Endogenous Myeloperoxidase Is a Mediator of Joint Inflammation and Damage in Experimental Arthritis

Authors


Abstract

Objective

Myeloperoxidase (MPO) is implicated as a local mediator of tissue damage when released extracellularly in many chronic inflammatory diseases. The purpose of this study was to explore the role of endogenous MPO in experimental rheumatoid arthritis (RA).

Methods

K/BxN serum–transfer arthritis was induced in C57BL/6 wild-type (WT) and MPO knockout (MPO−/−) mice, and disease development was assessed. MPO activity was measured in joint tissues from mice with or without K/BxN arthritis. Collagen-induced arthritis (CIA) was induced in WT and MPO−/− mice, and disease development and immune responses were examined. MPO expression was assessed in synovial biopsy samples from patients with active RA, and the effect of MPO on synovial fibroblasts was tested in vitro.

Results

MPO was up-regulated in the joints of mice with K/BxN arthritis, and MPO deficiency attenuated the severity of the disease without affecting circulating cytokine levels. In CIA, MPO−/− mice had enhanced CD4+ T cell responses and reduced frequency of regulatory T cells in the lymph nodes and spleen, as well as augmented interleukin-17A and diminished interferon-γ secretion by collagen-stimulated splenocytes, without an effect on circulating anticollagen antibody levels. Despite enhanced adaptive immunity in secondary lymphoid organs, CIA development was attenuated in MPO−/− mice. Intracellular and extracellular MPO was detected in the synovium of patients with active RA, and human MPO enhanced the proliferation and decreased the apoptosis of synovial fibroblasts in vitro.

Conclusion

MPO contributes to the development of arthritis despite suppressing adaptive immunity in secondary lymphoid organs. This suggests distinct effects of local MPO on arthritogenic effector responses.

Rheumatoid arthritis (RA) is a common chronic autoimmune disease characterized by inflammation and destruction of joints. Both adaptive and innate immunity contribute to the development of rheumatoid inflammation. Many cell types, including CD4+ T cells, B cells, macrophages, and neutrophils, accumulate in the inflamed joints of RA patients, where they influence disease development ([1-4]). The importance of CD4+ T cells in RA has been shown in humans with the disease as well in experimental arthritis, where CD4-directed strategies have successfully attenuated disease ([3, 5, 6]). Likewise, the critical role of B cells as producers of pathogenic autoantibodies has been demonstrated in humans with RA and in models of the disease ([4, 7]). Macrophages, one of the major cell types found in the inflamed synovial lining, are also well known for their proinflammatory effects in experimental and human RA ([1]). The role of neutrophils in RA has been less extensively investigated, but several reports suggest that they play an important role in the development of inflammatory joint injury. Neutrophils are the most abundant immune cell type found in rheumatoid synovial fluid ([2]), and the therapeutic efficacy of some antiarthritic drugs, such as methotrexate and tumor necrosis factor (TNF) inhibitors, is associated with reduced neutrophil accumulation and/or function in the joint ([8, 9]). A critical role of neutrophils as effector cells mediating joint inflammation and damage has been demonstrated in models of RA, including K/BxN serum–transfer arthritis and collagen-induced arthritis (CIA) ([10, 11]).

Myeloperoxidase (MPO) is a heme-containing enzyme stored in primary granules of neutrophils ([12]). It is the major neutrophil protein, making up ∼5% of the dry cell weight ([12]). MPO is also present to a lesser extent in monocytes and some macrophages ([12, 13]). In the presence of hydrogen peroxide (H2O2) and a halide (chloride, bromide, or thiocyanate), MPO catalyzes the formation of powerful reactive intermediates, including hypochlorous (HOCl), hypobromous, and hypothiocyanous acids, respectively, which can have large biologic effects by altering proteins, lipids, and/or DNA ([12, 13]).

The MPO/HOCl system plays an important role in intracellular microbial killing by neutrophils ([12, 13]). However, MPO can be also released to the extracellular milieu after leukocyte activation and, through the formation of reactive intermediates, can cause tissue damage at sites of inflammation ([12, 13]). Thus, MPO has been implicated in the pathogenesis of multiple inflammatory diseases, including atherosclerosis, cardiovascular disease, kidney disease, cystic fibrosis, and multiple sclerosis (MS) ([12, 13]). A role of MPO as a local mediator of tissue damage has also been demonstrated in models of cardiovascular, renal, and lung disease ([14-16]). These findings have implicated MPO as an important therapeutic target in the treatment of inflammatory conditions.

