Ablation of the Ccr2 gene exacerbates polyarthritis in interleukin-1 receptor antagonist–deficient mice

Abstract

Objective

The pathogenesis of rheumatoid arthritis (RA) involves cytokines and chemokines. Given the role of intraarticular macrophage infiltration in RA, this study was undertaken to address the pathogenic role of CCR2, a chemokine receptor that is abundantly expressed by macrophages, in Il1rn-deficient mice, a mouse model of RA.

Methods

Il1rn-deficient and Il1rn and Ccr2–double-deficient mice were subjected to clinical assessment of arthritis and histologic examination. Bone mineral density was measured with computed tomography. The types of cells infiltrating joints were determined by immunohistochemical analysis and flow cytometric analysis. Osteoclasts in joints were quantified after tartrate-resistant acid phosphatase staining. Cytokine and chemokine levels were measured by enzyme-linked immunosorbent assay and multiplex suspension array assay. The expression patterns of chemokines and osteoclastogenic factors were determined by double-color immunofluorescence analysis. Anti-mouse CXCR2 antibody was injected into Il1rn and Ccr2–double-deficient mice for blocking experiments.

Results

Ablation of the Ccr2 gene actually exacerbated arthritis and intraarticular osteoclastogenesis, while it enhanced intraarticular neutrophil but not macrophage accumulation in Il1rn-deficient mice. Infiltrated neutrophils expressed the osteoclastogenic factors RANKL and ADAM-8, thereby augmenting intraarticular osteoclastogenesis in Il1rn and Ccr2–double-deficient mice. Moreover, the double-deficient mice exhibited enhanced expression of the neutrophilic chemokines keratinocyte chemoattractant and macrophage inflammatory protein 2 (MIP-2), compared with Il1rn-deficient mice. Finally, neutralizing antibodies to CXCR2, the receptor for keratinocyte chemoattractant and MIP-2, dramatically attenuated arthritis in Il1rn and Ccr2–double-deficient mice.

Conclusion

Our findings indicate that CCR2-mediated signals can modulate arthritis in Il1rn-deficient mice by negatively regulating neutrophil infiltration.

Rheumatoid arthritis (RA), a chronic inflammatory disorder of synovium, cartilage, and bone, affects ∼1% of the population and is associated with substantial morbidity and increased mortality (1). In RA, various inflammatory cells, such as macrophages, neutrophils, and lymphocytes, infiltrate into the synovium and periarticular space. Among inflammatory cells, macrophages play a major pathogenic role in RA because they are a major cellular source of tumor necrosis factor α (TNFα), interleukin-1α (IL-1α), and IL-1β, the cytokines that are presumed to be crucially involved in the pathogenesis and/or progression of RA (1, 2). Indeed, macrophage numbers in the synovium correlate well with clinical symptoms and joint damage in RA (3). Moreover, various therapies have been shown to attenuate RA symptoms and reduce synovial macrophage numbers (4). Thus, inhibition of intrasynovial macrophage accumulation can prevent disease progression in RA.

Chemokines represent a large superfamily of small proteins that can regulate trafficking and activation of various types of cells, particularly leukocytes, under both inflammatory and homeostatic conditions. Chemokines can exert their actions by binding their cognate G protein–coupled receptors expressed on target cells. Thus, the types of expressed chemokine receptors can determine the responsiveness of leukocytes to particular chemokines (5). Macrophages express a number of chemokine receptors, particularly abundantly CCR2, a selective receptor for a major macrophage-tropic chemokine, CCL2/monocyte chemoattractant protein 1 (MCP-1). Mirroring abundant macrophage infiltration into joint cavities, CCL2 levels have been shown to be markedly increased in RA joints (6), and inhibition of either CCL2 or CCR2 has been shown to ameliorate symptoms and signs observed in various rodent models of arthritis (7, 8). Moreover, recent studies have demonstrated that CCL2 can promote osteoclastogenesis in vitro (9, 10). These observations raised the possibility that CCR2 could be a target for regulating arthritis. However, clinical intervention with anti-CCR2 or anti-CCL2 antibody failed to induce the expected improvement when given to RA patients (11, 12). Similarly, CCR2 antagonist failed to improve anticollagen antibody–induced arthritis (13). Furthermore, Quinones and colleagues (14) demonstrated that Ccr2 gene ablation even aggravated collagen-induced arthritis. These discrepancies may be explained by an observed dual role of CCR2 during the initiation and progression of collagen-induced arthritis (15), but require further investigation.

