Regulation of T-Lymphocyte CCL3 and CCL4 Production by Retinal Pigment Epithelial Cells

Abstract

Purpose.: Retinal pigment epithelial (RPE) cells have an important role in the immune suppression associated with the immune privilege of the eye. Some aspects of this remain unclear and this study aimed to determine how RPE cells could influence the production of chemokines by T lymphocytes.

Methods.: T lymphocytes, separated from peripheral blood of normal volunteers, and RPE cells, cultured from donor eyes, were cultured separately and together, either in contact or in transwells. Supernatants were analyzed for CCL3, CCL4, and soluble CD54 (sCD54) by ELISA. Blocking agents were used to determine which soluble mediators were involved.

Results.: Coculture of RPE cells with activated lymphocytes resulted in a reduction in CCL3 and CCL4 production by lymphocytes, primarily by soluble mediators. Soluble CD54 was markedly increased on coculture of lymphocytes with RPE cells. Soluble CD54 reduced CCL3 and CCL4 production by RPE cells, and inhibition of CCL3 and CCL4 on coculture with RPE cells was reduced by anti-CD54. Blocking prostaglandin E2 (PGE2) abrogated the inhibition of CCL4, but not CCL3, by RPE cells. Blocking TGFβ and nitric oxide production had no effect.

Conclusions.: RPE cells are able to down-regulate high levels of CCL3 and CCL4 production by T lymphocytes by using the soluble mediators sCD54 and PGE2. Reducing this production of CCL3 and CCL4 will dampen down the cascade effect and recruitment of more inflammatory cells, protecting the retina from an excessive immune response.

Introduction

Retinal pigment epithelial (RPE) cells, which have key roles in the maintenance of photoreceptor cells and as part of the blood–retina barrier, have long been known to be an important component of the immune suppression associated with the immune privilege of the eye. ¹ Although some of the ways in which RPE cells can act in this role have been identified, the extent of their action and the mechanisms involved are still sketchy. With the move to use RPE cells therapeutically, a better understanding of this RPE function will be important. 

The immunosuppressive effects of RPE cells vary from the production of TGFβ, a known immunosuppressive cytokine ² to the inhibition of lymphocyte proliferation. Early studies with rat RPE cells showed that they could suppress lymphocyte proliferation in response to antigen, mitogen, and IL-2, and this suppression was due in part to production of prostaglandin E2 (PGE2) ³ and nitric oxide (NO). ⁴ More recently, Gregerson et al. ⁵ have shown that immortalized murine RPE cells could present peptide via MHC class II to naive T cells, which led to the induction of anergy in these T cells but not an immunoregulatory phenotype. These immortalized RPE cells were also able to inhibit proliferation and IL-2 and IFNγ production by ConA-activated lymph node cells, regardless of whether RPE cells were in direct contact or in split wells or if RPE cell supernatant alone was used, indicating the role of soluble factors consistent with other studies. ^3,6 In studies on activated human T cells, RPE cell coculture also inhibited proliferation and both IL2 secretion and induction of IL2R as well as CD71 and cyclin A. ^7,8 It will be important for RPE cells to modulate the T cells they encounter because nonspecifically activated T cells are able to cross a normal noninflamed blood–retina barrier and are able to initiate limited and transient breakdown of the blood–retina barrier. ⁹  

The production of chemokines, small inducible chemoattractants, by T cells is key to the establishment of an inflammatory response in the retina; for example, in posterior intraocular inflammation (PII), autoimmune posterior uveitis. Chemokines are involved in a variety of immune and inflammatory responses and are particularly important in providing specific signals for leukocyte migration. ¹⁰ Chemokine production by activated T cells results in the recruitment of other cells such as effector macrophages, which are predominantly responsible for sight-threatening damage to the retina in PII. 