In contrast to its injurious local effects at sites of inflammation, recent reports have demonstrated that MPO can suppress the generation of adaptive immune responses in secondary lymphoid organs ([16-18]). In a model of MS, MPO inhibited lymphocyte proliferation in draining lymph nodes (LNs) and attenuated disease ([17]). Similarly, we have shown in experimental glomerulonephritis that MPO suppresses the induction of nephritogenic CD4+ T cell responses in the spleen ([16]).

Several lines of evidence have implicated MPO in the pathogenesis of RA as a local mediator of joint damage. MPO is released by activated neutrophils in RA synovial fluid, where enhanced levels of enzymatically active MPO correlate with the presence of HOCl-modified proteins ([19, 20]). Increased concentrations of active MPO are also found in the inflamed joints of mice with experimental RA ([21]), and injection of MPO into the joints increases the severity of arthritis induced by streptococcal cell wall fragments ([22]). However, the contribution of endogenous MPO to the pathogenesis of inflammatory arthritis is largely unknown. The purpose of the present study was to explore the role of endogenously expressed MPO in experimental RA.

We found that MPO plays a pathogenic role in joint inflammation and damage in neutrophil-dependent, T cell–independent K/BxN serum–transfer arthritis. Our findings also show that endogenous MPO contributes to the development of CIA despite suppressing adaptive arthritogenic immunity in secondary lymphoid organs. In addition, they demonstrate that intracellular and extracellular MPO is present in the RA synovial lining and that human MPO expands synovial fibroblasts in vitro. These findings point to a local role of MPO as a proinflammatory effector in arthritis, independent of its inhibitory effects on adaptive immune responses.

MATERIALS AND METHODS

Animals

Mice used in these experiments were 6–8-week-old males. C57BL/6 wild-type (WT) mice were obtained from Monash University Animal Services. MPO knockout (MPO−/−) mice were backcrossed onto the C57BL/6J background for ≥10 generations ([23]). Mice were bred and kept at Monash Medical Centre Animal Facilities under specific pathogen–free conditions. All studies adhered to National Health and Research Council of Australia guidelines for animal experimentation and were approved by the Monash University Ethics committee.

Induction of K/BxN serum–transfer arthritis

WT mice and MPO−/− mice (n = 6 per group) were injected intraperitoneally with K/BxN serum (from A. Cook, Department of Medicine, University of Melbourne, Melbourne, Victoria, Australia) on days 0 and 2 (38 μl/mouse), and clinical disease was monitored until day 7. Paw thickness was measured daily with calipers (Mitutoyo), and the results were expressed as the change in joint thickness (in mm) relative to that measured on day 0 (before serum injection). Each limb was scored daily on a scale of 0 (no redness or swelling) to 3 (severe redness and swelling). The scores in the 4 limbs were added to obtain the clinical index (maximum score 12). For histologic assessment, ankle joints were formalin-fixed, decalcified, and paraffin-embedded, and sagittal sections were stained with Safranin O (with fast green/iron hematoxylin counterstain). Sections were scored on a scale of 0–3 for joint tissue inflammation, synovitis, joint space exudate, cartilage degradation, and bone damage, and a total histologic score was calculated from the sum of these parameters, as described previously ([24]).

Induction of CIA

CIA was induced as described elsewhere ([25]). Native chick type II collagen (100 μg; Chondrex) dissolved in 0.05M acetic acid was emulsified with Freund's complete adjuvant (CFA; Sigma) containing 5 mg/ml of heat-killed Mycobacterium tuberculosis (H37Ra; Difco) and injected subcutaneously into the tail of WT mice (n = 10) and MPO−/− mice (n = 8). Mice were given a subcutaneous booster injection with chick collagen/CFA (day 21) and monitored daily for clinical features of arthritis (days 22–45). Clinical arthritis severity was graded by scoring each limb on a scale of 0–3, where 0 = no erythema or swelling, 0.5 = very mild swelling and erythema in 1–2 digits, 1 = mild swelling and erythema in >2 digits, 2 = moderate swelling and erythema involving the entire limb or multiple limbs, and 3 = pronounced swelling leading to incapacitated limbs. The total clinical score was determined by adding the individual scores for each limb. Arthritis was considered to be present if the total clinical score was ≥0.5.