Although IL-1α and IL-1β are transcribed from two distinct genes, they have similar biologic activities after binding to a common IL-1 receptor and can behave as essential mediators of tissue destruction in various inflammatory disorders, such as sepsis and RA (16, 17). IL-1 receptor antagonist (IL-1Ra; gene name Il1rn) is a member of the IL-1 family and exhibits a β-pleated sheet structure identical to that of IL-1α and IL-1β. IL-1Ra can bind to IL-1 receptor at the same sites with an affinity similar to that of IL-1α and IL-1β (18) but cannot associate with IL-1 receptor accessory protein, which is indispensable for the biologic activities of IL-1α and IL-1β (19, 20). Thus, IL-1Ra is a natural competitive antagonist of IL-1α and IL-1β and can negatively regulate their bioactivities in various inflammatory conditions. Combination therapy consisting of methotrexate and recombinant human IL-1Ra protein has been shown to provide significantly greater clinical benefits to patients with RA than methotrexate alone (21). Moreover, analysis of human IL1RN gene polymorphisms revealed that homozygosity for IL1RN*2 was associated with lower plasma IL-1Ra levels (22) and with an increased number of affected articular areas in RA patients (23). Furthermore, polyarthritis spontaneously develops in Il1rn-deficient BALB/c mice and resembles human RA (24). Thus, Il1rn-deficient BALB/c mice are a good animal model of RA.

Gene expression profile studies have revealed that the joints of Il1rn^−/− mice exhibited enhanced expression of messenger RNA (mRNA) for chemokines, including CCL2/MCP-1, CCL8/MCP-2, CCL7/MCP-3, CCL5, CXCL1, and CXCL12 (25). Moreover, further analysis demonstrated enhanced expression of the genes encoding receptors for these chemokines, such as CCR2, CCR5, CXCR2, and CXCR4 (25). In order to delineate the pathogenic roles of CCR2-mediated signals in polyarthritis, we generated mice lacking both Il1rn and Ccr2 genes, by mating Il1rn^−/− mice with Ccr2^−/− mice. Intriguingly, osteoclastogenesis and polyarthritis were enhanced in these double-deficient mice compared with Il1rn^−/− mice. Thus, CCR2-mediated signals could dampen pathologic osteoclastogenesis in arthritic joints.

MATERIALS AND METHODS

Mice.

BALB/c mice were purchased from Charles River Japan and were designated as wild-type mice. Il1rn^−/− mice were generated as previously described (26). Ccr2^−/− mice were a kind gift from Dr. William Kuziel (University of Texas San Antonio) (27). Ccr2^−/− mice were backcrossed to BALB/c mice for 8–10 generations. Il1rn^−/−Ccr2^−/− mice were generated by crossing Il1rn^−/− mice with Ccr2^−/− mice. All animal experiments were performed under specific pathogen–free conditions in accordance with the Guidelines for the Care and Use of Laboratory Animals of Kanazawa University.

Assessment of arthritis signs and histologic examination.

Each limb was scored weekly on a scale of 0–3, as previously described (28). Ankle joints were obtained from 8-, 12-, 16- and 20-week-old mice. Joints were fixed in 4% paraformaldehyde, decalcified in 10% EDTA-4Na, and embedded in paraffin. Sections were stained with hematoxylin and eosin. Histopathologic changes were scored as previously described (28) by an examiner without any previous knowledge of the genetic background of the mice. Joint sections were stained for tartrate-resistant acid phosphatase (TRAP) with a leukocyte acid phosphatase kit, according to the recommendations of the manufacturer (Sigma-Aldrich). Slides were counterstained with methyl green. Osteoclasts were identified as TRAP-positive cells with at least 2 nuclei and were determined separately in synovial tissue and subchondral bone marrow. Some paraffin-embedded sections were stained with rat anti-mouse F4/80, anti-mouse CD3 (Serotec), or anti-mouse Ly-6G antibodies (BD Biosciences). The sections were then incubated with biotinylated rabbit anti-rat IgG (Dako). The immune complexes were visualized using a catalyzed signal amplification system (Dako) or the Elite avidin–biotin–peroxidase and diaminobenzidine substrate kits (Vector). When rat IgG (Cosmo Bio) was used instead of specific primary antibodies as a negative control, no staining was observed (results not shown), indicating the specificity of the reactions. Positive cell numbers were determined in 5 randomly chosen fields at 400× magnification in each synovial tissue section.

Flow cytometric analysis.

Bone marrow cells were obtained from femoral and tibial bones of 8- or 20-week-old mice. Joints were obtained from 10- or 20-week-old animals and were digested with type I collagenase (Wako). The single-cell suspensions from bone marrow or joints were stained with various combinations of fluorescein isothiocyanate (FITC)–conjugated anti-CD3, phycoerythrin (PE)–conjugated anti-CD4, allophycocyanin (APC)–conjugated anti-CD8, PE-conjugated anti-B220, APC-conjugated CD11c, FITC-conjugated anti–Ly-6C (Abcam), PE-conjugated anti–Ly-6G, and APC-conjugated anti-CD11b monoclonal antibodies (eBioscience). Cell surface marker expression was analyzed using a FACSCalibur instrument (BD Biosciences) with CellQuest analysis software (BD Biosciences).

