| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
From the Department of Rheumatology,* University Medical Center Nijmegen, Nijmegen, The Netherlands; N.V. Organon, Oss, The Netherlands; the Institute of Pharmacology, School of Medicine, University of Messina, Messina, Italy; and the Laboratory of Host Defenses, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
, interleukin-1, matrix metalloproteinase (MMP)-3, MMP-9 and receptor activator of NF-
B ligand (RANKL) might contribute to the aggravated arthritis in the NADPH oxidase-deficient mice. This showed that the involvement of NADPH oxidase in arthritis is probably far more complex and that oxygen radicals might also be important in controlling disease severity, and reducing joint inflammation and connective tissue damage.
Autoimmune disorders such as lupus and polyarthritis in CGD patients9
prompted us to study the role of NADPH oxidase in experimental arthritis. In arthritis patients, increased reactive oxygen species are present at the sites of inflammation.10
Circulating neutrophils and monocytes have increased NADPH oxidase activity in rheumatoid arthritis (RA) patients11,12
and the superoxide release by the neutrophils isolated from synovial fluid of these patients is even further increased.13
NADPH oxidase-derived superoxide, a highly reactive molecule, may have effector functions in RA, which are detrimental for the joint. One approach to elucidating the role of superoxide in experimental arthritis is to treat animals with superoxide dismutase (SOD). SOD has a beneficial effect on arthritis that is induced by streptococcal cell walls, adjuvant, or via immunization with collagen type II.14-16
However, we previously found that local treatment with cationic SOD has no effect on antigen- and zymosan-induced arthritis in mice.17
Recently, Lida and Saito18
found that endotoxin-free CuZn SOD has no effect in three arthritis models. Even trace amounts of endotoxin contamination of SOD do carry the anti-inflammatory effects found in previous studies. Therefore, to identify the role of superoxide derived from NADPH oxidase in arthritis, we used gene knockout (KO) models of CGD. Zymosan was used as an irritant to elicit arthritis since it is a potent activator of p47phox-mediated superoxide production in polymorphonuclear leukocytes and macrophages via activation of the toll-like receptor (TLR)-2, or through nonopsonic phagocytosis using the complement receptor type 3.1,19-21
The findings of ZIA are confirmed in the passive immune-complex arthritis (ICA) model in mice.22
Immune complexes stimulate neutrophil NADPH oxidase-mediated superoxide production via their Fc-
receptors, a pathway distinct from TLR-2 signaling.23
Experimental arthritis was compared between wild-type (WT) and p47phox gene KO mice. We found worsening of joint inflammation and granulomatous synovitis with extensive cartilage and bone erosion in the p47phox gene KO mice. This identifies NADPH oxidase-dependent superoxide as a possible negative feedback/anti-inflammatory mediator in arthritis, which is contrary to previous reports of the anti-rheumatic effects of SOD. Furthermore, this study unexpectedly showed that connective tissue destruction during arthritis can proceed and is even enhanced in the absence of NADPH oxidase-produced superoxide.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
P47phox-/- and gp91phox-/- mice, generated as previously described,7,24 lack the cytosolic p47phox and membrane gp91phox subunits of the NADPH oxidase multicomponent system, respectively. The KO mice were backcrossed to the C57Bl6 background for 15 generations and C57Bl/6 (obtained from The Jackson Laboratory, Bar Harbor, ME) were used as controls. In some experiments p47phox-/- mice of intercross progeny (C57Bl/6x129Sv) were used with their proper controls. Colonies were maintained at the National Institutes of Health (Bethesda, MD). All mice were housed under specific pathogen-free conditions during breeding and experiments. Mice received autoclaved chow and acidified water at libitum. Only mice that were healthy were used in the experiments and were age-matched (10 to 20 weeks) and sex-matched for each set of experiments. To limit discomfort for the animals, experiments were terminated at day 7 of arthritis and mice were killed by cervical dislocation. All experiments were approved by local authorities of Animal Care and Use Committee (DEC 98.22) and performed by personnel certified by the Dutch Ministry of Health, Welfare, and Sport.