In this study we have chosen to investigate the influence of RPE cells on T lymphocyte production of the chemokines CCL3 (MIP-1α) and CCL4 (MIP-1β). CCL3 and CCL4 have been implicated in retinal inflammation, particularly at early T-cell–dependent stages. ^11–14 Although CCL3 and CCL4 were first purified from lipopolysaccharide-treated monocytic cell lines, T lymphocytes have been shown to be able to produce CCL3 and CCL4 in substantial quantities, particularly Th1 type lymphocytes. ¹⁵ These chemokines can be produced by T cells following T-cell receptor engagement or after nonspecific activation ¹⁶ and have been implicated in a range of inflammatory conditions ^17,18 including those with an autoimmune basis. ^19,20 However, T cells from some normal donors may also produce CCL3 and CCL4 constitutively, and T-cell clones have been established from normal donors that produce substantial levels of CCL3 and CCL4. ²¹ It is not known whether this spontaneous production is due to inherent genetic factors or whether it has an environmental basis. This production is physiologically relevant and has been shown to be a positive factor in resistance to HIV infection. ^22,23 However, since it will result in further recruitment of inflammatory cells and an amplifying cascade effect, it is likely to exacerbate autoimmune disease or be a predisposing factor in susceptible individuals. We have shown that treatment with neutralizing antibody to CCL3 can reduce the severity of experimental autoimmune uveitis, an animal model of PII. ²⁰ Regulation of CCL3 and CCL4 production by T cells infiltrating the retina is important to protect the retina from an inflammatory response, and the RPE cells are likely to be key to this regulation. 

In this study we show for the first time that RPE cells appear to be able to down-regulate high levels of CCL3 and CCL4 released into the medium by T lymphocytes and that this action is brought about by soluble mediators. Our results provide evidence that RPE cells use the soluble form of CD54 (intercellular adhesion molecule-1 [ICAM-1]) and PGE2 to bring about this reduction. The use of soluble CD54 (sCD54) is of particular interest. Its membrane-bound form is an important adhesion molecule, and sCD54 has often been associated with inflammatory disease. However, it is thought to also have an inhibitory role, acting as a competitive inhibitor of membrane-bound CD54, with overexpression of sCD54 shown to reduce inflammatory cell recruitment in vivo. ²⁴ To the best of our knowledge, this is the first time RPE cells have been shown to use sCD54 in this capacity. 

Materials and Methods

Cell Preparation

Mononuclear cells obtained from normal volunteers, according to the Declaration of Helsinki, were separated from peripheral blood using Histopaque 1077 and resuspended in RPMI containing 10 % (v/v) fetal calf serum, L-glutamine (200 μM), and 50 μM mercaptoethanol. All media, buffers, and additives used in this study were from Sigma-Aldrich Co. Ltd. (Poole, UK) unless stated otherwise. The cells were seeded into tissue culture flasks and incubated at 37°C for 1 hour to allow monocytes in the cell population to adhere to the plastic, leaving the lymphocytes in suspension. The nonadherent lymphocytes were removed, washed with Hanks' balanced salt solution (HBSS), and resuspended in serum-free RPMI for coculture experiments. Immunostaining of cytospin preparations showed the resulting cell suspension to be >95% T lymphocytes (CD3⁺) and 1 % B lymphocytes (CD20⁺). For some experiments lymphocytes were activated by incubation with phytohemagglutinin (PHA) at 2 μg/mL overnight and washed 3 times with HBSS before use in coculture experiments. 