Double-color immunofluorescence analysis.

Paraffin-embedded joint sections from 20-week-old Il1rn^−/−Ccr2^−/− mice were stained with various combinations of goat anti-RANKL (R&D Systems), rabbit anti–keratinocyte chemoattractant (BioVision), goat anti–macrophage inflammatory protein 2 (anti–MIP-2; R&D Systems), rat anti–Ly-6G, rat anti-CD3, and rat anti-F4/80 antibodies. Cells infiltrating the joints were isolated as described above and were centrifuged to prepare cytospin sections. After fixation in methanol, sections were incubated with goat anti–ADAM-8 (Santa Cruz Biotechnology), together with anti–Ly-6G, anti-CD3, or anti-F4/80 antibodies. After washing, the sections were incubated with Alexa Fluor 488–labeled donkey anti-rat IgG antibodies and Alexa Fluor 546–labeled donkey anti-goat IgG or Alexa Fluor 594–labeled donkey anti-rabbit IgG antibodies (Molecular Probes). Immunofluorescence was visualized in a dual channel mode on a fluorescence microscope.

Assessment of bone mineral density (BMD).

Total BMD of the 4 limbs of 20-week-old mice was determined using a Latheta computed tomography system, according to the recommendations of the manufacturer (Aloka).

Determination of cytokine and protein levels.

A cartilage oligomeric matrix protein (COMP) assay kit (AnaMar Medical) and TRAP5b assay kit (Immunodiagnostic Systems) were used to determine serum COMP and TRAP5b levels, respectively. Joint tissue samples were obtained from animals at the indicated time points and homogenized with radioimmunoprecipitation assay buffer (Santa Cruz Biotechnology) containing proteinase inhibitor cocktail and centrifuged to obtain supernatants. After determining total protein content with a BCA kit (Pierce), levels of macrophage colony-stimulating factor (M-CSF), RANKL (R&D Systems), and ADAM-8 (Cusabio Biotech) were determined using specific enzyme-linked immunosorbent assay kits for these molecules, according to the recommendations of the manufacturers. IL-17A, granulocyte colony-stimulating factor (G-CSF), keratinocyte chemoattractant, and MIP-2 levels were determined using a multiplex suspension array system according to the recommendations of the manufacturer (Bio-Rad). The data are expressed as the target molecule (pg) per total protein (mg) for each sample.

Treatment with neutralizing antibody against mouse CXCR2.

Anti-mouse CXCR2 rabbit IgG were prepared as previously described (29) and used for blocking experiments. Il1rn^−/−Ccr2^−/− mice were injected intraperitoneally with 200 μg of anti-mouse CXCR2 rabbit IgG or normal rabbit IgG, both twice per week from 12 weeks of age until 16 weeks of age. Mice were killed 24 hours after the last injection, and joint pathology and the number of infiltrating neutrophils were analyzed as described above.

Statistical analysis.

Data were analyzed using one-way analysis of variance followed by Fisher's protected least significant difference. Mann-Whitney U test or Kruskal-Wallis test was used in instances when the data was not normally distributed. P values less than 0.05 were considered significant.

RESULTS

Aggravation of spontaneously developed polyarthritis in Il1rn^−/− mice after Ccr2 deletion.

Consistent with previous observations (24), Il1rn-deficient mice spontaneously developed polyarthritis until 12 weeks of age. Ccr2^−/− mice did not exhibit any signs indicative of arthritis until 1 year of age (data not shown).

Body weight gain was retarded in Il1rn^−/−Ccr2^−/− mice after 14 weeks of age, compared to Il1rn^−/− mice (Figure 1A). The incidence and the time of onset of disease were comparable between these two strains (data not shown). However, after 16 weeks of age, Il1rn^−/−Ccr2^−/− mice had higher arthritis scores, with extensive swelling and severe ankylosis in multiple joints, compared to Il1rn^−/− mice, which had mild joint inflammation restricted to the hind paws (Figure 1A). Wild-type and Ccr2^−/− mice showed no histologic signs of arthritis until 1 year of age (data not shown). Histologic signs of synovial inflammation appeared in Il1rn^−/− and Il1rn^−/−Ccr2^−/− mice at 8 weeks of age, with no apparent differences between the two strains until 12 weeks of age. Signs of joint inflammation, such as cartilage damage, bone erosion, and cell infiltration, were augmented in Il1rn^−/−Ccr2^−/− mice ages 16 weeks and older compared with Il1rn^−/− mice (Figure 1B). Computed tomography consistently demonstrated greater decreases in BMD in Il1rn^−/−Ccr2^−/− mice than in Il1rn^−/− mice (Figure 1C). Moreover, the serum concentration of COMP, a marker of cartilage destruction, was increased to a greater extent in Il1rn^−/−Ccr2^−/− mice than in Il1rn^−/− mice (Figure 1D). These observations indicate that the lack of a Ccr2 gene exacerbated polyarthritis in Il1rn^−/− mice, together with enhanced intraarticular inflammation and bone destruction.

