CCR5 and its ligands (CCL3, CCL4, and CCL5) may play a role in inflammatory cell recruitment into the joint. However, it was recently reported that CCR5 on T cells and neutrophils acts as a decoy receptor for CCL3 and CCL5 to assist in the resolution of inflammation. The aim of this study was to determine whether CCR5 functions as a proinflammatory or antiinflammatory mediator in arthritis, by examining the role of CCR5 in proteoglycan (PG)–induced arthritis (PGIA).
Arthritis was induced by immunizing wild-type (WT) and CCR5-deficient (CCR5−/−) BALB/c mice with human PG in adjuvant. The onset and severity of PGIA were monitored over time. Met-RANTES was used to block CCR5 in vivo. Arthritis was transferred to SCID mice, using spleen cells from arthritic WT and CCR5−/− mice. The expression of cytokines and chemokines was measured by enzyme-linked immunosorbent assay.
In CCR5−/− mice and WT mice treated with the CCR5 inhibitor Met-RANTES, exacerbated arthritis developed late in the disease course. The increase in arthritis severity in CCR5−/− mice correlated with elevated serum levels of CCL5. However, exacerbated arthritis was not intrinsic to the CCR5−/− lymphoid cells, because the arthritis that developed in SCID mouse recipients was similar to that in WT and CCR5−/− mice. CCR5 expression in the SCID mouse was sufficient to clear CCL5, because serum levels of CCL5 were the same in SCID mouse recipients receiving cells from either WT or CCR5−/− mice.
These data demonstrate that CCR5 is a key player in controlling the resolution of inflammation in experimental arthritis.
Rheumatoid arthritis (RA) is a chronic, progressive autoimmune disease characterized by joint infiltration by several populations of leukocytes, including T cells, B cells, macrophages, and neutrophils (1). These infiltrating cells, in collaboration with resident cells, mediate cartilage destruction and bone erosion through a combination of cytokines, chemokines, and matrix metalloproteinases (2, 3). In synovial fluid from patients with RA, all 4 groups of chemokine families, CXC, CC, C, and CX3C (and their receptors) have been identified (3). The inhibition of infiltrating cells into the joint therefore represents an important step for therapeutic intervention.
CCR5 is a protein G–coupled chemokine receptor composed of 7 transmembrane helices, which is expressed by activated T cells, monocytes, dendritic cells, tissue macrophages, and neutrophils (4, 5). CCR5 is the natural receptor for the ligands CCL3 (macrophage inflammatory protein 1α [MIP-1α]), CCL4 (MIP-1β), and CCL5 (RANTES), all of which contribute to chemotactic activity in the synovial fluid of patients with RA (6–8). The accumulation of CCR5-positive T cells in the synovium of patients with RA suggests an important role in disease pathology (8).
However, the chemotactic responses of lymphocytes to RA synovial fluid were shown not to be dependent solely on a functional CCR5 receptor (9). CCR5-positive T cells express other chemokine receptors, such as CCR1, that also are capable of responding to CCL3 and CCL5 (9). An allelic form of the CCR5 gene contains a 32-bp deletion (Δ32CCR5) that results in a nonfunctional CCR5 receptor. Several studies have addressed whether a loss-of-function mutation could have a protective effect against RA. Two of these studies demonstrated a statistically significant negative correlation between Δ32CCR5 and RA (10, 11), whereas the results of 3 other studies were not significant (12–14). More recently, a meta-analysis of these 5 published studies demonstrated a significant negative association of Δ32CCR5 with RA (15), suggesting that, at least in the population of European ancestry, it may be protective.
Treatment with a modified form of RANTES known as Met-RANTES mildly inhibited arthritis in animals with collagen-induced arthritis (CIA) and adjuvant-induced arthritis (15–17); however, these studies did not exclude the possibility that the interaction of Met-RANTES with CCR1 mediates such inhibition. Use of a nonpeptide antagonist of CCR5 in CIA revealed intact cellular immune responses but impaired T cell migration (16). Contrary to these findings, CIA in CCR5-deficient mice (CCR5−/−) is similar to that in wild-type (WT) mice (17).
A novel role for CCR5 in the clearance of inflammation was recently elucidated. Early clues suggesting an antiinflammatory role of CCR5 showed that the CCR5 antagonist Met-RANTES affected the uptake of apoptotic cells in a model of glomerulonephritis (18). Similarly, a deficiency in CCR5 or in blocking CCR5 enhanced pathogenic inflammatory responses in mice with hepatic liver disease, pancreatitis, or glomerulonephritis (18–20). An important mechanistic explanation for these phenomena revealed that CCR5 may act as a novel decoy receptor on late apoptotic neutrophils and T cells. In this way, CCR5 can function to remove excess CCL3, CCL4, and CCL5 from tissue in the presence of proresolution lipid mediators (5).
