Surface expression of CC- and CXC-chemokine receptors on leucocyte subsets in inflammatory joint diseases

Abstract

Chemokine receptors play a crucial role in the recruitment of leucocyte subsets into inflamed tissue. Using FACS analysis we have studied the surface expression of different CC- and CXC-chemokine receptors on synovial fluid (SF) and peripheral blood leucocytes from 20 patients with various forms of arthritis. In the SF the majority T cells stained positive for CCR5 (93%) and CCR2 (57%), compared to the peripheral blood (36% and 25%). In addition, most of the T cells expressed CXCR4 in both compartments, with a somewhat higher percentage in the SF (90%) versus peripheral blood (83%). To date little information is available on chemokine receptor expression on monocytes in arthritis. We report a marked increase of CCR5⁺ monocytes in the SF (87%) compared to the peripheral blood (22%). In contrast, the frequency of CXCR1⁺, CXCR2⁺, CXCR4⁺ and CCR1⁺ monocytes was considerably lower in the SF than in the peripheral blood. Moreover, we report the expression CXCR4 on neutrophils in the SF. Approximately 60% of neutrophils stained positive for CXCR4 in the SF, while in the peripheral blood the number of CXCR4⁺ neutrophils was low (24%). Surface expression of CXCR1 and CXCR2 was significantly reduced on SF neutrophils (53% and 68%) compared to the peripheral blood. Chemokine receptors are differentially expressed on leucocyte subsets in arthritis. The identification of their pattern of expression might help to identify suitable targets for therapeutic intervention.

Introduction

Chemokines are small chemotactic proteins that play a central role in the recruitment of leucocytes into inflamed tissue. According to the relative position of the first cysteine residues, chemokines are divided into C-, CC-, CXC- and CX₃C-chemokines. Chemokine receptors belong to a family of G-protein coupled heptahelical receptors that are differentially expressed on leucocyte subpopulations. In vitro the chemokine system is characterized by a considerable redundancy, since most chemokines bind to several receptors and most chemokine receptors bind more than one chemokine. In vivo, however, the blockade of single chemokines or receptors has been shown to reduce significantly the disease activity in numerous animal models of inflammation [1]. In humans, a predominance of CCR5 expressing leucocytes has been described in inflammatory infiltrates present in arthritis [2–5], inflammatory renal diseases, including transplant rejection [6], inflammatory bowel disease [7] and multiple sclerosis [8]. Further, a high percentage of CXCR3 and CXCR4 expressing CD45RO⁺ T cells have been detected in the synovial tissue of patients with rheumatoid arthritis [2,9–11]. Little is known about the distribution of other chemokine receptors.

Chemokine receptor expression on leucocytes is tightly regulated during cell differentiation and cytokine activation. T cell maturation from naive to memory cells is paralleled by a reduction of CXCR4 and an increase in CCR5 expression [12]. The differentiation of helper T cells into a TH1 phenotype is associated with CXCR3 and CCR5 expression, while TH2 cells predominantly express CCR3, CCR4 and CCR8 [13,14]. Further, genetics contribute to a variability in chemokine receptor expression on leucocytes between individuals. Twenty per cent of the Caucasian population are heterozygous for a 32 basepair deletion in the CCR5 allele resulting in a diminished CCR5 expression on peripheral blood lymphocytes. Homozygous deletion precludes the expression of CCR5 on the cell surface [15]. The impact of a diminished or missing CCR5 expression on the incidence of rheumatoid arthritis is under discussion [16,17].

In various animal models of inflammation chemokine receptor knock-out mice show a lower disease activity. CCR5 and CCR2 knock-out mice are partially protected from dextran-sulphate mediated colitis [18]. Similarly, CCR2 knock-out mice are less affected by experimental autoimmune encephalomyelits, artheriosclerosis and allergy-induced bronchial hypersensitivity [19,20].

In patients with arthritis much attention has been focused on the expression of CCR5, CXCR3 and CXCR4 by T cells. In order to identify other potentially relevant chemokine receptors in inflammatory joint diseases, we have expanded our analysis to CCR1, CCR2, CXCR1 and CXCR2 by infiltrating T cells, monocytes and neutrophils. These receptors may be of critical importance, as ligands for these receptors, RANTES, MIP-1α, MCP-1, IL-8 and Gro-α, have been detected in the synovial fluid and tissue of patients with rheumatoid arthritis [21–27]. Only recently, the CXCR4 ligand SDF-1α has been described in the synovial tissue of patients with rheumatoid arthritis [10,11]. We have further studied the expression of CXCR4 on mononuclear cells and neutrophils in the context of inflammation. The presence of CXCR4 on neutrophils is controversial and no information is available on the expression of CXCR4 by neutophils under inflammatory conditions.