Figure 1. Exacerbation of arthritis in Il1rn^−/− mice by Ccr2 gene ablation. A, Body weight and arthritis score were determined weekly. Values are the mean ± SD (n = 40 Il1rn^−/− mice and 50 Il1rn^−/−Ccr2^−/− mice). ∗ = P < 0.05; # = P < 0.01. B, Left, Mouse ankle joints were removed at the indicated time intervals, fixed, and stained with hematoxylin and eosin. Results shown are representative of 5 animals for each genotype. Bars = 100 μm. Right, Cartilage damage, bone erosion, and cell infiltration were scored in each mouse. Values are the mean ± SD (n = 5 animals per group). ∗ = P < 0.05. C, Total bone mineral density (BMD) of the fore limbs of 20-week-old mice was measured using computed tomography. Values are the mean ± SD (n = 16–55 for each genotype). ∗ = P < 0.05; ∗∗∗ = P < 0.001. D, Serum levels of cartilage oligomeric matrix protein (COMP) in 20-week-old mice were determined by enzyme-linked immunosorbent assay. Horizontal lines show the mean; diamonds represent individual animals (n = 6–10 animals per group). ∗ = P < 0.05. WT = wild-type.

Enhanced neutrophil infiltration in Il1rn^−/− mice with Ccr2 deficiency.

The enhanced intraarticular cell infiltration in Il1rn^−/−Ccr2^−/− mice prompted us to investigate the types of cells infiltrating joints. Immunohistochemical analyses revealed that a small number of CD3-positive T cells and F4/80-positive macrophages infiltrated the joints in Il1rn^−/− and Il1rn^−/−Ccr2^−/− mice 8 weeks after birth and that their numbers increased marginally and to a similar extent in Il1rn^−/− and Il1rn^−/−Ccr2^−/− mice thereafter (Figure 2A). (Additional results are available online at http://kanazawa-u.ac.jp/∼ganken/binsiseitai/Suppl-Fig-IL-1ra.pdf.) In contrast, Ly-6G–positive neutrophils markedly infiltrated the joints in Il1rn^−/− and Il1rn^−/−Ccr2^−/− mice beginning 12 weeks after birth, and the increase was more evident in Il1rn^−/−Ccr2^−/− mice than in Il1rn^−/− mice (Figure 2A). (Additional results are available online at http://www.kanazawa-u.ac.jp/∼ganken/bunsiseitai/Suppl-Fig-IL-1ra.pdf.) Flow cytometric analyses of infiltrating cells in the joints consistently revealed an augmented increase in the relative and absolute numbers of Ly-6G–positive neutrophils, but not F4/80-positive macrophages, CD4-positive T lymphocytes, CD8-positive T lymphocytes, B220-positive B lymphocytes, or CD11c-positive dendritic cells in Il1rn^−/−Ccr2^−/− mice, compared with Il1rn^−/− mice (Figures 2B and C). These observations indicate that the absence of the Ccr2 gene selectively augmented intraarticular neutrophil infiltration.

Figure 2. Enhanced intraarticular neutrophil accumulation in Il1rn^−/−Ccr2^−/− mice. A, The numbers of CD3+, F4/80+, and Ly-6G+ cells in mouse ankle joints were enumerated as described in Materials and Methods. Values are the mean ± SD (n = 6 animals per group). ∗ = P < 0.05. B, Single-cell suspensions were obtained from the joint cavities of 10-week-old or 20-week-old Il1rn^−/− mice and Il1rn^−/−Ccr2^−/− mice and were stained with various combinations of antibodies and analyzed by fluorescence-activated cell sorting as described in Materials and Methods. Ly-6G and Ly-6C profiles in CD11b+ cells from joints were determined. Results shown are representative of 5 animals per group. SSC = side scatter. C, The absolute cell number in each fraction was calculated in Il1rn^−/− and Il1rn^−/−Ccr2^−/− mice. Values are the mean ± SD (n = 5 animals per group). ∗ = P < 0.05.

Increased osteoclast numbers in Il1rn^−/− mice with Ccr2 deficiency.