T cell and B cell activation, proliferation, and apoptosis

Single-cell suspensions were obtained from LNs and spleen. CD4+ T cell and B cell (CD19+) activation was measured by flow cytometric analysis of CD44 and CD69 expression, as described previously ([16]). Apoptosis was measured by annexin V and propidium iodide (PI) staining ([16]). Treg cells were identified as FoxP3+CD4+CD25+ cells by flow cytometry ([26]). For assessment of lymphocyte proliferation, splenocytes were cultured for 72 hours in the presence of chick collagen (10 μg/ml). Proliferation was measured by the incorporation of 3H-thymidine (2.5 μCi/ml, which was added during the last 18 hours). For specific measurement of CD4+ T cell and B cell proliferation, LN and spleen cells were cultured for 72 hours with chick collagen (20 μg/ml). Cells labeled with phycoerythrin-conjugated anti-CD4 and allophycocyanin–Cy7–conjugated anti-CD19 were fixed/permeabilized according to manufacturer's instructions (eBioscience) and stained using fluorescein isothiocyanate–conjugated anti–Ki-67 (clone 35; BD).

MPO activity in joints

Potassium phosphate buffer (50 mM; pH 6) containing 1% cetyltrimethylammonium bromide (Sigma) was added (50 μl of buffer/mg of tissue) to joint tissues that had been isolated from coarsely mashed (mortar and pestle) frozen joints. Homogenized tissues were incubated for 30 minutes at 37°C, centrifuged, and the supernatant was collected. MPO activity in supernatants was assessed by measuring the change in absorbance at 460 nm upon oxidation of o-dianisidine dihydrochloride in the presence of H2O2 ([16]).

Cytokines

Splenocytes were cultured with chick collagen (20 μg/ml; 72 hours), and concentrations of interleukin-17A (IL-17A), interferon-γ (IFNγ), and IL-4 in supernatants were measured by enzyme-linked immunosorbent assay (ELISA) ([18]). Levels of TNF, IL-12p70, IL-23p19, and IL-17A in cetyltrimethylammonium bromide–solubilized joint tissue lysates and/or serum were measured by ELISA ([18, 26, 27]). Total protein concentrations in joint tissue lysates were determined with a bicinchoninic acid protein assay kit (Thermo Scientific), and results were expressed in picograms of cytokine per microgram of protein.

Autoantibodies

Serum was collected by cardiac puncture or cheek bleeding, and levels of IgG, IgG1, IgG2a, and IgG3 anticollagen were measured by ELISA ([16, 28]), using mouse type II collagen (Chondrex)–coated plates.

Real-time polymerase chain reaction (PCR) analysis

RNA was extracted from frozen joints ([29]), and real-time PCR for measurement of FoxP3 and β-actin messenger RNA (mRNA) was performed ([30]).

Confocal microscopy

Formalin-fixed, paraffin-embedded (4-μm–thick) sections of synovial biopsy tissue from 6 patients with active RA ([31]) were stained with mouse anti-human CD45 (Dako) and rabbit anti-human MPO (Thermo Fisher), after undergoing heat-induced epitope retrieval (0.1M Tris/0.01M EDTA buffer, pH 9, for 10 minutes under pressure) and blocking (10% chicken serum/5% bovine serum albumin/phosphate buffered saline). Sections were then incubated with Alexa Fluor 488–labeled chicken anti-mouse Ig and Alexa Fluor 594–labeled chicken anti-rabbit Ig and counterstained with DAPI Prolong Gold (all from Molecular Probes).

Images were acquired using a Nikon C1 confocal laser scanner, captured in a line-sequential manner, and converted/analyzed using ImageJ software (National Institutes of Health). To determine the amount of intracellular and extracellular MPO, analysis was performed based on the spatial localization of DAPI and CD45 fluorescence signals, using ImageJ software. MPO+ pixels that had colocalization with DAPI+ and/or CD45+ pixels were considered intracellular, whereas MPO+ pixels that did not colocalize with either CD45+ or DAPI+ pixels were considered extracellular. A perimeter of 3 pixels beyond the physical edge of a CD45+ cell was also defined as part of a CD45+ cell.

In vitro experiments

Fibroblast-like synoviocytes (FLS) from the joints of patients with RA were cultured as previously described ([31]). FLS, used between passages 5 and 8, were cultured for 48 hours in the presence of recombinant human TNF (10 ng/ml; R&D Systems) and with or without purified native human MPO (Sigma). Bromodeoxyuridine (BrdU; Sigma) was added at 20 μM. Proliferation and apoptosis of FLS were measured by flow cytometric analysis of intracellular BrdU incorporation and annexin V/PI staining, respectively ([18]).

Statistical analysis

Results are expressed as the mean ± SEM. Unpaired t-test or Mann-Whitney U test was used for comparison of parametric and nonparametric data, respectively. All statistical analyses were performed using GraphPad Prism software. P values less than 0.05 were considered significant.