Zymosan-Induced Arthritis (ZIA)
A homogeneous suspension of 300 mg of zymosan A (from Saccharomyces cerevisiae; Sigma, St. Louis, MO), dissolved in 10 ml of endotoxin-free saline was obtained by repeated boiling, followed by sonic emulsification. The suspension was autoclaved for and stored in 0.5 ml aliquots at -20°C. Arthritis was induced by intraarticular injection of 180 µg of zymosan through the suprapatellar ligament into the joint cavity. In specified experiments the contralateral knee-joints received an equal amount of saline (6 µl) as control.
Immune-Complex Arthritis
Immune complex arthritis was induced as described before.22
In brief, arthritis was induced by injecting 3 µg of poly-L-lysine coupled lysozyme into the right knee-joint cavity of mice that had, the night before, received intravenously 0.2 ml specific rabbit antisera directed against lysozyme. To prevent infection both antigen and antisera were -radiated with 25kGy before use.
Depletion of Mice for Leukotrienes and Complement
To inhibit 5'-lipoxygenase, mice were treated with Zileuton (Abbott, Abbott Park, IL). A 600-mg tablet was emulsified in 100 ml of 0.5% hydroxyethyl cellulose in saline and stored at 4°C. Mice received by gavage an oral dose of 50 mg/kg daily, starting 1 week before induction of arthritis until the end of the experiment. To deplete animals of complement, cobra venom factor (CVF) (200U/kg; Quidel Corp., San Diego, CA) was injected i.p. the night before arthritis induction. The dose has previously been shown to reduce C3 levels in plasma to 3% of normal values within 4 hours after CVF administration and complement C3 values returned to normal 9 days later.25
Assessment of Joint Swelling
Animals were injected subcutaneously with 14 µCi of 99mTechnetium pertechnetate in 0.2 ml of saline in the neck region. After 10 to 15 minutes the mice were sedated by 4.5% chloral hydrate injected i.p., 0.01 ml/g of body weight. The accumulation of the isotope due to increased bloodflow and edema in the knee was determined by external gamma counting using a fixed position of the Na-I-scintillation crystal. A ratio of the 99mTc uptake in the inflamed over contralateral knee-joint of >1.1 indicated joint swelling.
Histology
Knee joints were dissected, fixed in phosphate-buffered formalin (pH 7.4) for 4 days, decalcified in 5% formic acid, and embedded in paraffin wax. Serial coronal (frontal) sections of whole knee-joint were made and stained with safranin-O and counterstained with fast green. Glycosaminoglycan depletion was scored using a linear scale from 0 (normal red staining) to 3 (complete loss of red staining) in the metabolically active upper-layer of articular cartilage. Inflammation was evaluated according to the extent and composition of the influx of inflammatory cells into the synovium (infiltrate) and joint cavity (exudate) and scored separately using a linear scale from 0 (no inflammation) to 3 (full-blown inflammation). Seven-eight serial sections taken every 70-µm apart were scored per joint.
NIMP-R14 Staining
Influx of polymorphonuclear neutrophils was analyzed on knee-joint sections. Briefly, sections were deparaffinized and preincubated for 15 minutes at room temperature with 20% normal rabbit serum. Thereafter, sections were incubated with rat monoclonal antibodies to mouse neutrophils (NIMP-R14, a 25- to 30-kd epitope mainly present on neutrophils26 ) for 1 hour. After incubation with a second peroxidase conjugated rabbit anti-rat IgG antibodies for 30 minutes, sections were incubated with 3-amino-9-Ethylcarbazole (Vector Laboratories, Burlingame, CA) substrate in the dark at 37°C for 10 minutes. Thereafter, sections were stained with hematoxylin for 30 seconds. As a negative control, sections were incubated with normal rat Ig instead of NIMP-R14.
Tartrate-Resistant Acid Phosphatase (TRAP) Staining
Whole knee-joints were fixed for 2 days in 2% formalin, followed by decalcification in 10% ethylenediaminetetraacetate (EDTA) in 1 mmol/L Tris-HCl pH 7.4 for up to 2 weeks at 4°C. Decalcified specimens were processed for paraffin embedding. Staining of tissue sections (7 mm) for TRAP was performed by a leukocyte acid phosphatase kit, a cell-staining kit for the detection of TRAP (Sigma).