In this study, we use an established model of RA, proteoglycan (PG)–induced arthritis (PGIA), to examine the function of CCR5 in disease. We observed that CCR5 function is important in the clearance of CCL5, thereby promoting the resolution of inflammation.
MATERIALS AND METHODS
CCR5−/− mice on the BALB/c background were originally generated by Dr. Rodrigo Bravo (21) and were subsequently backcrossed to BALB/c mice for 10 generations. CCR5−/− mice were generously supplied by Dr. Don Moser (The Scripps Research Institute, La Jolla, CA). Mice were genotyped using primers specific for CCR5. Wild-type BALB/cJ mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Mice were maintained at the Rush University Medical Center facility. All animal experiments were approved by the institutional Animal Care and Use Committee at Rush University Medical Center.
Induction and assessment of arthritis.
Cartilage obtained during human joint replacement surgery was acquired via the Orthopedic Tissue and Implant Repository of Rush University Medical Center, with the approval of the institutional review board. PG (aggrecan) was isolated as previously described (22). Age-matched female BALB/c mice, 12–14 weeks of age, were used in all experiments. Mice were immunized intraperitoneally with 150 μg of human PG (measured as protein) in dimethyldioctadecylammonium bromide (DDA; Sigma-Aldrich, St. Louis, MO), as previously described (23). Booster immunizations with 100 μg of PG in DDA were given at 3 weeks and 6 weeks. Mice were monitored for arthritis twice weekly, and arthritis was scored in a blinded manner. Paw swelling was scored, based on an established scoring system, on a scale of 1–4, as follows: 0 = normal, 1 = mild erythema and swelling of several digits, 2 = moderate erythema and swelling, 3 = more diffuse erythema and swelling, and 4 = severe erythema and swelling of the complete paw, with ankylosis. The incidence of arthritis denotes the percentage of mice in which PGIA develops, with scores ranging from 0 to 16, based on individual paw scores of 0–4. The hind ankle joints of immunized mice were isolated. The joints were fixed in formalin, decalcified in 5% formic acid, embedded in paraffin, and stained with hematoxylin and eosin. Cellular infiltration was measured on a scale of 0–4 in a blinded manner.
Transfer of arthritis to SCID mice.
Spleen cells (2.5 × 107/mouse) from immunized WT or CCR5−/− mice and 100 μg of PG were injected intraperitoneally into SCID mice. Mice were monitored under blinded conditions for the onset and severity of disease, as described above. SCID mice were treated with Met-RANTES (100 μg) in phosphate buffered saline (PBS) or with PBS alone every third day for 9 days, starting on day 6 or day 9, depending on the experiment. Met-RANTES was generously provided by Dr. Amanda Proudfoot (Geneva Research Centre, Merck Serono SA, Geneva, Switzerland).
Assessment of T cell activation and autoantibody production.
CD4+ T cells were isolated from the spleens of PG-immunized mice by negative selection, using CD4 isolation kits and an AutoMACS automated separation system (Miltenyi Biotech, Auburn, CA). CD4+ T cells (2.5 × 105/well) were stimulated in the presence or absence of PG (10 μg/ml) and naive irradiated spleen cells (2.5 × 105/well; 2,500 rad) in 96-well plates (Fisher Scientific, Fair Lawn, NJ) in serum-free HL-1 medium (Fisher Scientific) containing 100 μg/ml penicillin, 100 μg/ml streptomycin, and 2 mML-glutamine (complete medium). Cells were cultured in quadruplicate. Cultures were incubated at 37°C in 5% CO2 for 5 days and pulsed with 3H-thymidine (0.5 μCi/well) for the last 18 hours. Cells were harvested (Tomtec, Orange, CT), and incorporated 3H-thymidine was assessed by scintillation counting (Wallac, Gaithersburg, MD).
Serum was obtained from immunized mice and assessed for antibodies to mouse PG. For tissue culture enzyme immunoassay, Costar Half-Area Plates (Corning, Corning, NY) were coated with 0.75 μg of native mouse PG in carbonate buffer. Sera were serially diluted in PBS containing 0.5% Tween 20. Samples were incubated with the immobilized mouse PG, and mouse PG–specific autoantibodies were detected using peroxidase-conjugated rabbit IgG against mouse IgG1 and IgG2a (Zymed, South San Francisco, CA). Secondary antibodies were detected with the substrate o-phenylenediamine. The relative concentration was determined using a standard curve of known concentrations of unlabeled murine IgG1 and IgG2a. Data represent the mean ± SEM values for IgG1 and IgG2a autoantibodies from 7–10 mice.