RESULTS

Increased MPO activity in joints of mice with passive K/BxN arthritis

To determine whether MPO levels in joints are up-regulated in mice with K/BxN serum–transfer arthritis, we measured MPO activity in joint tissue lysates from normal and arthritic mice. Levels of enzymatically active MPO were markedly increased in the joints of mice with K/BxN arthritis (Figure 1A). No MPO activity was detected in joints from arthritic MPO−/− mice.

Figure 1.

Attenuation of clinical disease by myeloperoxidase (MPO) deficiency in passive K/BxN arthritis. A, MPO activity was measured in lysates of joint tissues from wild-type (WT) normal mice and mice with K/BxN arthritis (day 7 of disease). B, K/BxN arthritis was induced in WT and MPO−/− mice by intraperitoneal injection of K/BxN serum on days 0 and 2, and clinical symptoms of disease were monitored daily until day 7. Clinical scores are shown only for the arthritic mice. C, Paw thickness was measured in WT and MPO−/− mice developing K/BxN serum–transfer arthritis. Data are representative of 2 independent experiments. Values are the mean ± SEM. = P < 0.05; ∗∗ = P < 0.01 versus WT mice.

Attenuation of K/BxN serum–transfer arthritis in MPO−/− mice

To explore the role of endogenous MPO in experimental arthritis, we induced neutrophil-dependent, T cell–independent ([11]) K/BxN serum–transfer arthritis in WT and MPO−/− mice. WT mice developed severe arthritis, as indicated by joint swelling and erythema (Figures 1B and C). The development of arthritis was significantly attenuated by MPO deficiency (Figures 1B and C). Both the clinical arthritis score (Figure 1B) and paw swelling (Figure 1C) were significantly decreased in mice lacking MPO. Histologic assessment of ankle joints from WT mice on day 7 of disease showed typical features of severe arthritis, including marked joint tissue inflammation, synovial hypercellularity, presence of leukocytes in the joint space, as well as cartilage and bone damage (Figures 2A and C). All parameters of joint inflammation (synovitis, joint tissue inflammation, joint space exudate) and damage (cartilage degradation, bone damage) were significantly attenuated in the absence of MPO (Figures 2A and C). Correspondingly, the total arthritis score was significantly reduced in MPO−/− mice (Figure 2B).

Figure 2.

Decreased histologic parameters of disease and local levels of cytokines in K/BxN serum–transfer arthritis caused by a lack of myeloperoxidase (MPO). K/BxN arthritis was induced in wild-type (WT) mice and MPO−/− mice by intraperitoneal injection of K/BxN serum on days 0 and 2. Ankle joints were collected on day 7 and assessed histologically for arthritis severity. A, Joint sections were scored on a scale of 0–3 for each of 5 parameters: joint tissue inflammation, synovitis, joint space exudate, cartilage degradation, and bone damage. B, The total histologic score was calculated from the sum of the 5 parameters in A. C, Representative photomicrographs show arthritis severity in WT and MPO−/− mice with K/BxN arthritis on day 7 of disease. Sagittal sections of formalin-fixed paraffin-embedded ankle joints were stained with Safranin O (fast green/iron hematoxylin counterstain). Original magnification × 50. D, Levels of tumor necrosis factor (TNF) and interleukin-12 (IL-12) in joint tissue lysates or serum collected from WT and MPO−/− mice on day 7 of disease were measured by enzyme-linked immunosorbent assay. Values in A, B, and D are the mean ± SEM. ∗ = P < 0.05 versus WT mice.

Reduction of local, but not systemic, cytokine levels by MPO deficiency in K/BxN arthritis

To assess how MPO affected the local and systemic expression of proinflammatory cytokines in K/BxN arthritis, we measured the concentrations of TNF and IL-12 in joints and serum from WT and MPO−/− mice on day 7 of the disease. TNF and IL-12 were up-regulated in the joint tissues of WT mice with K/BxN arthritis compared with normal mice (mean ± SEM TNF level 831 ± 160 pg/ml versus 2,828 ± 643 pg/ml [P < 0.05] and mean ± SEM IL-12 level 325 ± 5 pg/ml versus 801 ± 220 pg/ml [P < 0.05]). MPO deficiency reduced the joint tissue levels of TNF and IL-12 in K/BxN arthritis (Figure 2D). TNF and IL-12 concentrations were also enhanced in the serum of WT mice with K/BxN arthritis compared with normal animals (mean ± SEM TNF level 23 ± 23 pg/ml versus 1,277 ± 259 pg/ml [P < 0.05] and mean ± SEM IL-12 level 0 ± 0 pg/ml versus 444 ± 136 pg/ml). However, there was no difference in serum TNF and IL-12 concentrations between arthritic WT and MPO−/− mice (Figure 2D).