Enhanced bone destruction in Il1rn^−/−Ccr2^−/− mice prompted us to determine the numbers of osteoclasts, the cell which is mainly responsible for bone resorption. TRAP staining demonstrated that osteoclast numbers in joints were increased progressively in Il1rn^−/− and Il1rn^−/−Ccr2^−/− mice and that the increase was more evident in Il1rn^−/−Ccr2^−/− mice than in Il1rn^−/− mice (Figure 3A). (Additional results are available online at http://www.kanazawa-u.ac.jp/∼ganken/bunsiseitai/Suppl-Fig-IL-1ra.pdf.) Moreover, the serum concentration of TRAP5b, an indicator of osteoclast activity, was increased in Il1rn^−/−Ccr2^−/− mice to a greater extent than in Il1rn^−/− mice (Figure 3B).

Figure 3. Increased numbers of intraarticular osteoclasts in Il1rn^−/−Ccr2^−/− mice. A, The numbers of tartrate-resistant acid phosphatase (TRAP)–positive osteoclasts were counted as described in Materials and Methods. Values are the mean ± SD (n = 5 animals per group). B, The serum TRAP5b concentration in 20-week-old mice was determined by enzyme-linked immunosorbent assay. Values are the mean ± SD (n = 6–15 animals per group). ∗ = P < 0.05. WT = wild-type.

Clinical studies of RA patients have implicated bone edema as an important parameter for predicting prognosis (30). Because bone edema is pathologically characterized by osteoclast infiltration into subchondral bone marrow (31), we determined TRAP-positive cell numbers in subchondral bone marrow. Indeed, TRAP-positive cell numbers were increased in the subchondral bone marrow in Il1rn^−/−Ccr2^−/− mice to a greater extent than in Il1rn^-/- mice (Figure 4A). (Additional results are available online at http://www.kanazawa-u.ac.jp/∼ganken/bunsiseitai/Suppl-Fig-IL-1ra.pdf.) Several lines of evidence indicate the essential roles of Ccr2-mediated signals in the egress of macrophages, a precursor of osteoclasts, from bone marrow. Ly-6G^lowLy-6C–positive macrophage numbers in bone marrow were consistently increased in Il1rn^−/−Ccr2^−/− mice compared with Il1rn^−/− mice (Figure 4B). (Additional results are available online at http://www.kanazawa-u.ac.jp/∼ganken/bunsiseitai/Suppl-Fig-IL-1ra.pdf.) These observations indicate that Ccr2 deficiency increased the numbers of intra–bone marrow macrophages, a precursor of osteoclasts, and simultaneously increased the number of osteoclasts in the synovium and bone marrow.

Figure 4. Increased numbers of osteoclast (OC) precursors and osteoclastogenic factors in Il1rn^−/−Ccr2^−/− mice. A, The numbers of tartrate-resistant acid phosphatase–positive OCs in mouse subchondral bone marrow (BM) regions were enumerated as described in Materials and Methods. Values are the mean ± SD (n = 5 animals per group). ∗ = P < 0.05; ∗∗∗ = P < 0.01. NS = not significant; WT = wild-type. B, Bone marrow cells were obtained from 8-week-old or 20-week-old mice and were stained with a combination of phycoerythrin-labeled anti-CD11b, allophycocyanin-labeled anti–Ly-6G, and fluorescein isothiocyanate–labeled anti-Ly-6C antibodies. The proportion of CD11b+Ly-6G^intermediateLy-6C+ monocytes was enumerated based on the results of fluorescence-activated cell sorting analysis. Values are the mean ± SD (n = 5 animals per group). ∗ = P < 0.05; # = P < 0.01. C, Macrophage colony-stimulating factor (M-CSF), RANKL, and ADAM-8 levels were determined in joints obtained from 20-week-old animals by using specific enzyme-linked immunosorbent assay as described in Materials and Methods. Values are the mean ± SEM (n = 15–20 animals per group). ∗ = P < 0.05; # = P < 0.01. D, Ankle joints were removed from 20-week-old Il1rn^−/−Ccr2^−/− mice and were stained with the combination of anti-RANKL and anti–Ly-6G, anti-RANKL and anti-CD3, or anti-RANKL and anti-F4/80 antibodies, as described in Materials and Methods. Cytospin sections were prepared from joint exudates from 20-week-old Il1rn^−/−Ccr2^−/− mice and were stained with the combination of anti–ADAM-8 and anti–Ly-6G, anti–ADAM-8 and anti-CD3, or anti–ADAM-8 and anti-F4/80 antibodies. Results shown are representative of 3 animals per group. Bars = 20 μm.

Increased expression of osteoclastogenic factors by neutrophils.