Immunohistochemistry of VDIPEN
Immunostaining of VDIPEN, a neo-epitope of a matrix metalloproteinase (MMP)-induced cleavage site in aggrecan, was performed as described.27 In short, dissected knee-joints were fixed in 4% paraformaldehyde and subsequently decalcified in 10% EDTA in phosphate buffer (pH 7.4) for 7 days. Thereafter, joints were embedded in paraffin wax and coronal sections (7 µm) were cut, dewaxed, hydrated and digested with proteinase-free chrondroitinase ABC (0.25U/ml in 0.1 mol/L Tris-HCL, pH 8.0; Sigma) for 1 hour at 37°C. Sections were treated with Nakane’s fixative, exposed to 1% hydrogen peroxide in methanol for 20 minutes, then 5 minutes with 0.1% TritonX-100 in PBS. After blocking with 1.5% normal goat serum for 20 minutes, sections were incubated with affinity-purified rabbit anti-VDIPEN IgG (a gift from Merck Research Laboratories, Rahway, NJ) overnight at 4°C. This antibody has been extensively characterized and detects the VDIPEN C-terminal sequence in the core protein of aggrecan after cleavage by MMPs and cathepsin. Sections were then incubated with biotinylated goat anti-rabbit IgG, and the VDIPEN immunostaining was detected by avidin-peroxidase staining using nickel enhancement development. Orange G was used as a counterstain.
Immunohistochemical Localization of Poly(ADP-Ribose) Polymerase (PARP)
Paraffin-embedded tissue sections were deparaffinized, and endogenous peroxidase was quenched with 3% H2O2 in 60% methanol for 30 minutes. The sections were permeabilized with 0.1% Triton X-100 in phosphate-buffered saline for 20 minutes. Non-specific adsorption was minimized by incubating the section in 2% normal goat serum in phosphate-buffered saline for 20 minutes. Endogenous biotin or avidin binding sites were blocked by sequential incubation for 15 minutes with avidin and biotin (DBA, Milan, Italy). The sections were then incubated overnight with 1:500 dilution of primary anti-PARP antibody (DBA) or with control solutions. Control included buffer alone or non-specific purified rabbit IgG. Specific labeling was detected with a biotin-conjugated goat anti-rabbit IgG and avidin-biotin peroxidase (DBA).
Radiology
At the end of the experiments, knee joints were isolated and used for X-ray analysis as a marker for bone destruction. X-ray photographs (Kodak Industrex film) were carefully examined using a stereo microscope. Bone erosions were scored on a scale ranging from 0 to 5: 0, no erosions; 1, minor bone destruction, 1 spot; 2, moderate changes, 2 to 4 spots, 1 area; 3, marked changes, 2 to 4 spots, more areas; 4, severe erosions afflicting the joint; 5, complete destruction of joint, new bone formations. Osteolysis was also scored as loss of trabecular bone organization and bone density using a scale from 0 to 3, ranging from normal cortical and trabecular bone structure to complete loss of organization, cracks and occasional fractures.
RNA Isolation and Reverse Transcriptase-Polymerase Chain Reaction
RNA extraction and RT-PCR were performed as described.28
Synovial biopsies were taken from tissue adjacent to the suprapatelar ligament with a biopsy punch (diameter of 3 mm) and immediately frozen in liquid nitrogen. RNA was isolated with TRIzol reagent according to the protocol of the manufacturer (Life Technologies, Breda, The Netherlands), and reverse-transcribed to complementary DNA (cDNA) using oligo(dT) primers and reverse transcriptase (Life Technologies). Equal amounts of cDNA, normalized by the housekeeping gene GAPDH, were used for standard PCR reactions using Taq-DNA polymerase (Life Technologies). Primers for murine MCP1, macrophage inflammatory protein (MIP)1, MIP2, interferon (IFN)-, tumor necrosis factor (TNF)-, interleukin (IL)-1, IL-6, IL-10, IL-1Ra, MMP2, MMP3, MMP9, tissue inhibitor of metalloproteinase (TIMP)-1, osteoprotegerin ligand (OPGL), and Osteoprotegerin (OPG) were designed by Primer3 (www-genome.wi.mit.edu/genome software/other/primer3.html). Specificity was controlled using FASTA search databank.29
Samples of the PCR reaction (5-µl aliquots) were taken after an initial round of 20 cycles, and thereafter every 2 to 3 cycles until the end of the PCR reaction (cycle 35 to 40). The size of the PCR product was analyzed on an 1.6% agarose gel containing ethidium bromide. Photoscanning analysis of the agarose gels was performed to determine the difference in the number of cycles between the expression in normal synovium (contralateral joint) with the inflamed synovium (zymosan-injected joint) of the same animal. Synovial tissue samples of six mice were evaluated individually and the mean difference in gene expression was presented.