Attenuation of CIA in MPO−/− mice

To examine the effect of endogenous MPO in another widely accepted model of RA, we induced CIA, which is neutrophil-dependent as well as T cell–, B cell–, and autoantibody-dependent ([6, 7, 10]), in WT and MPO−/− mice. Consistent with the model being initiated in C57BL/6 mice, a strain relatively resistant to CIA, WT mice developed mild arthritis, with clinical signs of the disease beginning around day 27 (Figure 3) and increasing thereafter. CIA was significantly attenuated by MPO deficiency (Figure 3). Both the incidence (Figure 3A) and clinical severity (Figure 3B) of arthritis were decreased in MPO−/− mice compared with WT mice. In addition to significant differences in disease severity at multiple time points, analysis of the area under the curve for disease severity was also significantly reduced in MPO−/− mice compared with WT mice (mean ± SEM 16.2 ± 2.7 versus 6.5 ± 0.7; P < 0.05). In association with reduced disease, levels of proinflammatory cytokines, TNF, IL-17A, and IL-23 tended (P = 0.07) to be decreased in the joints of MPO−/− mice (Figure 3C).

Figure 3.

Suppression of collagen-induced arthritis (CIA) development in myeloperoxidase (MPO)–deficient mice. CIA was induced in wild-type (WT) and MPO−/− mice by immunization with chick type II collagen in Freund's complete adjuvant on days 0 and 21. Disease development was monitored until day 45. A, Incidence of CIA in WT and MPO−/− mice. B, Clinical arthritis scores in WT and MPO−/− mice developing CIA. Clinical scores are shown only for the arthritic mice. C, Levels of tumor necrosis factor (TNF), interleukin-17A (IL-17A), and IL-23 in joint tissue lysates from WT and MPO−/− mice on day 45 of CIA, as determined by enzyme-linked immunosorbent assay. Results are representative of 2 independent experiments. Values in B and C are the mean ± SEM. = P < 0.05 versus MPO−/− mice.

Enhancement of CD4+ T cell responses in secondary lymphoid organs by MPO deficiency in CIA

To assess how the lack of MPO affected CD4+ T cells in secondary lymphoid organs, we examined T cell responses in the LNs and spleen of WT and MPO−/− mice with CIA. T cell activation, as indicated by CD44 and CD69 expression (Figure 4A), was enhanced in the LNs and spleen of MPO−/− mice. Proliferation of T cells, as measured by Ki-67 staining of collagen-stimulated CD4-labeled spleen cells, was significantly augmented by MPO deletion, while a trend toward increased T cell proliferation was observed in the LNs of MPO-deficient mice (Figure 4B). These results were confirmed by increased collagen-specific lymphocyte proliferation in the spleen cells of MPO−/− mice, as measured by 3H-thymidine incorporation (mean ± SEM 2,236 ± 121 counts per minute in WT mice versus 2,690 ± 154 cpm in MPO−/− mice; P < 0.05). T cell apoptosis was decreased in the spleen, while a trend toward increased T cell apoptosis was observed in LNs, of MPO−/− mice (Figure 4C). The proportion of FoxP3+CD25+ CD4 Treg cells in secondary lymphoid organs was decreased by the lack of MPO (Figure 4D). To assess whether increased numbers of Treg cells infiltrated the joints of MPO−/− mice, we measured FoxP3 mRNA in the joints of WT and MPO−/− mice with CIA; however, FoxP3 mRNA could not be detected in joints of any animals (data not shown).

Figure 4.

Enhanced CD4+ T cell responses in secondary lymphoid organs of MPO−/− mice with collagen-induced arthritis (CIA). CIA was induced in wild-type (WT) and MPO−/− mice by immunization with chick type II collagen in Freund's complete adjuvant on days 0 and 21, and lymph nodes (LNs) and spleen were isolated on day 45 for assessment of CD4+ T cell responses by flow cytometry and cytokine production by enzyme-linked immunosorbent assay. A, T cell activation was measured by CD44 and CD69 expression. B, For assessment of T cell proliferation, LN cells and spleen cells were cultured for 72 hours with chick type II collagen, CD4-labeled, fixed, permeabilized, and then stained intracellularly for Ki-67. C, CD4+ T cell apoptosis was measured by annexin V and propidium iodide (PI) staining and analysis by flow cytometry. D, The proportion of regulatory FoxP3+CD25+CD4+ T cells was measured. E, Interleukin-17A (IL-17A) and interferon-γ (IFNγ) concentrations were measured in supernatants collected from splenocytes that had been cultured for 72 hours in the presence of chick type II collagen. Results are representative of 2 independent experiments. Values are the mean ± SEM. ∗ = P < 0.05; ∗∗∗ = P < 0.001 versus WT mice.