Small molecular antagonists have been found to efficiently block chemokine receptors [28] and modified antibodies proved to be effective in eliminating chemokine receptor expressing cells in vitro [29]. In view of impending clinical trials with those reagents, the expression pattern of chemokine receptors needs to be defined under physiological and inflammatory conditions.

Materials and methods

Patients

Synovial fluid and blood samples were obtained simultaneously from 20 patients who presented with arthritis of the knee for diagnostic or therapeutic arthrocentesis. Diagnoses included rheumatoid arthritis (seven), reactive arthritis (three), undifferentiated arthritis of the knee (four), psoriatic arthritis (three), activated osteoarthritis (two) and ankylosing spondylitis (one) according to ACR criteria, where applicable. Written informed consent was obtained from all patients. Synovial fluid leucocytes were differentiated routinely by light microscopy after Jenner–Giemsa staining (Table 1).

Table 1.

Clinical data of 20 patients with different forms of arthritis

Diagnosis	Patients	Mean duration of effusion	Mean SF leucocyte count ×109/l	Mean SF cell type % (PMN/Ly/Mo + Mφ)	NSAID/steroids/ DMARD %
RA	7	7 weeks	11 500	69/17/14	4/3/3
AS	1	4 weeks	4670	55/21/41	0/0/0
ReA	3	5 weeks	13 550	41/23/36	1/0/1
OA	2	12 weeks	4500	41/23/36	1/0/0
PsA	3	24 weeks	69 075	69/8/23	0/0/1
Undiff. A	4	3 weeks	35 016	81/7/12	2/0/0

RA: rheumatoid arthritis (n = 7), AS: ankylosing spondytitis (n = 1), ReA: reactive arthritis (n = 3), OA: activated osteoarthritis (n = 2), PsA: psoriatic arthritis (n = 3), Undiff. A: undifferentiated arthritis of the knee (n = 4). Synovial fluid analysis was performed by light microscopy after Giemsa staining. Mean total leucocyte counts are given in the table and cell differentiation is described as a percentage of polymorphonuclear neutrophils (PMN), lymphocytes (Ly) and monocytes/macrophages (Mo/Mφ). The number of patients receiving oral medication (NSAID, prednisone >7·5 mg or equivalent, DMARD) is detailed.

Cell preparation, staining and FACS-analysis

Immediately after arthrocentesis SF leucocytes were isolated by two washing steps with 5% PBS in NaCl 0·9%. Synovial fluid cells and whole blood (containing 1 mm EDTA) were incubated on ice with monoclonal antibodies against chemokine receptors and the appropriate isotype controls at a concentration of 10 µg/ml. The antibodies were for CCR5: MC-1 [30], for CCR2: DOC-3, which specifically binds to CCR2 [3], for CCR1: Clone 53504 (R & D-Systems), for CXCR1: 5A12 (PharMingen), for CXCR2: 6C6 (PharMingen), and for CXCR4: 12G5 (PharMingen), IgG1-, IgG2a- and IgG2b-isotype controls (Sigma). After two washing steps cells were incubated for 30 min on ice with a PE-conjugated rabbit-antimouse F(ab)₂ fragment (R439, DAKO). Cells were washed twice and incubated with 10% mouse serum followed by a combination of anti-CD4-FITC, anti-CD8-PECy5 and anti-CD14-APC (Immunotech). After lysis of erythrocytes, cells were analysed immediately by flow cytometry (Becton-Dickinson). Calculations were performed with Cell Quest analysis software. Helper T cells, cytotoxic T cells, monocytes and neutrophils were identified by their light-scatter properties and the expression or absence of CD4, CD8 and CD14. In separate staining PE-conjugated antileucocyte alkaline phosphatase MoAb (1B12, PharMingen) was used to confirm gating for peripheral blood and synovial fluid neutrophils. As established previously by staining with propidium iodide, dead cells were excluded from the analysis by virtue of their light-scatter properties and altered expression of CD4, CD8 and CD14. Chemokine receptor expression was calculated after defining a cut-off according to the isotype control. Two-sided paired Student's t-test was used for statistical analysis.

