-10⁺ T cells in AH compared with PB of patients with isolated uveitis but not those with associated systemic disease. The fact that increased levels of IL-10⁺ T cells were found in AH but not PB of patients with isolated uveitis may reflect the findings of previous studies in which increased numbers of activated CD4 T lymphocytes have been found in the AH of patients with idiopathic uveitis [7].

The current findings relate to those of a previous study [5], which reported that IL-10 levels measured in AH samples from patients with FHC tended to be higher that those with idiopathic anterior uveitis. Although IL-10⁺ T cells may secrete IL-10 into the AH during AU, the presence of increased numbers of IL-10⁺ T cells in the AH may not always have an anti-inflammatory effect within the eye, as the outcome within any environment depends not only on the T cells present, but also on the profile of cytokines secreted by other cell types [14] and the genetic make-up of the individual [21,22]. However, the finding of a relatively higher proportion of IL-10⁺ T cells in the AH of most of the patients with FHC compared to those with AAU, and of similar levels of AH IFN-γ⁺ T cells between these patient groups, would suggest that IL-10 T cells have a modulatory effect and are involved in disease chronicity in certain presentations of AU.

For ethical reasons, we could investigate samples of AH only once during the course of each patient's disease, and while our results show that percentages of IFN-γ⁺ T lymphocytes are similar in AH and in PB in all patients, the proportion of IL-10⁺ T cells is increased in AH from patients with chronic as well as acute AU. Topical application of corticosteroid is used to reduce the inflammatory response in acute AU, but it is not known whether this treatment targets specific T cell subsets. Because long-term steroid use may lead to sight-threatening complications such as cataract, glaucoma and herpetic keratitis, alternative approaches to therapy would be valuable and it has been suggested that IL-10 has a role as a therapeutic agent for immunological diseases characterized by a Th-1-type cytokine pattern [24]. Thus, inhibition of Th1-derived cytokines and manipulation of the balance of T cell cytokine profiles present in the eye by the activation of IL-10⁺ T lymphocytes or the introduction of IL-10 as a topical agent could be considered appropriate strategies for the down-regulation of inflammatory eye disease in the future.