To enhance physiological relevancy, RPE cells were isolated from normal donor eyes (obtained from Manchester Eye Bank, Manchester Royal Eye Hospital with local ethical approval) rather than using an immortal RPE cell line. The isolated RPE cells were cultured and confirmed to be 100% epithelial in nature by cytokeratin staining, as described previously, ²⁵ and by staining for RPE-65. Cytospin preparations of harvested cells were air-dried, incubated with 0.25% Triton-X 100 in Tris-buffered saline for 10 minutes, and stained with mouse monoclonal antibody to RPE-65 (Novus Biologicals, Cambridge, UK) using the Envision detection system (Dako UK Ltd., Cambridgeshire, UK) according to manufacturer instructions (Fig. 1). The cells were used at passages 3 to 6. For analysis of sCD54 production by RPE cells, cultures were washed as already described, pre-incubated in serum-free medium for 16 hours before addition of TNFα and/or IL-1β at a range of concentrations (R&D Systems Europe Ltd., Abingdon, UK) for 24 hours. Medium was harvested and centrifuged at 500g for 10 minutes, and the supernatant was removed and stored at −20°C prior to use. Protein content of cells was determined by Coomassie Blue binding. ²⁵ For the preparation of conditioned medium, RPE cells were grown to confluence in six-well plates and washed three times in HBSS, after which the medium was replaced with 1 mL serum-free RPMI for 16 hours. Supernatant was harvested as already described. 

Figure 1. 

 

Expression of RPE-65 by cultured RPE cells (passage 4). Cytospin preparations of RPE cells were stained with RPE-65 and the Envision horseradish peroxidase detection system (brown stain): (a) control, no primary antibody, (b) RPE-65. Magnification ×40.

A549 cells, human alveolar epithelial cells, and LX2 cells, human hepatic stellate cells that are liver-specific pericytes, were also used for comparison. A549 cells were cultured in RPMI 1640 containing 10% (v/v) fetal calf serum (FCS), 200 μM L-glutamine, 100 IU/mL penicillin, and 100 μg/mL streptomycin, and LX2 cells were cultured in Dulbecco's modified Eagle's medium (high glucose) containing 10% (v/v) FCS, 2 mM L-glutamine, 100 IU/mL penicillin, and 100 μg/mL streptomycin. 

Coculture

Prior to coculture with lymphocytes, confluent RPE, A549, and LX2 cell cultures were washed three times in HBSS and cultured with serum-free RPMI for 16 hours. Lymphocytes (1 × 10⁶ cells/mL) were cocultured with cell monolayers in six-well plates for 24 hours. Any effects due to mismatch of histocompatibility antigens were unlikely over this time period. In certain experiments, lymphocytes were cultured with RPE cell-conditioned medium (1 mL/well) in place of cells. RPE cells were also grown to confluence in transwell plates (24-mm diameter, 3-μm pore size; Merck Biosciences Ltd., Nottingham, UK), with lymphocytes added to the upper compartment. Antibodies or inhibitory factors were added to the coculture plates at the same time as the lymphocytes. Anti-human CD54 (10 μg/mL) and anti-human TGFβ1,β2,β3 (0.6 μg/mL), neutralizing antibodies (R&D Systems), were used at the concentrations required for neutralization as tested by the supplier. Anti-CD54 was checked for functional activity using an aggregation assay ²⁶ ; at 10 μg/mL it was shown to inhibit the formation of clusters of fresh human lymphocytes in response to PHA by 78%. The prostaglandin inhibitor, indomethacin (Sigma-Aldrich) was used at 20 μM ²⁷ and the NO inhibitor, NMMA ⁴ (NG-monomethyl-L arginine, Sigma-Aldrich) at 2 mM both concentrations selected as previously shown to result in over 50% inhibition. Soluble CD54 was added at 0.1 ng/mL. Following coculture, supernatant was harvested as previously described. 

Enzyme-Linked Immunosorbent Assay

ELISAs for CCL3, CCL4, and sCD54 used R&D Systems capture and detection antibodies according to manufacturer and Maxisorp microtiter plates (ThermoScientific Nunc, Roskilde, Denmark). Briefly, standards and test samples were incubated for 2 hours at room temperature on plates that had been coated with capture antibody. After washing, biotinylated detection antibody was added and plates were incubated for a further 2 hours at room temperature. Detection used Streptavidin-HRP (Zymed Laboratories, Inc., San Francisco, CA) followed by 3,3′,5,5′-tetramethylbenzidine (Sigma-Aldrich) substrate. The plate was read at 450 nm with background correction at 570 nm in a microplate reader. 