Given the observed enhanced osteoclastogenesis, we next determined the intraarticular levels of major osteoclastogenic factors, such as M-CSF, RANKL, and ADAM-8, in the mouse strains. There were no significant differences in intraarticular M-CSF levels among wild-type, Ccr2^−/−, Il1rn^−/−, and Il1rn^−/−Ccr2^−/− mice (Figure 4C). In contrast, intraarticular RANKL levels were elevated in Il1rn^−/− and Il1rn^−/−Ccr2^−/− mice, compared with wild-type and Ccr2^−/− mice, and RANKL levels were higher in Il1rn^−/−Ccr2^−/− mice than in Il1rn^−/− mice (Figure 4C). Similar observations were obtained for intraarticular ADAM-8 levels (Figure 4C). We determined the type(s) of cells expressing RANKL and ADAM-8. With double-color immunofluorescence, RANKL was detected mainly in Ly-6G–positive neutrophils and CD3-positive T cells, but not in F4/80-positive macrophages in the synovium in Il1rn^−/−Ccr2^−/− mice (Figure 4D). Likewise, ADAM-8 protein was detected mainly in Ly-6G–positive neutrophils and to a lesser extent in F4/80-positive macrophages in mouse synovial exudates (Figure 4D). These observations suggest that the absence of Ccr2 augmented the infiltration of neutrophils, a major source of RANKL and ADAM-8, and that neutrophil-derived RANKL and ADAM-8 in turn promoted osteoclast differentiation from macrophages, which were retained in subchondral bone marrow as well as in synovial tissue.

Ccr2 deficiency and augmented production of neutrophil-tropic chemokines.

In order to delineate the molecular mechanisms underlying enhanced intraarticular neutrophil infiltration, we assessed the intraarticular levels of IL-17A, G-CSF, keratinocyte chemoattractant, and MIP-2, which can potently mobilize neutrophils. IL-17A and G-CSF levels were increased marginally (<20 pg/mg protein) in both Il1rn^−/− and Il1rn^−/−Ccr2^−/− mice (Figure 5A). In contrast, keratinocyte chemoattractant and MIP-2 levels were markedly increased in Il1rn^−/− mice as compared to wild-type or Ccr2^−/− mice. Moreover, the increment was further augmented in Il1rn^−/−Ccr2^−/− mice. Double-color immunofluorescence revealed that keratinocyte chemoattractant and MIP-2 proteins were expressed on Ly-6G–positive neutrophils (Figures 5B and C) and on F4/80-positive macrophages but not on CD3-positive T cells, suggesting that neutrophils attracted more neutrophils by producing these chemokines in a paracrine manner. (Results are available online at http://www.kanazawa-u.ac.jp/∼ganken/bunsiseitai/Suppl-Fig-IL-1ra.pdf.) These observations indicate that the absence of the Ccr2 gene amplifies the intraarticular keratinocyte chemoattractant and MIP-2 expression induced by Il1rn gene deficiency and that the net result is enhanced neutrophil infiltration in Il1rn^−/−Ccr2^−/− mice.

Figure 5. Augmented chemokine production in Il1rn^−/−Ccr2^−/− mice. A, Interleukin-17A (IL-17A), granulocyte colony-stimulating factor (G-CSF), keratinocyte chemoattractant (KC), and macrophage inflammatory protein 2 (MIP-2) levels were determined in joints obtained from 20-week-old animals as described in Materials and Methods. Values are the mean ± SEM (n = 8–16 animals per group). ∗ = P < 0.05; # = P < 0.01. WT = wild-type. B and C, Ankle joints were removed from 20-week-old Il1rn^−/−Ccr2^−/− mice and were stained with the combinations of anti–keratinocyte chemoattractant and anti–Ly-6G (B) or anti–MIP-2 and anti–Ly-6G (C) antibodies as described in Materials and Methods. Results shown are representative of 3 animals per group. Bars = 20 μm.

Amelioration of arthritis in Il1rn^−/−Ccr2^−/− mice by anti-CXCR2 antibody.

Finally, to determine the involvement of neutrophils in arthritis that developed spontaneously in Il1rn^−/−Ccr2^−/− mice, we injected antibodies against CXCR2, the predominant receptor for keratinocyte chemoattractant and MIP-2, into Il1rn^−/−Ccr2^−/− mice between 12 and 16 weeks of age, after arthritis became prominent. Anti-mouse CXCR2 antibodies markedly decreased the severity of arthritis in mice (Figures 6A and B) and reduced the infiltration of neutrophils into the joint (Figure 6B). Flow cytometric analysis consistently revealed that the number of neutrophils in the joint was dramatically decreased in anti-CXCR2–treated mice (Figure 6C). Moreover, anti-CXCR2 antibodies markedly reduced osteoclast numbers (Figure 6D). These observations indicate that CXCR2-mediated neutrophil infiltration crucially contributed to the progression of arthritis and pathologic osteoclastogenesis in Il1rn^−/−Ccr2^−/− mice.