Expression of apoptosis genes (BAX, Bcl-2, Bcl-xL, Bcl-xS, caspase-3) was evaluated by multiplex PCR using the mouse apoptosis genes set 2 (Apo-M152; AMS Biotechnology, Abingdon, Oxon, UK) according to the manufacturer’s protocol. Products were run on 2% agarose gel and analyzed with the Multi-Analyst system (Bio-Rad, Hercules, CA).
Patella Assay
Patellae were dissected with a minimal amount of surrounding soft tissue and incubated in RPMI 1640 (Glutamax; Invitrogen, Carlsbad, CA) with 1% pyruvate (100 mmol/L) and Penstrep 100 IU/ml with or without murine recombinant IL-1 (10 ng/ml) for 48 hours. Thereafter glycosaminoglycan (GAG) synthesis was assessed by incorporation of 35S-sulfate (20 µCi/ml) during the last 3 hours. Patellae were fixed overnight (o/n) in 10% formalin and subsequently decalcified in 5% formic acid for an additional 4 hours. Thereafter, patellae were punched out of their surrounding tissue and dissolved in 0.5 ml Lumasolve (Lumac, Groningen, The Netherlands) o/n at 60°C. The 35S-sulfate content of each patella was measured by liquid scintillation counting and expressed as counts per minute (cpm). More than 95% of the radiolabel was incorporated into the GAGs of proteoglycans. The concentration of NO2- (a stable breakdown product of NO) in the 24-hour culture medium was determined by the Griess reaction, using NaNO2 as a standard. One hundred µl of culture medium was mixed with 100 µl of Griess reagent (0.05% naphtylene diaminedihydrochloride and 0.05% sulfanilamide in 2.5% H3PO4) in a flat-bottom microtiterplate and adsorbance was read at 545 nm using an ELISA-plate reader.
Statistical Analysis
Statistical comparison between groups was performed with Student’s t-test. Histological scores were analyzed using Mann-Whitney Rank Sum Test. Values of P < 0.05 were considered significant.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
An intraarticular injection of sterile zymosan particles causes an acute joint inflammation and compared to WT mice, joint swelling was markedly increased (fivefold) in p47phox-/- mice (Figure 1)
. Histological evaluation of the knees at day 7 of ZIA showed exudation of polymorphonuclear leukocytes (PMN) into the joint cavity residing in close proximity of the cartilage surfaces (Figure 2, E and F)
. Furthermore, zymosan-induced joint inflammation developed into a granulomatous synovitis in both p47phox-/- and gp91-/- mice (Table 1
; Figure 2
). Immunohistological staining of NIMPR4 confirmed that a large percentage (approximately 40%) of infiltrating cells were PMNs (Figure 2, C and D)
. Interestingly, the caseous necrotic centers of granulomata also stained positive for NIMP-R14, although cells could not be identified morphologically. A mononuclear cell layer (predominantly macrophages) which stained negative for NIMP-R14 surrounded the granulomas. The relative paucity of lymphocytes could suggest that the inciting zymosan was either an inert foreign body or that at day 7 an acquired immune response against zymosan had not yet developed. No granulomatous synovitis was seen in immune-complex mediated arthritis (ICA), although NADPH oxidase-deficient mice also showed a significant increased joint inflammation as compared to the WT littermates (Table 1)
.