In CIA, the absence of MPO increases IL-17A production but decreases IFNγ production by splenocytes

Joint injury in CIA is dependent on IL-17A, while IFNγ is protective in the disease ([32-34]). To assess how these cytokines were affected systemically by MPO deletion, we measured IL-17A and IFNγ levels in supernatants from collagen-stimulated splenocytes derived from WT and MPO−/− mice with CIA. Levels of IL-17A were enhanced, while IFNγ production (Figure 4E) was attenuated by MPO deficiency. This resulted in a markedly increased IL-17A:IFNγ ratio in MPO−/− mice (mean ± SEM 0.02 ± 0.002 in WT mice versus 0.11 ± 0.01 in MPO−/− mice; P < 0.0001). IL-4 production by splenocytes from MPO−/− mice was enhanced (mean ± SEM 0.7 ± 0.3 pg/ml in WT mice and 10.2 ± 2.5 pg/ml in MPO−/− mice; P < 0.001).

The effect of MPO on humoral immunity in CIA

To determine the effect of MPO on humoral immunity, we measured B cell responses in secondary lymphoid organs and collagen-specific autoantibodies in the serum of WT and MPO−/− mice with CIA. B cell activation was increased in LNs, while a trend toward augmented B cell activation was detected in the spleen, of MPO−/− mice (Figure 5A). B cell proliferation, as measured by Ki-67 staining of collagen-stimulated CD19-labeled LN cells or splenocytes, was also enhanced by MPO deletion (Figure 5B). B cell apoptosis was increased in LNs, but decreased in the spleen, of MPO−/− mice (Figure 5C). MPO deficiency did not alter circulating levels of collagen-specific IgG at any time point during CIA development (Figure 5D). Similarly, serum levels of IgG1, IgG2a, and IgG3 anticollagen were not affected by the lack of MPO (Figure 5D).

Figure 5.

Effect of MPO deletion on B cells and serum autoantibodies in collagen-induced arthritis (CIA). CIA was induced in wild-type (WT) and MPO−/− mice by immunization with chick type II collagen in Freund's complete adjuvant on days 0 and 21, and lymph nodes (LNs) and spleens were collected on day 45. Serum was obtained on days 14, 28, and 45 for assessment of anti-mouse type II collagen antibodies (Ab) by enzyme-linked immunosorbent assay. A, B cell activation was measured in LNs and spleen by flow cytometric analysis of CD69 expression. B, For assessment of B cell proliferation, LN cells and spleen cells were cultured for 72 hours with chick type II collagen, CD19-labeled, fixed, permeabilized, and then stained intracellularly for Ki-67 and analyzed by flow cytometry. C, B cell apoptosis was measured by annexin V and propidium iodide (PI) staining and analysis by flow cytometry. D, Circulating IgG anticollagen levels (1:10,000 dilution) were determined at various time points during CIA development, and collagen-specific IgG subtypes IgG1 (1:100 dilution), IgG2a (1:50 dilution), and IgG3 (1:100 dilution) were determined on day 45 of CIA. Results are representative of 2 independent experiments. Values are the mean ± SEM. ∗ = P < 0.05; ∗∗ = P < 0.01; ∗∗∗ = P < 0.001 versus WT mice.

Intracellular and extracellular MPO in RA synovium

To examine MPO expression in RA synovium, we stained synovial biopsy tissues from patients with active RA, using antibodies against MPO and CD45 (pan-leukocyte surface marker) and analyzed them by confocal microscopy. As shown in Figure 6, abundant MPO was demonstrated in the inflamed RA synovium. MPO was detected mainly in the intimal lining (Figures 6A–C), and only occasional MPO+ cells were found within lymphoid aggregates (Figure 6D). The majority of MPO in the synovial lining was detected inside leukocytes, but importantly, MPO (∼10%) was also present in the extracellular space (Figures 6A, B, C, and E).

Figure 6.