Flow Cytometry

Lymphocytes were stained with fluorochrome-conjugated specific antibodies to human CD3 (FITC; Caltag-Medsystems Ltd., Buckingham, UK, mouse IgG_2a), CD4 (PerCP, BD Biosciences, Oxford, UK, mouse IgG₁), and CD69 (allophycocyanin; Caltag-Medsystems, mouse IgG_2a) in 1% (w/v) BSA/PBS at 4°C for 20 minutes. In a separate experiment, apoptosis was assayed as previously described ²⁸ using a propidium iodide (PI) annexin V kit (Bender MedSystems, Vienna, Austria) according to the manufacturer's protocol. For flow cytometric analysis 10,000 events from each sample were captured (FACS Calibur; BD Biosciences) and data analyzed with CellQuest software (BD Biosciences). The lymphocyte population was selected and gated using forward scatter, which relates to size and can be used to exclude dead and aggregated cells, and side scatter, which relates to granularity. 

Data Analysis

Data were analyzed by one-way ANOVA using Tukey's multiple comparison test to compare individual columns of data. Probability values of P < 0.05 were considered significant. Analysis was carried out using Prism software (GraphPad, Inc., San Diego, CA). 

Results

T-Lymphocyte Activation Is Associated with CCL3 and CCL4 Production

T lymphocytes were prepared from normal human donors (>95% CD3⁺), and their activation state was determined by CD69 expression. Association of CD69 expression with CCL3 and CCL4 production was examined (Table). Amounts of CCL3 or CCL4 produced by different donors varied, but higher CCL3 production corresponded to higher CCL4 (r ² = 0.7840). CD3⁺ CD4⁺ cells were gated, and expression of CD69 was analyzed by flow cytometry. An increased number of CD69⁺ cells in the population was associated with increased CCL3 and CCL4 production (r ² = 0.9874 for CCL3, r ² = 0.9820 for CCL4) indicating that the production of CCL3 and CCL4 was linked to the degree of activation of the lymphocytes. Lymphocytes from donor 1 with the lowest activation were treated with PHA to compare with a maximal mitogenic response. This increased the percentage of CD69⁺ cells in the CD3⁺ CD4⁺ gated population markedly and increased CCL3 and CCL4 production to the level seen for unstimulated cells from donor 3. 

Table. 

 

Association of CD69 Cell Surface Expression with CCL3 and CCL4*

	% CD69 of CD3+ CD4+	CCL3, pg/mL ± SEM	CCL4, pg/mL ± SEM
Donor 1	1.7	53.7 ± 4.7	267.0 ± 41.0
Donor 2	13.7	2734.0 ± 79.2	2854.0 ± 126.0
Donor 3	29.3	5851.0 ± 130.1	7584.0 ± 129.9
Donor 1 stimulated with PHA	90.3	5935.0 ± 127.9	7448.0 ± 213.1

*  Peripheral blood mononuclear cells were prepared as described. CD3⁺ CD4⁺ cells were gated and expression of CD69 analyzed by flow cytometry. CCL3 and CCL4 production by cells from these samples was assayed by ELISA (n = 6) after 24 hours. The percentage of CD69⁺ cells in the population was associated with increased CCL3 and CCL4 production (r ² = 0.9874 for CCL3 and r ² = 0.9820 for CCL4).

RPE Cells Down-Regulate CCL3 and CCL4 Production by T Lymphocytes Using Soluble Factors

The effect of RPE cells on T lymphocyte CCL3 and CCL4 production was determined in cocultures. RPE cell cultures alone produced negligible CCL3 and CCL4 (data not shown). When RPE cells were included in lymphocyte cultures, CCL3 and CCL4 in the culture medium were significantly reduced (p < 0.001) compared to when the lymphocytes were cultured alone (Figs. 2a, 2b). CCL3 could be reduced by over 90% with unstimulated donor lymphocytes and CCL4 by 60%. An equally significant reduction (P < 0.001) was seen with RPE cells that had been cultured from eyes from different donors. For donor 3 lymphocytes, CCL3 was decreased by 8240 ± 29 pg/mL when they were cultured with RPE cells from donor 15 and by 8190 ± 51 pg/mL when cultured with RPE cells from donor 24; CCL4 by 5781 ± 57 and 5626 ± 66 pg/mL when cultured with RPE15 and RPE24, respectively (± SEM). CCL3 and CCL4 were also significantly (P < 0.05) down-regulated when lymphocytes were cocultured with A549 alveolar epithelial cells but not LX2 pericytes (Fig. 3). 