Assessment of cytokines/chemokines.
Spleens were harvested from PG-immunized mice. CD4+ T cells were purified using CD4 isolation beads (Miltenyi Biotech) and cultured with naive irradiated spleen cells (2,500 rad) in the presence or absence of PG (10 μg/ml) in complete medium. Supernatants were harvested on day 4 and examined for cytokines by enzyme-linked immunosorbent assay (ELISA), using the OptEIA mouse interferon-γ (IFNγ) or interleukin-4 (IL-4) kit (BD PharMingen, San Diego, CA) and a mouse IL-17 ELISA kit (R&D Systems, Minneapolis, MN). Synovial fluid was obtained from the ankle joints of arthritic mice, and samples from 3 mice were pooled. Serum samples from individual mice were assessed for CCL3, CCL4, and CCL5 by ELISA (R&D Systems).
The Mann-Whitney U test was used to compare nonparametric data for statistical significance. P values less than 0.05 were considered significant.
Expression of CCR5 ligands in synovial fluid from arthritic mice.
We first examined the expression of the CCR5 ligands CCL3, CCL4, and CCL5 in arthritic mice (Figure 1). Synovial fluid was obtained from the ankle joints of arthritic hind paws and was pooled and examined by ELISA for the expression of CCL3, CCL4, and CCL5. CCL3, CCL4, and CCL5 were expressed in the synovial fluid of arthritic mice (Figure 1A). Arthritic ankle joints were collagen digested, and cells were stained with antibodies to CD4, Ly-6G/C (Gr-1), and CCR5 (Figure 1B). Synovial tissue contained CD4low- and CD4high-expressing cells, and CCR5 was expressed predominantly on the CD4low cells. Neutrophils are the predominant cell population in synovial tissue. Neutrophils, which were identified by the Gr-1 antibody, were detected as a population of Gr-1low and Gr-1high cells, with CCR5 expressed on the Gr-1low cells. The presence of CCL3-, CCL4-, CCL5-, and CCR5-expressing CD4+ T cells and neutrophils suggests a functional role for these chemokines and chemokine receptors in PGIA.
Transient inhibition of arthritis by CCR5 blockade.
In order to determine whether CCR5 plays a role in PGIA, we used a peptide antagonist, Met-RANTES, to transiently block the engagement of CCR5 and CCR1 with their natural ligands. To facilitate the effective use of the peptide antagonist, we used the adoptive transfer model of PGIA. In this system, transfer of spleen cells from arthritic mice and transfer of PG intraperitoneally into naive SCID mice produced rapid and reproducible arthritis within a few weeks. Recipient mice were treated, starting on day 6 after cell transfer, with 100 μg of Met-RANTES in PBS or with PBS alone, every third day for 9 days. Our results showed a brief, significant delay in the onset and severity of disease in mice treated with Met-RANTES in comparison with PBS-treated mice (Figures 2A and B). These data suggest that CCR5 and CCR1 contribute to the inflammatory process in PGIA. Interestingly, as arthritis progressed, swelling and erythema appeared to be enhanced in the Met-RANTES–treated mice.
To assess the possibility that blocking CCR5/CCR1 might have a negative affect on inflammation, we repeated the adoptive transfer experiment. In this experiment, spleen cells from arthritic mice were transferred into SCID mouse recipients, but treatment with Met-RANTES was delayed until day 12 (Figures 2C and D). Beginning on day 12, 100 μg of Met-RANTES was administered every 3 days for 9 days. The early suppression of arthritis was blunted, probably because of the delay in Met-RANTES treatment. However, when the development of arthritis was monitored over several weeks, disease severity was significantly elevated in the Met-RANTES–treated mice. The normal waxing and waning of inflammation observed in the PBS-treated mice did not occur in the Met-RANTES–treated mice. These data demonstrate that CCR5/CCR1 had an important regulatory function in PGIA.
Sustained inflammation in CCR5−/− mice as disease waned in WT mice.
To begin to understand the role of CCR5 in regulating PGIA, we used mice deficient in CCR5. We immunized groups of age-matched WT and CCR5−/− BALB/c mice with PG in adjuvant on days 0, 21, and 42 and monitored disease onset and severity over time (Figures 3A and B). CCR5−/− mice succumbed to disease early, with the onset and severity of arthritis being similar to that in WT mice. However, after week 11 postimmunization, disease became more severe in CCR5−/− mice compared with WT mice, and disease was significantly more severe by week 13. These data support a role for CCR5 in the regulation of inflammation rather than in the recruitment of inflammatory cells late in disease. Histologic assessment of ankle joints from arthritic WT and CCR5−/− mice revealed leukocyte infiltration and synovial lining proliferation, with cartilage loss and bone erosion (Figures 3C and D). Although the clinical score in the CCR5−/− mice was more severe late in disease development compared with WT mice, the histologic appearance of ankle joints from these mice was not significantly different. This observation may reflect the fact that the histologic appearance was very severe in both CCR5−/− and WT mice.