Determination of the CCR5 genotype

Genomic DNA was prepared from frozen blood samples by affinity chromatography (Roche Diagnostics). Subsequently a fragment of the CCR5 gene containing the potential 32 basepair deletion was amplified by a 40 cycle PCR with Taq polymerase. The primers were P1: 5′tttaccagatctcaaaaagaag and P2: 5′ggagaaggacaatgttgtagg. Differences in the length of the PCR fragments (274 or 242 bp) allowed to identify CCR5-wild-type and CCR5-Δ32 alleles [3].

Results

Chemokine receptor expression on T cells

By FACS-analysis we compared the expression of various CC- and CXC chemokine receptors on synovial fluid and peripheral blood leucocytes from 20 patients with different causes of joint effusions. Approximately 88% of CD4⁺ T cells and 93% of the CD8⁺ T cells from the synovial fluid stained positive for CCR5. Similarly, a major proportion of CD8⁺ and CD4⁺ T cells in the SF expressed CCR2 (66% and 48%) for all types of arthritis (Figs 1,2). In contrast, only a minority of T cells in the peripheral blood expressed the chemokine receptors CCR5 or CCR2. The enrichment in the synovial fluid was most pronounced for the CCR5⁺ helper-T cells (blood: SF ratio = 1: 4). Further, the majority of CD4⁺ and CD8⁺ T lymphocytes stained positive for CXCR4. The percentage of CD4⁺/CXCR4⁺ and CD8⁺/CXCR4⁺ T cells was somewhat increased in the synovial fluid (90% and 90%) compared to the peripheral blood (88% and 78%) (Fig. 3). Whereas differences in the expression of CCR2 and CCR5 on T cells from the SF compared to the peripheral blood were highly significant (P < 0·0001, two-sided paired Student's t-test), differences in CXCR4 expression did not reach statistical significance. CXCR1, CXCR2 and CCR1 were expressed only by a minor and variable percentage of T cells. With the restriction of a limited patient number no obvious differences could be detected in relation to the underlying diagnoses, duration of joint effusion or pretreatment.

Fig. 1.

Surface expression of chemokine receptors on leucocyte subsets in the peripheral blood (□) and the synovial fluid (▪) of patients with arthritis. For the constitutively expressed chemokine receptors CXCR1 and CXCR2 on neutrophils the mean fluorescence intensity (MFI) is given, otherwise the mean percentage of positive T-cells, monocytes and neutrophils is depicted. Dots represent individual patient data.

Fig. 2.

FACS dot plots showing the expression of CCR5 and CCR2 on leucocytes in the peripheral blood (left) and synovial fluid (right) of one patient with rheumatoid arthritis. The cut-offs were set according to the isotype controls and are shown as vertical lines.

Fig. 3.

FACS dot plots showing the expression of CXCR4 on leucocytes in the peripheral blood (left) and synovial fluid (right) of one patient with rheumatoid arthritis. The cut-offs were set according to the isotype controls and are shown as vertical lines.

Chemokine receptor expression on monocytes

Consistent with our previous data, the majority of monocytes in the SF expressed CCR5 (87%) versus a minority in the peripheral blood (21%), P < 0·0001 (two-sided paired Student's t-test). In addition, we now report a reduced percentage of synovial fluid monocytes expressing CXCR1, CXCR2, CXCR4 and CCR1 compared to the peripheral blood (Fig. 1). Notably, the level of expression determined by mean fluorescence intensity (MFI) was considerably lower for the CXC-chemokine receptors 1, 2 and 4 on synovial fluid monocytes (Fig. 4). In particular, CXCR1 expression (MFI) on synovial fluid monocytes was reduced in all patients (SF: mean 65 versus blood: 202), reaching statistical significance with P = 0002 (two-sided paired Student's t-test) (Fig. 5). No statistical significance was reached for CXCR2 and CXCR4.

Fig. 4.

Histogram plots depicting differences in chemokine receptor expression level (mean fluorescence intensity) between peripheral blood and synovial fluid monocytes of one patient with rheumatoid arthritis. Matched isotype controls are shown as thin lines while chemokine receptor expression is given in bold.