Figure 2. 

 

Effect of RPE cells on CCL3 and CCL4 production by lymphocytes. Lymphocytes were either from donor 3 or donor 1 stimulated by PHA. (a, b) Lymphocytes (lymph, 1 × 10⁶ cells/mL) were cultured either alone (clear bars) or together with RPE cells in the same well (hatched bars). (c, d) Lymphocytes were cultured with RPE-conditioned medium (med, cross-hatched bars) or with RPE cells separated from lymphocytes (upper compartment) in transwell plates (filled bars). Cells were cultured for 24 hours before supernatant was analyzed by ELISA for CCL3 (a, c) and CCL4 (b, d). RPE cells alone produced negligible CCL3 or CCL4 (not shown). Results shown are the mean of six experiments ± SEM. **P < 0.001.

Figure 3. 

 

Comparison of effect of RPE, A549, and LX2 cells on CCL3 and CCL4 production by lymphocytes. Lymphocytes were from donor 1 stimulated by PHA. Lymphocytes (lymph, 1 × 10⁶ cells/mL) were cultured either alone or together with either RPE, A549, or LX2 cells in the same well; CCL3 (hatched bars) and CCL4 (clear bars). Results shown are the mean of three experiments ± SEM. **P < 0.005.

Production of CCL3 and CCL4 by lymphocytes cultured with RPE cells was compared to culture with conditioned medium from RPE cell cultures or with the culture of lymphocytes and RPE cells separated in a transwell culture system. CCL3 and CCL4 were also significantly down-regulated when RPE conditioned medium or transwell conditions were used (P < 0.001) indicating that a soluble factor or factors were responsible rather than a cell contact effect (Figs. 2c, 2d). 

The percentage of lymphocytes that were annexin V positive (early apoptotic) was 26.83% for lymphocytes cultured alone for 24 hours and 11.35% for lymphocytes that had been cocultured with RPE cells for 24 hours. The percentage of cells positive for annexin V and PI (late apoptotic) was 3.60% for lymphocytes cultured alone and 3.02% for lymphocytes that had been cocultured with RPE cells. These findings indicated that lymphocyte apoptosis was not responsible for the reduction in CCL3 and CCL4 seen for lymphocytes cocultured with RPE cells. 

Role of RPE-Produced Soluble Factors, NMMA, TGFβ, and PGE2 in Down-Regulation of CCL3 and CCL4 Production

Since the decrease in CCL3 and CCL4 production upon coculture with RPE cells depended on soluble factors to a greater extent than cell contact, we investigated which soluble factors produced by RPE cells might be responsible. The effect of blocking mediators produced by RPE cells that have been shown to have immunosuppressive effects was examined by using lymphocytes from donor 3. NMMA is a competitive inhibitor of NO synthase, blocking NO production. ^4,29 Indomethacin blocks PGE2 production. ²⁷ Neutralizing antibody to human TGFβ was also tested because TGFβ is produced constitutively by RPE cells. ² None of these inhibitors influenced the reduction in CCL3 (P > 0.05) seen when lymphocytes were cocultured with RPE cells (Figs. 4a–c). For CCL4, NMMA and anti-TGFβ similarly had no effect (Figs. 4d, 4e). Indomethacin, however, did return CCL4 to the levels seen for lymphocytes alone. There was a significant difference (P = 0.03) between CCL4 levels from lymphocyte/RPE cell coculture with indomethacin compared to without indomethacin (Fig. 4f). There was no significant difference between CCL4 levels from lymphocytes alone compared to levels from lymphocyte/RPE cell coculture with indomethacin (P = 0.9). Interestingly, indomethacin treatment also resulted in a significant increase (P < 0.05) in CCL3 with lymphocytes alone (Fig. 4c), but this was not the case for CCL4. 