Myeloperoxidase (MPO) in rheumatoid arthritis (RA) patients. A–D, Formalin-fixed paraffin-embedded RA synovial sections were stained for CD45 and MPO (DAPI counterstained) and analyzed by confocal microscopy. A, Representative photomicrographs showing abundant MPO in the synovial lining. Original magnification × 20. B, Higher-magnification view of boxed area in A. Original magnification × 600. C, Higher-magnification view of boxed area in B. Arrows indicate extracellular MPO. Original magnification × 2,000. D, Occasional MPO+ cells in lymphoid aggregates. Original magnification × 20. E, Intracellular (IC) and extracellular (EC) MPO in the RA synovial lining. F and G, RA FLS were cultured with TNF and with or without MPO. F, Effect of MPO on fibroblast-like synoviocyte (FLS) proliferation, as determined by the incorporation of bromodeoxyuridine (BrdU) and analysis by flow cytometry. G, Effect of MPO on FLS apoptosis, as determined by annexin V/propidium iodide (PI) staining and analysis by flow cytometry. Values in E–G are the mean ± SEM. ∗ = P < 0.05; ∗∗ = P < 0.01; ∗∗∗ = P < 0.001.

Human MPO increases the expansion of RA FLS

To test how MPO affects FLS function, one of the major effector cell types in the RA synovial lining ([35]), we cultured RA FLS in the presence of TNF and with or without human MPO. MPO significantly increased the proliferation of FLS (Figure 6F) and decreased the apoptosis of FLS (Figure 6G).

DISCUSSION

RA is thought to be mediated by adaptive and innate immune mechanisms. It is widely accepted that CD4+ T cells and B cells are important in the pathogenesis of RA ([3, 4]). More recently, it was shown that neutrophils, the main source of MPO and the dominant inflammatory cell type in arthritic synovial fluid, also play a critical role in models of rheumatoid inflammation ([2, 10, 11]). Extracellularly, MPO contributes to tissue damage in many human chronic inflammatory diseases by forming reactive oxidants ([13]). This potent enzyme is also implicated in RA since it is up-regulated in the synovial fluid of patients with the disease ([19, 20]). Using 2 models of this disease in the current study, we demonstrated that endogenous MPO plays a pathogenic local role in experimental joint inflammation and damage. Our data showing that extracellular MPO is present in the RA synovial lining and that MPO expands FLS suggest that MPO may also play a similar role in human RA. The converse effects of MPO deficiency on adaptive immunity raise the prospect that MPO inhibition in RA might limit inflammatory effector events without constraining adaptive immune responses that are essential for protection from infection and cancer.

To assess the potential role of MPO in RA, we first induced passive K/BxN arthritis, an acute and severe, neutrophil-dependent T cell–independent model of RA ([11]), in WT and MPO−/− mice. Levels of active MPO were elevated in the arthritic joints of WT mice with K/BxN arthritis compared with normal animals, suggesting that MPO may contribute to disease development. Joint swelling, inflammation, and damage were significantly reduced in MPO−/− mice, indicating that MPO plays a proinflammatory injurious role in this model of RA. Decreased levels of TNF and IL-12, proinflammatory cytokines important for the development of K/BxN arthritis ([36, 37]), in the joints but not the serum of MPO-deficient mice support the conclusion that MPO contributes to disease development through local, rather than systemic, effects in the joint.

It was recently demonstrated that neutrophils play a key role as effector cells in joint inflammation and injury in CIA, a well-characterized and widely used CD4+ T cell/B cell–dependent autoimmune model of RA ([10]). Enzymatically active MPO is up-regulated in the joints of mice with CIA ([21]), indicating that MPO may contribute to CIA development. In the current studies, similar to the effects observed in K/BxN arthritis, MPO deficiency protected mice from developing clinical signs of CIA, demonstrating a pathogenic role of endogenous MPO in CIA development. We note that the severity of CIA in C57BL/6 mice was mild, as has been reported by multiple other investigators; this could conceivably result in a protective effect of MPO deletion that may not be translated to more severe disease. However, the lack of MPO was also protective in the severe K/BxN serum–transfer model, as discussed above. The effects of MPO on human RA synovial cells, as discussed below, also support an important potential contribution of MPO in human RA.