Fig. 5.

Significantly decreased surface expression of CXCR1 on neutrophils (P = 0·0001) and monocytes (P = 0·002) from the synovial fluid compared to the peripheral blood from patients with different forms of arthritis (PsA: n = 3, RA: n = 2, ReA: n = 1, OA: n = 1, undiff.A: n = 1). The mean fluorescence intensity (MFI) for CXCR1 expression on neutrophils (left plot) and monocytes (right plot) in the peripheral blood and synovial fluid is given for each patient.

Chemokine receptor expression on neutrophils

Acute inflammatory joint effusions are characterized by a high number of neutrophils in the synovial fluid. We investigated the chemokine receptor expression on neutrophils in the synovial fluid potentially involved in the recruitment of this cell population. Interestingly, CXCR4 was found to be expressed on a large fraction of neutrophils (60%) from the synovial fluid of patients with acute and chronic arthritis, while a significantly lower expression was found in the peripheral blood of these patients (24%) (Figs 1 and 3), P < 0·001 (two-sided paired Student's t-test).

CXCR1 and CXCR2 are constitutively expressed on all neutrophils, but the expression level of both chemokine receptors was significantly reduced on neutrophils in the synovial fluid compared to the peripheral blood. As shown in Figs 1 and 5 the mean fluorescence intensity of CXCR1 was 1890 on synovial fluid neutrophils and 3560 on peripheral blood neutophils, P = 0·0001. Similarly, CXCR2 surface expression was reduced on SF neutrophils (mean 763) compared to the peripheral blood (mean 1116), P = 0·03. As expected, no expression was found for CCR5 and CCR2 on neutrophils.

Recently, it was reported that CCR1 induced on neutrophils after activation with IFN-γ [31] is involved in the recruitment of neutrophils in response to MIP-1α in vivo [32]. However, CCR1 was found to be expressed by only a minority of neutrophils in the synovial fluid and peripheral blood.

Homozygous CCR5 Δ32-deletion in one patient with undifferentiated gonarthritis

PCR-based genotype analysis revealed that 23% of our patients were heterozygous for the Δ32 CCR5 allele, a frequency that did not differ significantly from a local control population [33]. Interestingly, one patient with a recurrent undifferentiated oligoarthritis in both knees and ankles was homozygous for the 32 bp deletion in the CCR5 allele. In this patient CCR2⁺/CD4⁺ and CCR2⁺/CD8⁺ T cells were highly enriched in the synovial fluid (53% and 65%) compared to the peripheral blood (17% and 14%), resembling the chemokine expression pattern observed in CCR5 wild type and heterozygous patients.

Discussion

In 20 patients with different forms of arthritis we compared the expression of chemokine receptors on leucocytes from the peripheral blood and the synovial fluid. Our analysis included the chemokine receptors CCR1, CCR2, CCR5, CXCR1, CXCR2 and CXCR4. Chemokine receptor surface expression on synovial fluid leucocytes was characterized by: (1) a predominance of CCR5, CCR2 and CXCR4 positive T cells; (2) an increased percentage of CCR5⁺ and a decreased percentage of CXCR1⁺, CXCR2⁺, CXCR4⁺ and CCR1⁺ monocytes; and (3) an increased expression of CXCR4 and decreased expression of CXCR1 and CXCR2 on neutrophils.

In different forms of rheumatic joint disease the percentage of CD4⁺ and CD8⁺ T cells expressing the chemokine receptor CCR5 was strongly increased. These data are in good agreement with previous reports that have suggested a strong expression of TH1-associated chemokine receptors on infiltrating T cells and the presence of their respective ligands in the synovial fluid of RA patients [2–5,9]. Some controversy exists on the significance of the of Δ32 CCR5 allele on the incidence and severity of rheumatoid arthritis. Gomez-Reino et al. suggested that homozygosity for the Δ32 CCR5 allele reduces the incidence of rheumatoid arthritis by preventing functional expression of CCR5 on the cell surface [16]. Including previous data [3,5], we have identified one homozygous patient among 40 patients with arthritis. Together with data from several other groups and reports by Cooke et al. who found two homozygous RA patients out of 278 [17] it now seems questionable that the CCR5 genotype plays a dominant role in the pathogenesis of arthritis. In all patients, including the CCR5-deficient patient, the percentage of CD4⁺ and CD8⁺ T cells that expressed CCR2 was increased in the synovial fluid. This finding suggests that as well as CCR5 other chemokine receptors are involved in rheumatic joint inflammation. In MRL/lpr mice, adjuvant arthritis could be prevented and partially cured by an MCP-1 analogue that is active against CCR2 [1]. To our knowledge, no data on arthritis models in CCR knock-out mice are available to date.