Figure 4. 

 

CCL3 and CCL4 production by lymphocytes (donor 3) and RPE cells either alone or in coculture in response to NMMA (a, d), anti-TGFβ (b, e), or indomethacin (c, f). Cells were cultured alone or together, with (hatched bars) or without inhibitor (clear bars), for 24 hours before supernatant was analyzed by ELISA for CCL3 (a–c) and CCL4 (d– f). NMMA was at 2 mM, indomethacin at 20 μM, and anti-TGFβ1,2,3, at 0.6 μg/mL. Results shown are the mean of three experiments ± SEM. *P < 0.05.

Production of Soluble ICAM-1 (CD54) by RPE Cells

It has been shown that RPE cells can produce soluble ICAM-1, or CD54, ^30,31 and it has been suggested that this might have a suppressive effect. ²⁴ We therefore examined the production of sCD54 by our RPE cells both constitutively and in response to the pro-inflammatory cytokines IL-1β and TNFα. Constitutive production of sCD54 was confirmed, and significantly increased production was seen in response to both TNFα and IL-1β (P < 0.05) (Fig. 5a) at concentrations that are in the physiological range and had been tested previously. ³⁰ There was no synergistic effect when IL-1 and TNFα were added together. There was a substantial increase (P < 0.001) in sCD54 in the culture medium when RPE cells and lymphocytes were cocultured compared to their culture alone (Fig. 5b). 

Figure 5. 

 

Production of sCD54 by RPE cells. Soluble CD54 production in response to TNFα (1, 5, and 10 ng/mL) and IL-1β (0.5 and 5 ng/mL) alone and in combination as measured by ELISA (a). Soluble CD54 production by RPE cells and lymphocytes (donor 3) alone and cultured together as measured by ELISA (b). Results shown are the mean of three experiments ± SEM. *P < 0.05; **P < 0.001.

CD54 Is Responsible in Part for RPE Down-Regulation of Lymphocyte CCL3 and CCL4 Production

To determine whether sCD54 was responsible for the down-regulation of CCL3 and CCL4 production by activated lymphocytes, PHA-stimulated lymphocytes from donor 1 were used to ensure that activation levels were consistent. PHA was used in preference to activation with anti-CD3 in combination with phorbol myristate acetate or CD28 because it results in a similar cytokine secretion profile and is responsive to the same inhibitors but gives a less variable response. ³²  

Soluble CD54 was shown to decrease CCL3 production by PHA-treated lymphocytes in a dose-dependent manner (Fig. 6a). As can be seen, RPE cells can produce approximately 10 ng of sCD54 per 10⁶ cells over a 24-hour period after stimulation with a high concentration of pro-inflammatory cytokines. ³⁰ However, the lower concentration of 0.1 ng/mL sCD54 was chosen for further experiments because this was considered to be more representative of cell number and inflammatory status in the eye at the time of early infiltration of activated T cells. Soluble CD54 at 0.1 ng/mL significantly decreased CCL3 and CCL4 production by PHA-treated lymphocytes from donor 1 (P < 0.0001; Figs. 6b, 6c). CCL3 and CCL4 production by lymphocytes cocultured with RPE cells was decreased further by sCD54 at 0.1 ng/mL (P < 0.001; Figs. 6b, 6c). Addition of neutralizing antibody to CD54 to the lymphocyte RPE cell coculture significantly reduced the degree of inhibition of CCL3 (P < 0.0001) and CCL4 (P = 0.027) (Figs. 6b, 6c). However, it did not bring about complete abrogation of the inhibition and the reduction was less marked for CCL4 (Fig. 6c). Control antibody had no effect on the production of CCL3 or CCL4 by lymphocyte/RPE cell cocultures (Figs. 6b, 6c). Anti-CD54 or control antibody had no direct effect on CCL3 and CCL4 production by lymphocytes (data not shown). 