|
|
|
The leukotriene LTB4 and complement C3 are potent chemokines but are sensitive for NADPH oxidase-mediated oxidative degradation. Impaired inactivation of these chemokines might cause the enhanced inflammatory response in NADPH oxidase-deficient mice. Treatment with Zileuton, an inhibitor of 5'-lipoxygenase had no effect on the extent of zymosan-induced joint inflammation in either WT or p47phox-/- mice (Table 2)
. Treatment of mice with CVF significantly suppressed infiltration (-64%) and exudation (-83%) in WT animals. The same CVF treatment, however, did not decrease joint inflammation in the p47phox-/- mice (Table 2)
. This suggests that an impaired oxidative degradation of the chemokines LTB4 and complement C3 was not the underlying mechanism of the derailed inflammatory response during experimental arthritis in p47phox-/- animals.
|
Whole knee sections from p47phox-/- mice showed enhanced loss of glycosaminoglycan (safranin-O staining) content in the articular cartilage matrix of patella and tibial plateau at day 7 of arthritis in both models (Figure 2, G and H
; Table 1
). The differences in GAG content of the patellar groove (femur) were statistically significant between WT and NADPH oxidase-deficient mice only in the ICA model (Table 1)
. We previously showed that IL-1 played a major role in the suppression of chondrocyte proteoglycan synthesis via a nitric oxide-dependent mechanism during experimental arthritis. We tested whether the articular chondrocytes with NADPH oxidase deficiency became more vulnerable to IL-1-induced inhibition of proteoglycan synthesis. Patellar cartilage was exposed to murine recombinant IL-1 (10 ng/ml) for 48 hours. We found that p47phox deficiency did not affect IL-1-induced inhibition of chondrocyte proteoglycan synthesis and did not significantly increase the level of nitrite (stable end-product of nitric oxide in aqueous solutions) in culture media (Table 3)
. More importantly, we observed enhanced expression of the neo-epitope VDIPEN, a MMP cleavage site in the core protein of proteoglycans, and increased chondrocyte death in the non-calcified cartilage layer (above tidemark) suggests that NADPH oxidase deficiency leads to irreversible cartilage damage during ZIA (Table 1
; Figures 1 and 3
).
|
Histology showed pronounced erosions of the cortical bone in knee joints from gp91-/- and p47phox-/- mice at day 7 of ZIA (Figure 2B
; Table 1
). X-ray analysis showed significant changes of 2 to 4 erosion spots per whole knee-joint section in 9 out of 15 p47phox-/- mice evaluated (mean 2.78 ± 1.48) and all 15 WT animals examined were without erosions (Figure 4)
. Furthermore, marked osteolysis (loss of trabecular bone) below the femoral growth plate in the femur was found in all p47phox-/- mice at day 7 of arthritis (mean 2.1 ± 0.7) (Figures 4 and 5)
. The overall integrity of the bone structure was not compromised in the arthritic joints of WT mice, indicating a clear difference with p47phox-/- mice. Histochemistry showed TRAP-positive osteoclast-like cells at the sites of bone erosion (Figure 6)
in the p47phox-/- mice. No osteoclast-like cells were present in the inflamed synovia and at the bone margins in the periosteum of the arthritic joints in WT mice. Only moderate bone erosion was observed in the arthritis joints of gp91-/- mice with ICA (Table 1)
.
|
|
Synovial tissues were dissected 24 hours after zymosan injection and mRNA expression was analyzed by RT-PCR. Compared to the contralateral non-inflamed joints, the arthritic joints showed enhanced mRNA expression of chemokines (MCP1, MIP-1
), cytokines (IFN-, TNF, IL-1) and metalloproteinases (MMP-2, -3, -9) (Figure 7)
. The natural protective proteins IL-1Ra and TIMP were also up-regulated during inflammation. In the p47phox-/- mice, the expression of most of the above-mentioned genes was above the expression in WT mice (Figure 7)
. The shift toward a more severe process in NADPH oxidase-dependent superoxide deficiency was underlined by changes in IL-1/IL-1Ra mRNA expression ratio (0.6 in WT to 1.5 in p47phox-/- and the MMP3/TIMP or MMP9/TIMP ratios (1.4 in WT to 2.2 in p47phox-/- mice. Substrate polyacrylamide gel electrophoresis of synovial tissue extract detected enzymes with gelatinolytic activity in arthritic joints of p47phox-/- but not WT mice (not shown). The genes for MIP2 and RANKL (also known as OPGL) were expressed exclusively in the arthritic joints of p47phox-/- mice. The antagonist OPG was not detected in the synovial mRNA samples. The above data underlined the misbalance of cytokines, enzymes, growth factors, and their antagonists in NADPH oxidase-deficient mice that could explain the exaggerated inflammatory response and marked connective tissue destruction during arthritis in these mice.