Figure 6. Amelioration of arthritis development in Il1rn^−/−Ccr2^−/− mice after treatment with anti-CXCR2 antibody (Ab). A, Il1rn^−/−Ccr2^−/− mice received either anti-mouse CXCR2 or control antibodies intraperitoneally twice a week from 12 weeks of age until 16 weeks of age (arrows). Arthritis scores were determined weekly. Values are the mean ± SEM (n = 7 animals per group). ∗ = P < 0.05. B, The hind feet of the IgG-treated and anti-mouse CXCR2 antibody–treated mice were examined and were stained with hematoxylin and eosin at 16 weeks of age. Results shown are representative of 7 animals. Bars = 50 μm. C, Single-cell suspensions were obtained from the joint cavity of 16-week-old mice and were stained with the combination of CD11b and Ly-6G antibodies and analyzed by fluorescence-activated cell sorting as described in Materials and Methods. Absolute cell numbers in the CD11b+Ly-6G+ fraction were calculated. Values are the mean ± SEM (n = 7 animals per group). # = P < 0.01. D, The numbers of tartrate-resistant acid phosphatase–positive osteoclasts were counted as described in Materials and Methods. Values are the mean ± SD (n = 5 animals per group). ∗ = P < 0.05.

DISCUSSION

TNF antagonists have been widely used to treat RA patients, with a remarkable response rate compared with conventional therapy. However, not all RA patients benefit sufficiently from TNF antagonist treatment and, therefore, an additional therapeutic strategy for RA is needed. TNF antagonists reduce the biologic activities of TNFα produced mainly by macrophages infiltrating the synovium of RA patients. Because macrophages can produce additional factors involved in the progression of arthritis (2), a drug that targets macrophages may be promising for the treatment of RA. CCR2, a receptor for a major macrophage-tropic chemokine, CCL2, may also be an attractive target for the treatment of RA, but the role of CCR2 in arthritis is still a subject of controversy. Thus, we examined the effects of Ccr2 gene ablation on arthritis that develops spontaneously in Il1rn^−/− mice, to delineate the roles of CCR2-mediated signals during the course of the disease. We determined that Ccr2 gene deletion, to our surprise, markedly augmented arthritis in Il1rn^−/− mice, as evidenced by enhanced bone destruction, together with enhanced intraarticular neutrophil infiltration. Moreover, infiltrating neutrophils expressed two potent osteoclastogenic factors, RANKL and ADAM-8.

Ccr2^−/− mice exhibit reduced macrophage infiltration in various types of macrophage-mediated inflammatory responses (32–34). However, in this model, Ccr2 gene ablation had few effects on macrophage infiltration into the synovium. Quinones and colleagues (14) actually observed enhanced F4/80-positive cell infiltration into the synovium of Ccr2^−/− mice in a collagen-induced arthritis model. They further observed that Ccr2^−/− mice exhibited enhanced intraarticular expression of CCL3, another potent macrophage-tropic chemokine. We also observed that intraarticular expression of CCL5, the chemokine that shares CCR1 and CCR5 with CCL3, was increased in Il1rn^−/−Ccr2^−/− mice compared with Il1rn^−/− mice (Fujii H, et al: unpublished observations). Thus, compensatory CCL5 production may prevent a decrease in macrophage infiltration. Alternatively, increases in infiltrating neutrophils can maintain macrophage infiltration by producing LL37, an α-defensin and heparin-binding protein, which can attract monocytes (35), in a compensatory manner.

Osteoclasts are the main cell type capable of removing calcium from bone and degrading bone matrix, and they play crucial roles in inflammatory bone erosion. In the first step of inflammatory bone erosion, osteoclast precursor cells, monocyte/macrophages, accumulate near the site of bone erosion (36). Il1rn^−/− mice retained a larger number of macrophages in the bone marrow than did wild-type mice. Genetic deletion of Ccr2 resulted in further retention of macrophages in bone marrow. This may mirror the biologic activity of the CCL2–CCR2 axis to induce macrophage migration from bone marrow (37). The accumulated macrophages may serve as osteoclast precursor cells. The survival and differentiation of osteoclast progenitor cells depend on several factors, particularly M-CSF and RANKL (38). Intraarticular levels of RANKL, but not M-CSF, were markedly increased in Il1rn^−/− mice, and the increase was further augmented in Il1rn^−/−Ccr2^−/− mice. In addition to RANKL and M-CSF, a member of the ADAM family, ADAM-8, can induce osteoclast differentiation, particularly acting during its later phase (39). Moreover, ADAM-8 expression is enhanced in RA pannus adjacent to bone erosion (40), suggesting that it crucially contributes to bone erosion in RA. Intraarticular ADAM-8 levels were increased in Il1rn^−/− and Il1rn^−/−Ccr2^−/− mice, similar to RANKL. Thus, it is probable that RANKL and ADAM-8 enhance osteoclastogenesis in Il1rn^−/− mice and more so in Il1rn^−/−Ccr2^−/− mice, thereby augmenting bone destruction in these mice.