Figure 6. 

 

Involvement of sCD54 in CCL3 and CCL4 production by lymphocytes and lymphocyte/RPE cell coculture. Decrease in CCL3 production by PHA-stimulated lymphocytes in response to sCD54 at 0.1 and 10 ng/mL (a). CCL3 (b) and CCL4 (c) production following culture of lymphocytes (LY are donor 1 lymphocytes treated with PHA) and RPE cells either alone or together with sCD54 (0.1 ng/mL) or antibody to CD54 (aCD54, 10 μg/mL) or control antibody (10 μg/mL). CCL3 and CCL4 production measured by ELISA. Results shown are the mean of 6 experiments ± SEM. *P < 0.05; **P < 0.001; ***P < 0.0001.

Discussion

This study shows that CCL3 and CCL4 are produced by lymphocytes from some normal donors and that this is related to the activation state of the lymphocytes as determined by CD69 expression. This may reflect the immune status of the individual. It has previously been reported that immortalized CD4 and CD8 T-cell clones from some normal donors and with a normal functional phenotype spontaneously produce substantial levels of CCL3, CCL4, and CCL5, whereas clones from other normal donors may produce none. ²¹  

CCL3 and CCL4 production by T lymphocytes entering tissue will be advantageous in many inflammatory situations due to amplifying inflammatory cell recruitment and facilitating rapid and efficient clearance of a pathogen, but it will need tight regulation to avoid damaging inflammatory effects. We demonstrate for the first time that down-regulation of lymphocyte CCL3 and CCL4 by RPE cells may be a way of dampening the inflammatory reaction and protecting the vulnerable retina from a damaging autoimmune response. Importantly this may also be a more general mechanism employed by epithelial cells at sites where T cells enter tissue. 

There was no evidence to indicate that additional lymphocyte apoptosis was occurring when lymphocytes were cultured with RPE cells under these conditions. Studies by Jorgensen et al. ³³ in 1998 suggested apoptosis was occurring when activated T cells and RPE cells were cocultured and that this was due to expression of CD95L by the RPE cells. However, a study using fetal RPE cells and the Jurkat T-cell line has indicated that only IFNγ-stimulated RPE cells caused apoptosis in these T cells and this was unrelated to CD95L, which could not be detected on these cells. ³⁴  

Investigation of the mode of action of the RPE cells in reducing CCL3 and CCL4 production by the lymphocytes showed that the majority of the effect was the result of soluble mediators with little cell-contact dependency. Blocking NO and TGFβ production, two known immunomodulatory mediators produced by RPE cells, ^2,35,36 had no effect on either CCL3 or CCL4 production. NO and TGFβ have previously been shown to modulate T cells in various ways, for example, proliferation and cytokine secretion. ^37,38 The lack of inhibition of CCL3 and CCL4 by NO is in line with findings indicating that NO can amplify CCL3 production by lymphocytes and that NO synthase inhibitors decrease CCL3 and CCL4 production by PHA-stimulated lymphocytes. ^39,40 Blocking PGE2 production did, however, return CCL4 production, but not CCL3 production, to that seen for lymphocytes alone. This is consistent with a report that PGE2 induces inhibition of CCL4 promoter activity in human peripheral blood T cells via inducible cAMP early repressor. ²² Blocking PGE2 did, however, result in a significant decrease in CCL3 for lymphocytes alone, which may relate to the finding that low concentrations of PGE2 may enhance T-cell proliferation and be able to potentiate Th1 and Th17 responses. ⁴¹  