|
. Immunodetection of poly (ADP-ribose) polymerase (PARP), another marker of apoptosis, showed comparable staining intensity in the inflamed synovial tissue between both mouse strains (Figure 9)
. This suggests that the enhanced inflammation in the NADPH oxidase-deficient mice was probably not due to diminished apoptosis.
|
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
These results in NADPH oxidase-deficient mice are distinct from those obtained by experiments using reactive oxygen species (ROS) inhibitors, indicating that the use of various KO mouse models will help in delineating the role of NADPH oxidase-derived oxygen radicals.14-16 It is possible that the acute loss of NADPH oxidase-derived superoxide production was compensated by the action of other mediators. For instance, in the absence of gp91phox, NO may provide some resistance to indigenous bacteria.30 The latter could be of importance, as the combined inhibition of superoxide and NO was by far more effective in the treatment of experimental arthritis.31 We did not study possible compensatory mechanisms in detail, but NO production as induced by IL-1 in synovial explants was not further increased with p47phox deficiency. We previously showed that NO derived from inducible NOS does not significantly contribute to zymosan-induced joint inflammation.32 It is therefore unlikely that unchecked NO in the absence of superoxide in p47phox-/- mice was responsible for worsening of arthritis.
The superoxide scavengers used in previous studies failed to discriminate between the recently discovered new members of the NADPH oxidase family like the gp91phox homologues called Nox/Duox present in non-neutrophil cells. Although the rate of oxidant production is quite low by these other cell types, evidence is emerging that the Nox/Duox-derived oxidants serves a cell signaling function, eg, angiogenesis33,34 but others are speculated.3
The exacerbated joint inflammation in the NADPH oxidase-deficient mice could be caused by impaired phagocytosis of the immune complexes and zymosan used to elicited arthritis. Neutrophils from CGD patients are impaired in the phagocytosis of immune-complexes. The formation of granulomatous synovitis also suggests that zymosan in the NADPH oxidase-deficient mice was not cleared properly from the joint due to impaired phagocytosis. Evidence to support this is the exuberant splenic granulomata formed in p47phox-/- mice after administration of Schistosoma mansoni eggs.35
Excessive inflammatory response to sterile, heat-killed fungal particles in the absence of a respiratory burst is reported in the lung. In X-linked (gp91phox-/-) CGD mice, killed Aspergillus fumigatus injected into the lung causes significantly more alveolar neutrophils and increases expression of IL-1ß and TNF.36
Enhanced local levels of MIP2 and keratinocyte-derived chemokine also accompanied neutrophil sequestration in gram-negative lung inflammation in gp91phox-/- and p47phox-/- mice.37
We found that the exacerbated zymosan-induced inflammation in p47Phox-/- mice was accompanied with increased gene expression of TNF, IL-1, MIP1, and MIP2. Although ROS are able to provoke cell signaling by the activation of NF-B, Erk, and the JAK-STAT pathway,38,39
our study suggests that this is probably not the major route of cell activation during arthritis.