Recent advances in magnetic resonance imaging have allowed investigators to determine that bone edema precedes the onset of bone erosion in RA (30). Histologic studies demonstrated that bone edema represents a cellular infiltrate comprising osteoclasts and macrophages within bone, particularly in subchondral bone marrow areas (41). Likewise, in the present study osteoclasts and macrophages accumulated in bone marrow in Il1rn^−/−Ccr2^−/− mice and to a lesser degree in Il1rn^−/− mice. Thus, enhanced RANKL and ADAM-8 expression can also induce osteoclast generation in bone marrow, which contained a large number of macrophages, an osteoclast progenitor cell.

Previous studies have shown that neutrophil infiltration was augmented in various disease models in mice deficient in either Ccr2 or its ligand, CCL2 (42, 43), consistent with the findings of the present study. Several independent groups demonstrated that the expression of the neutrophil-tropic chemokines keratinocyte chemoattractant/CXCL1 and MIP-2/CXCL2 were enhanced in the absence of either CCR2 or CCL2 (14, 42, 43). We detected the enhanced expression of IL-17A and G-CSF, the cytokines with neutrophil-mobilizing activity, in addition to CXCL1 and CXCL2 in the joints of Il1rn^−/− mice. However, only CXCL1 and CXCL2 expression was enhanced to a greater degree in Il1rn^−/−Ccr2^−/− mice than in Il1rn^−/− mice. This may mirror our observation that the intraarticular expression of IL-1β, a potent inducer of CXCL1 and CXCL2 expression, was enhanced in Il1rn^−/−Ccr2^−/− mice and to a lesser degree in Il1rn^−/− mice (Fujii H, et al: unpublished observations).

Circulating RANKL levels were not increased in this model (Fujii H, et al: unpublished observations), indicating that RANKL was produced locally in joints, as has similarly been observed in Streptococcus pyogenes–induced septic arthritis (44), which is characterized by a massive intraarticular neutrophil infiltration. Thus, infiltrating neutrophils can produce RANKL. This assumption is supported by previous observations that RANKL is expressed abundantly by RA patient–derived neutrophils and lipopolysaccharide-stimulated neutrophils from normal volunteers (45, 46). Similarly, intraarticularly infiltrating neutrophils were a rich source of RANKL in the present model.

In association with the exacerbation of active RA, a large number of neutrophils frequently infiltrate synovial space via the pannus (47) and can promote joint damage by releasing reactive oxygen species and lysosomal enzymes. Moreover, ADAM-8 mRNA has been detected in human granulocytes and monocytes (48, 49), while ADAM-8 protein is constitutively present on the cell surface and in intracellular granules of human neutrophils and upon activation can be mobilized from the granules to the plasma membrane (49). Likewise, we detected ADAM-8 expression mainly in neutrophils and to a lesser degree in macrophages. The results of the present study indicate that the infiltration of neutrophils was amplified in the absence of CCR2 and that infiltrating neutrophils may additionally produce two potent osteoclastogenic molecules, RANKL and ADAM-8, and eventually induce osteoclastogenesis, thereby accelerating bone destruction.

Several lines of evidence indicate that CCL2 can directly act on osteoclasts or osteoclast progenitor cells to induce osteoclastogenesis (9, 10). Thus, bone destruction in Il1rn^−/− mice should hypothetically be alleviated by genetic ablation of Ccr2, a single and specific receptor for CCL2. However, ablation of the Ccr2 gene markedly increased the infiltration of neutrophils, a cellular source of reactive oxygen species and osteoclastogenic molecules, and appears to overcome the direct effect of the CCL2–CCR2 axis on osteoclastogenesis. Moreover, neutrophils are prone to become apoptotic, particularly after being activated and infiltrating, and should be removed by macrophages to prevent tissue injury. This is supported by the results of our previous study indicating that macrophages can ingest dead neutrophils to promote tissue repair in pneumonia arising from Pseudomonas aeruginosa infection, in a CCL2-dependent manner (50). Thus, it is probable that macrophages can remove intraarticular infiltrating activated neutrophils, thereby limiting joint damage in a CCR2-dependent manner. All of these observations may indicate a protective role for CCR2-expressing macrophages in inflammatory joint disease.

AUTHOR CONTRIBUTIONS

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. Mukaida 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. Baba, Mukaida.

Acquisition of data. Fujii, Baba, Mukaida.

Analysis and interpretation of data. Fujii, Baba, Ishida, Kondo, Yamagishi, Kawano, Mukaida.

Acknowledgements

We express our gratitude to Dr. Joost J. Oppenheim (National Cancer Institute–Frederick, Frederick, MD) for critical review of the manuscript. We thank Dr. William Kuziel for providing us with Ccr2-deficient mice.