Cytokines are important in the development of PGIA. Thus, it is possible that CCR5−/− mice may have an altered cytokine profile later in the course of disease, when they exhibit severe arthritis. To examine this possibility, we harvested spleen cells from WT and CCR5−/− mice 16 weeks after the initial immunization. We determined whether expression of the T cell inflammatory cytokines IFNγ, IL-17, and IL-4 was increased in CCR5−/− mice. CD4+ T cells were purified from the spleens and restimulated in the presence of naive antigen-presenting cells and PG, and supernatants were assessed for the production of cytokines, using ELISA (Figures 4A–C). In addition, the frequency of CD4+ T cell–expressing cytokines was measured by intracellular staining for IFNγ, IL-17, and IL-4. We observed no differences between WT and CCR5−/− mice in the production or the frequency of CD4+ T cells expressing IFNγ, IL-17, and IL-4 (Figures 4A–C, and results not shown). T cell activation was also measured by the ability of T cells to specifically respond to PG in in vitro culture. No difference was observed between WT and CCR5−/− mice in their CD4+ T cell recall response to PG (Figure 4D).
Autoantibodies specific for mouse PG are a hallmark of PGIA. To determine whether the concentration of PG-specific autoantibodies is increased in CCR5−/− mice in comparison with WT mice, blood samples were obtained after the third immunization with PG. There was no difference in the levels of autoantibodies to PG in WT and CCR5−/− mice (Figure 4E). These data demonstrate that the increase in disease severity in CCR5−/− mice was not governed by an altered T helper cytokine concentration, PG-specific T cell activation, or autoantibody level.
Significantly elevated levels of CCL5 in naive and PG-immunized CCR5−/− mice.
CCR5 deficiency in mice and humans is associated with elevated concentrations of CCL5 in tissue (24–26). Due to the redundancy of the chemokine/chemokine receptor system, it is possible that elevated levels of CCL5 in CCR5-impaired persons (or mice) could promote an enhanced influx of leukocytes via CCR1, which is also expressed on T cells (19, 25, 26). To determine whether this occurs in CCR5−/− mice, we examined the serum concentrations of CCL3, CCL4, and CCL5 in naive mice, 1 week after each PG immunization. CCL3 and CCL4 were undetectable in the serum of WT and CCR5−/− mice; however, the level of CCL5 was significantly elevated in CCR5−/− mice compared with WT mice at each time point tested (Figure 5).
Resolution of severe disease in CCR5−/− mice upon transfer of SCID mouse cells.
CCR5 has emerged as an important chemokine receptor on T cells and neutrophils that is involved in the clearance of inflammation by scavenging CCL3, CCL4, and CCL5 (5). In the setting of PGIA, CD4+ T cells and B cells are required for the development of disease, although T cells are a minor cell population in the inflamed joint, where neutrophils dominate. To determine whether the severe disease that was observed late in CCR5−/− mice was attributable to a defect in CD4+ T cells or a defect in both CD4+ T cells and neutrophils, we used the SCID mouse transfer model of PGIA. In this model, SCID mouse nonlymphoid cells express CCR5, whereas the transferred spleen cells are CCR5 negative.
We transferred spleen cells from WT or CCR5−/− mice into SCID mice and monitored the recipients for disease onset and severity (Figures 6A and B). The onset and severity of disease in SCID mice were similar regardless of whether spleen cells from WT or CCR5−/− mice were transferred. These data suggest that the presence of CCR5 on neutrophils or other non–T cells or non–B cells in the SCID mouse was sufficient to allow for resolution of inflammation similar to that observed in WT recipients. To assess the ability of CCR5-sufficient cells to scavenge chemokines in the SCID mouse recipient, we again examined the CCL5 concentrations in serum from arthritic mice, using ELISA (Figure 6C). In contrast to the elevated concentrations of CCL5 in arthritic CCR5−/− mice, serum concentrations of CCL5 were similar in the SCID mouse recipients of WT or CCR5−/− spleen cells. These data demonstrate that the expression of CCR5 on nonlymphoid cells was able to control the levels of circulating CCL5, supporting the association between elevated levels of CCL5 and sustained inflammation.