In vitro, reactive oxidants can inactivate proinflammatory chemotactic factors, including leukotrienes and C5a.40,41 We previously showed that in p47phox-/- mice, the enhanced thioglycollate-induced peritonitis was in part due to reduced oxidant-induced LTB4 inactivation.42 Pretreatment of the p47phox-/- mice with zileuton to prevent LTB4 generation did reduce the thioglycollate-induced peritonitis but had no effect on zymosan-induced joint inflammation. CVF is known to bind factor B of the alternative complement pathway, forming a stable C3bBb complex and a markedly depleted C3, preventing the chemoattractive activity of C3b and C5a. Zymosan activates complement via the alternative pathway and in this study we showed that CVF pretreatment of mice significantly reduced zymosan-induced joint inflammation in WT mice. However, CVF treatment of the p47phox-/- mice failed to suppress the exacerbated zymosan–induced joint inflammation, showing complement independence. This showed that the enhanced joint inflammation in NADPH oxidase-deficient mice was not due to an impaired inactivation of the chemotactic factors LTB4, C3b, and C5a at the site of inflammation by lack of reactive oxidant inactivation.
Diminished ROS-mediated apoptosis of inflammatory neutrophils may also account for the derailment of the inflammatory reaction in the arthritic NADPH oxidase-deficient mice.43 Reactive oxygen intermediates, especially in combination with NO, can damage lipids, proteins, and nucleic acids causing cells to die by apoptosis. Our finding that NADPH oxidase-deficient mice developed gaseous necrotic centers of granuloma and that reduced expression of caspase-3 and Bcl-xS suggests that there is reduced apoptosis in the arthritic tissues of gp91phox and p47phox deficient mice. The unchanged expression of PARP, an enzyme involved in DNA repair,44 does not support this conclusion.
Frustrated phagocytosis of neutrophils due to large foreign bodies or insoluble immune-complexes trapped on cartilage leads to an increased respiratory burst and degranulation.45
Phagocytosing neutrophils from CGD patients releases 2 to 3 times more active lysozyme and ß-glucuronidase.46
It is also shown that granulomatous tissue exerts a strong cartilage destructive effect. Implantation of cotton-wrapped cartilage into a murine air-pouch results in granuloma formation and increased gelatinase activity, and accelerates loss of proteoglycans and collagen from cartilage.47
We also found increased MMP expression and gelatinolitic activity in the synovial tissue of p47phox-/- mice. The levels of neutrophil-derived proteinases probably exceeds the threshold raised by antiproteinases present in the synovial fluid. Furthermore, the neutrophils in the arthritic joints of p47phox-/- mice were in close proximity to cartilage (Figure 2F)
, increasing the chance that neutrophil products (eg, secreted (metallo)proteinases) can escape scavenging and inhibition.45
In the p47phox-/- mice, this resulted in the generation of the VDIPEN neo-epitope, a MMP cleavage site in aggrecan. Previously, we reported that VDIPEN expression only occurred in experimental arthritis models in which immune complexes were involved (eg, antigen-, immune-complex-, and collagen-induced arthritis) but not in the non-immunologically mediated zymosan-induced arthritis.27,48,49
A possible explanation is that immune-complex deposition on the cartilage surface brings neutrophils in close contact with cartilage and directs destruction there. Furthermore, our study supports previous reports on oxygen-independent cartilage destruction evoked by neutrophils or conditioned synovial media in vitro.50,51
We previously showed that endogenous IL-1 suppresses chondrocyte GAG synthesis and this contributes significantly to articular cartilage damage seen during experimental arthritis.52
NO-mediated the suppression of GAG synthesis by during arthritis.53
This study showed that NADPH oxidase-dependent superoxide played no part in the suppression of GAG synthesis. Although we only studied arthritis until day 7, it is highly unlikely that the cartilage would recover in the p47phox-/- mice at later stages, as there was extensive chondrocytes death (Table 1)
. The study of Blanco et al54
shows a balance in NO/ROS for maintaining cell viability and that inhibition of IL-1-induced superoxide causes chondrocyte apoptosis, and inhibition of NO leads to necrosis.