In the peripheral blood CXCR4 is predominantly expressed by naive CD45RA⁺ T cells, whereas CD45RO⁺ memory cells express mainly CCR5 [12]. Recently, SDF-1, the only known ligand of CXCR4 was found to be functionally active on CD45RO⁺ synovial fluid T cells in chemotaxis and integin mediated adhesion [10,11]. Similarily, in our study we detected a high percentage of synovial fluid T cells that stained positive for CXCR4, indicating that in the context of inflammation, CXCR4 expression is also found on activated effector T cells. The co-localization of CD45RO⁺ T cells expressing CCR5, CXCR4 and CCR2 with their respective chemokine ligands in the SF could suggest a selective recruitment of this lymphocyte population into the site of inflammation via their chemokine receptors.

The chemokine receptor profile on monocytes differed significantly between the peripheral blood and the synovial fluid. In the synovial fluid the vast majority of monocytes expressed the chemokine receptor CCR5. The expression of CXCR1, CXCR2, CXCR4 and CCR1 was decreased in the synovial fluid compared to the peripheral blood. Reduction in chemokine receptor expression on synovial fluid monocytes was most pronounced for CXCR1. Although the lower expression of these receptors could merely reflect a differentiation of monocytes into macrophages, it could also indicate that some of the chemokine receptors have been engaged during chemotaxis and were subsequently removed from the cell surface by ligand-induced receptor internalization [30]. This is consistent with a clear expression of IL-8, ENA-87, Gro-α, SDF and RANTES in synovial tissue of RA patients [21]. In addition, cytokines differentially regulate chemokine receptor expression on monocytes and could therefore have a strong impact on the selective expression of chemokine receptors on leucocytes in the synovial fluid. With regard to CXCR4, receptor down-modulation on monocytes has been observed after stimulation of cells with IFN-gamma [34]. The expression of CCR5 was strongly increased on SF monocytes. CCR5 is up-regulated during the maturation of macrophages as well as after stimulation of monocytes with IL-10. Thus, high expression levels of this chemokine receptor on SF monocytes could be attributed to the state of maturation and activation at the site of inflamation, rather than selective cell migration.

The IL-8 receptors CXCR1 and CXCR2 are constitutively expressed on neutrophils in the peripheral blood and synovial fluid. However, the expression level of these receptors was decreased by almost 50% on synovial fluid neutrophils. IL-8 could down-regulate the surface expression of CXCR1 and CXCR2 by receptor internalization, suggesting that both receptors have been engaged during the recruitment of neutrophils. Alternatively, the reduced expression of CXCR1 and CXCR2 could be explained by the presence of TNF-α, which leads to a proteolytic degradation of IL-8 receptors [35]. Apart from soluble cell mediators, triggering of CD45 tyrosine phosphatase activity has been shown to down-regulate IL-8 receptors on neutrophils via tyrosine kinase activation in vitro [36]. Different sentivivity of chemokine receptors to regulatory intracellular signalling pathways or proteolytic clipping could contribute to the differential expression of receptors on activated leucocytes at the site of inflamation.

Controversial data exist on the expression of CXCR4 on neutrophils. While some authors were unable to detect expression of CXCR4 [37,38], others have described binding of SDF1-α to neutrophils, induction of Ca-flux and migration [39,40]. We describe for the first time the expression of CXCR4 on neutrophils in synovial fluid, demonstrating that neutrophils can up-regulate CXCR4 during inflammation. Interestingly, the surface expression of CXCR4 seems to be regulated differently on neutrophils and monocytes, since receptor surface expression was reduced on monocytes whereas it was up-regulated on neutrophils in the synovial fluid. Further work is needed to clarify the role of CXCR4 expression on neutrophils.

Taken together, our analysis provides new information on the surface expression of different chemokine receptors in arthritis and might help to identify suitable targets for therapeutic intervention in inflammatory joint diseases.