The present study was undertaken to determine the function of CCR5 in PGIA, based on the findings that CCR5 ligands are expressed in the synovial fluid of arthritic mice, and that CCR5 is expressed on CD4+ T cells and neutrophils in synovial tissue. In this study, we identified CCR5 as an important molecule in the resolution phase of inflammation. In mice deficient in CCR5, arthritis is similar to that in WT mice early in the disease process. These data suggest that the early infiltration of leukocytes into the joint may be mediated by other CXC, CC, C, and CX3C family members, such as CXCL5, which is known to precede the development of arthritis in adjuvant-induced arthritis (27). However, as arthritis began to wane in WT mice, arthritis severity increased in CCR5−/− mice. Similarly, exacerbation in the late phase of arthritis severity was observed in Met-RANTES–treated mice. However, treatment with Met-RANTES was also inhibitory in the early phase of disease. Because Met-RANTES is not specific for CCR5 but also inhibits CCR1, it is possible that the difference between Met-RANTES treatment and CCR5−/− mice is the inhibition of CCR1. It has been reported that another compound capable of inhibiting CCR1 also suppresses the development of CIA (28). In support of an immunoregulatory role for CCR5, several other models of inflammation demonstrate that a deficiency in CCR5 leads to an increase in inflammation. In models of T cell–mediated hepatitis, influenza A virus, graft-versus-host disease, and Mycobacterium tuberculosis infection, CCR5−/− mice exhibit exacerbated disease (19, 26, 29–31).
A role for CCR5 in the resolution of inflammation has recently been reported (5). The mechanism for this down-regulation of inflammation involves the use of CCR5 as a decoy receptor on the surface of late apoptotic T cells and neutrophils. Binding of local chemokines to CCR5, without subsequent signaling, has the effect of reducing the concentration of chemokines and thereby assisting the resolution of inflammation. We showed that CCL5 was detectable in the serum of both naive and PG-immunized CCR5−/− mice, whereas CCL5 was almost undetectable in WT mice. The presence of elevated levels of CCL5 in CCR5−/− mice correlated with the increase in arthritis severity exhibited in the late phase of disease. These data suggest that the inability to clear CCL5 in CCR5−/− mice contributes to enhanced inflammation.
To determine whether clearance of CCL5 correlates with a reduction in arthritis severity, we used an adoptive transfer model. In this model, spleen cells from arthritic CCR5−/− mice were transferred into CCR5-positive SCID mice. Spleen cells from CCR5−/− mice were unable to transfer accelerated arthritis, indicating that the enhanced disease was not intrinsic to the transferred cell population. Thus, a deficiency in CCR5 expression on spleen cells is not the cause of the enhanced arthritis observed in CCR5−/− mice but rather is attributable to another factor such as high levels of CCL5 in CCR5−/− mice. Moreover, the level of CCL5 in the serum of SCID mice was similar whether the cells were transferred from WT mice or CCR5−/− mice. These data suggest that CCR5-positive myeloid, natural killer, and dendritic cells in the SCID mouse cleared CCL5, and that clearance of CCL5 prevented the development of enhanced arthritis.
Treg cells play an important role in the prevention of autoimmunity. Treg cell migration to sites of inflammation involves several different chemokine receptors, including CCR5 (32–34). A deficiency in Treg cell migration to the joints might account for the increase in arthritis in CCR5−/− mice. If this were the case, we would predict that the transfer of lymphocytes from CCR5−/− mice would transfer Treg cells that could not migrate to arthritic joints; however, we observed no difference in the development of arthritis in SCID mouse recipients, regardless of whether they received cells from WT or CCR5−/− mice. In addition, if Treg cells were involved in exacerbated arthritis in CCR5−/− mice, we would have expected an increase in PG-specific T cell and B cell responses; however, these responses were similar in arthritic WT and CCR5−/− mice.
The ability to remove chemokines in a timely manner to dampen chemokine activity and stop leukocyte infiltration could be therapeutically useful in RA for the resolution of inflammation. Treatment with soluble CCR5 or with cells engineered to express nonsignaling CCR5 might function as a chemokine-scavenging device for the treatment of RA or other chronic autoimmune inflammatory diseases.
In conclusion, we have demonstrated that CCR5 is unnecessary for the initiation of PGIA but rather plays a role in the clearance of proinflammatory chemokines to resolve inflammation.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Finnegan had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Doodes, Finnegan.
Acquisition of data. Doodes, Cao, Hamel, Wang, Rodeghero, Kobezda.
Analysis and interpretation of data. Doodes, Finnegan.
We wish to thank Dr. Tibor T. Glant for providing purified human proteoglycan and for his valuable discussions during the performance of this study.