We then investigated whether sCD54 produced by the RPE cells was responsible for the inhibition of lymphocyte CCL3 production. CD54 is an adhesion molecule, which plays a central role in cell–cell adhesion or cell–matrix interaction and has a key role in the recruitment of cells from the circulation into an inflammatory site. It is expressed by a range of different cell types including leukocytes and endothelial and epithelial cells. ⁴² Soluble CD54 is thought to arise from differential mRNA splicing and cleavage from membrane-bound CD54 (mCD54). ⁴³ It is known to be produced by different cell types but its function is under debate. The fact that sCD54 ³¹ is associated with inflammatory disease and that it is up-regulated in vitro by proinflammatory cytokines such as TNFα, IL-1β, and IFNγ ⁴⁴ have led to the suggestion that it is pro-inflammatory and may activate lymphocytes and contribute to immunopathological processes in inflammatory diseases. However, there is evidence that it may be immunosuppressive. It can act as a direct competitive inhibitor of mCD54 ⁴⁵ and has been shown to inhibit lymphocyte binding to endothelial cells. ⁴⁶ Studies with transgenic mice overexpressing sCD54 also suggest that it can interfere with cell–cell interaction and leukocyte recruitment, reducing neutrophil and macrophage recruitment after intraperitoneal thioglycollate injection and the inflammatory cell infiltrate in contact hypersensitivity. ²⁴  

Soluble CD54 is associated with many pathological conditions including some inflammatory conditions of the eye. ⁴⁷ Human RPE cells have been shown to express CD54 at intercellular junctions, but in addition to this, CD54 has been reported to be secreted by RPE cells constitutively and markedly increased by treatment with IFNγ, TNFα, and IL1β. ^30,31 In this study, we confirmed constitutive production and up-regulation of sCD54 by TNFα and IL1β in our cell lines. Soluble CD54 significantly reduced CCL3 and CCL4 production by lymphocytes, and antibody to CD54 significantly reduced the inhibitory effect of RPE cell coculture on CCL3 and CCL4. 

Inhibition of CCL3 and CCL4 production by sCD54 may act by blocking LFA-1–CD54 interaction between T cells. LFA-1 has been shown to contribute to T-cell activation and T-cell differentiation in addition to T-cell adhesion, ⁴⁸ and it regulates the production of Th1 cytokines in vivo. ⁴⁹ It has been shown that the cross-linking of LFA-1 by anti-LFA-1 induces the secretion of high levels of CCL3 and CCL4 by pre-activated lymphocytes. ⁵⁰ CD54 can cross-link LFA-1 in a comparable way, leading to similar signaling events. ⁵¹ In addition, it is possible that some of the inhibitory effect may be due to inhibition of T-cell interaction with residual myeloid cells present in the preparation. 

Thus it seems likely that RPE cells utilize sCD54, as part of their arsenal for regulating the activated T lymphocytes they encounter, to down-regulate CCL3 and CCL4 production. Recently, use of sCD54 as a competitive inhibitor of the membrane-bound form has been shown to reduce inflammatory foci without altering T-cell composition in a mouse model of Sjogren's syndrome, but this was only if treatment was given just prior to the infiltration of inflammatory cells into the salivary gland. ⁵²  

Regulation of CCL4 by RPE cells, unlike that of CCL3, is also achieved through PGE2. It is not surprising that CCL3 and CCL4 are regulated in different ways by RPE cells. Although they are often co-expressed, they do have distinct actions, can be antagonistic, and are modulated by different transcription factors. ^53,54  

There are numerous soluble mediators produced by RPE cells, ² and in this study we have only been able to examine a selection; others, such as galectin-1, may also prove to be involved. ⁶ Down-regulation of lymphocyte CCL3 and CCL4 is another of the many immunosuppressive effects employed by RPE cells to protect the retina from a damaging immune response. Understanding how this is brought about will be important for insight into how CCL3 and CCL4 may be down-regulated in damaging proinflammatory reactions, not only in the retina but also elsewhere. 

Footnotes

 Supported by The Guide Dogs for the Blind Association, UK and NHS Grampian Endowment Trust.

Footnotes

 Disclosure: C.A. Wallace, None; G. Moir, None; D.F.G. Malone, None; L. Duncan, None; G. Devarajan, None; I.J. Crane, None