CD4+ T cells and B cells are known to play key roles in CIA ([3, 4]). Experiments in models of multiple sclerosis and glomerulonephritis have shown that MPO can suppress the generation of T cell and B cell responses in secondary lymphoid organs ([16, 17]), most likely because neutrophils infiltrate LNs and spleen, where, through interactions with antigen-presenting cells and/or T cells, they affect adaptive immunity ([38, 39]). Consistent with the findings in other disease models ([16, 17]), we found that CD4+ T cell responses were enhanced in the LNs and spleen of MPO−/− mice with CIA. Correspondingly, the proportion of Treg cells, which are protective in CIA ([40]), was decreased by the lack of MPO. Furthermore, in association with the augmented T cell responses and consistent with previous reports ([16, 17]), B cell activation and proliferation in LNs and spleen were enhanced by MPO deficiency. Circulating levels of collagen-specific autoantibodies were, however, not different between WT and MPO−/− mice. This may be due to opposing effects of MPO deficiency on B cell apoptosis in LNs and spleen, possibly arising because of different levels of MPO in these organs ([16, 18]).

Collectively, these results demonstrate that decreased CIA severity in MPO−/− mice cannot be attributed to diminished generation of injurious CD4+ T cell or B cell responses in secondary lymphoid organs or to decreased levels of pathogenic autoantibodies. In fact, decreased CIA severity was observed in MPO−/− mice despite enhanced generation of adaptive immune responses, strongly suggesting that the local antiinflammatory effects of MPO deficiency in the joints override the amplifying effects of systemic MPO deletion on adaptive immunity. In antigen-induced arthritis, enhanced T cell responses in the spleen due to MPO deficiency are associated with exacerbated disease in the joints ([18]). This is most likely due to the possibility that this nonautoimmune, T cell–driven model of RA ([41]) is not as neutrophil (and thus MPO)–dependent in the effector phase as CIA and K/BxN arthritis.

In CIA, IL-17A is pathogenic, while IFNγ is generally inhibitory ([32-34]). Attenuated CIA due to MPO deficiency cannot be explained by changes in systemic IL-17A or IFNγ levels. MPO−/− mice had increased IL-17A and decreased IFNγ production by splenocytes, resulting in an enhanced IL-17A:IFNγ ratio, which is known to cause CIA exacerbation ([33]). Increased IL-4 in MPO−/− mice is not a likely explanation for the decreased disease in these animals, as endogenous IL-4 does not affect CIA ([33, 42]). These results, together with the reduced levels of proinflammatory cytokines, including IL-17A, in the joints of MPO−/− mice further support our conclusion that MPO acts as a local, and not systemic, mediator of joint inflammation in CIA.

The present studies are the first to show that abundant MPO is present in the inflamed synovium of patients with active RA. The majority of the MPO was detected in the intimal lining, suggesting that it is derived from synovial lining macrophages, since neutrophils are known to be rare in the synovium ([35]). Other studies have shown increased MPO expression in RA synovial fluid, where neutrophils, due to their abundance, are thought to be the major source of this enzyme ([19, 20]). Importantly, here, MPO was detected not only within, but also outside of, leukocytes, suggesting that MPO that is released into the extracellular milieu of the synovium may contribute to joint inflammation and damage in RA by affecting the local cells. Our in vitro studies indicate that one of the mechanisms by which MPO may locally augment joint injury is by increasing the expansion (via enhanced proliferation and diminished apoptosis) of FLS, synovial lining cells which are well-known for their RA-promoting effects ([35]). Previous studies have reported similar antiapoptotic effects of MPO on other cell types ([43]). MPO may also, in an autocrine/paracrine manner, activate local macrophages, which contribute to joint inflammation and damage in RA ([1, 44]), by increasing their respiratory burst and proinflammatory cytokine production ([45, 46]).

In conclusion, using 2 models of RA, passive K/BxN arthritis and CIA, the current study demonstrates that endogenous MPO plays an important local pathogenic role in inflammatory arthritis. In CIA, these injurious effects of MPO are sufficient to override its suppressive effects on the generation of arthritogenic adaptive immunity in secondary lymphoid organs. These results from animal studies are supported by results in humans showing the presence of extracellular MPO in the inflamed RA synovium and its effects on synovial fibroblasts. Collectively, our findings suggest MPO as a potential therapeutic target in the treatment of RA. The observation that the absence of MPO suppresses joint inflammation and damage without impairing systemic adaptive immunity raises the possibility that MPO-targeting therapies could reduce local joint inflammation without inhibiting adaptive immune responses that are required to fight infection and malignancy.

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. Odobasic 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. Odobasic, Yang, Morand, Holdsworth.

Acquisition of data. Odobasic, Yang, Muljadi, O'Sullivan, Kao, Smith.

Analysis and interpretation of data. Odobasic, Yang, Muljadi, O'Sullivan, Kao, Morand, Holdsworth.

Acknowledgments

The authors thank Prof. M. Leech for help with analyzing stained synovial sections from RA patients.

Ancillary