The bone resorbing activity of the specialized, highly differentiated osteoclasts mediates bone erosion. We found TRAP-positive multinucleated cells, characteristic of osteoclast-like cells, at sites of trabecular (osteolysis) and cortical bone erosion in NADPH oxidase-deficient mice. Furthermore, we found increased osteoprotegerin ligand (OPGL/RANKL) expression in the inflamed synovium of p47phox-/- mice, a key mediator in osteoclast differentiation.55,56 The NADPH oxidase complex, including the p47phox protein, is present in the ruffled border of osteoclasts actively resorbing bone.57 Superoxide does not stimulate bone resorption directly but leads to hydrogen peroxide production, which in turn stimulates osteoclast formation and enhances bone-resorbing activity.58 However, osteoclast from gp91 KO mice produces equal amounts of superoxide and these mice are not osteopetrotic, indicating normal bone turnover. Yang et al59 showed that NOX4, a member of the newly discovered Nox/Duox family of NADPH oxidases in the osteoclasts generates superoxide and mediates bone resorption. Another contributor to the osteoclast-generated superoxide is TRAP. TRAP is a binuclear metalloenzyme that can react with hydrogen peroxide to generate ROS, which can destroy collagen and other proteins.60 Physiologically, TRAP is required for normal bone matrix resorption, as mice lacking TRAP have disrupted endochondral ossification and develop mild osteopetrosis.61 In this study we found that TRAP-positive cells are present in the areas of bone resorption, and bone resorption occurred in the absence of functional NADPH oxidase. Our study showed that for full osteoclast activity leading to bone resorption, NADPH oxidase is not required.
It is clear from our study that in non-infectious joint inflammation elicited by an irritant (zymosan) or immune complexes, the NADPH oxidase-derived oxidants might well serve to self-limit the disease process. Future research to define the mechanism of the control of inflammation by this crucial inflammatory mediator should indicate novel pathways for therapeutic exploration.
|
|
|
Footnotes |
|---|
Supported by the Dutch Arthritis Association (01–1-304).
Accepted for publication June 27, 2003.
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
in patients with inflammatory and autoimmune rheumatic diseases. Inflammation 1996, 20:427-438[Medline]
B activation, apoptosis and reactive oxygen species production. J Endotoxin Res 2001, 7:287-291[Medline]
RI to nicotinamide adenine dinucleotide phosphate (reduced form) oxidative burst and immune complex trafficking requires the activation of phospholipase D1. Blood 2001, 98:3421-3428-1-proteinase inhibitor. Br J Exp Pathol 1987, 68:81-88[Medline]
, and interleukin-6 in cartilage proteoglycan metabolism and destruction. Effect of in situ blocking in murine antigen- and zymosan-induced arthritis. Arthritis Rheum 1995, 38:164-172[Medline]
This article has been cited by other articles:
 |
|
 |
|
  R. Rajakariar, T. Lawrence, J. Bystrom, M. Hilliard, P. Colville-Nash, G. Bellingan, D. Fitzgerald, M. M. Yaqoob, and D. W. Gilroy Novel biphasic role for lymphocytes revealed during resolving inflammation Blood, April 15, 2008; 111(8): 4184 - 4192. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. H. Segal, B. A. Davidson, A. D. Hutson, T. A. Russo, B. A. Holm, B. Mullan, M. Habitzruther, S. M. Holland, and P. R. Knight III Acid aspiration-induced lung inflammation and injury are exacerbated in NADPH oxidase-deficient mice Am J Physiol Lung Cell Mol Physiol, March 1, 2007; 292(3): L760 - L768. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Bedard and K.-H. Krause The NOX Family of ROS-Generating NADPH Oxidases: Physiology and Pathophysiology Physiol Rev, January 1, 2007; 87(1): 245 - 313. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. J. Snelgrove, L. Edwards, A. E. Williams, A. J. Rae, and T. Hussell In the Absence of Reactive Oxygen Species, T Cells Default to a Th1 Phenotype and Mediate Protection against Pulmonary Cryptococcus neoformans Infection J. Immunol., October 15, 2006; 177(8): 5509 - 5516. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Spallarossa, P. Altieri, S. Garibaldi, G. Ghigliotti, C. Barisione, V. Manca, P. Fabbi, A. Ballestrero, C. Brunelli, and A. Barsotti Matrix metalloproteinase-2 and -9 are induced differently by doxorubicin in H9c2 cells: The role of MAP kinases and NAD(P)H oxidase Cardiovasc Res, February 15, 2006; 69(3): 736 - 745. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Kawahara, D. Ritsick, G. Cheng, and J. D. Lambeth Point Mutations in the Proline-rich Region of p22phox Are Dominant Inhibitors of Nox1- and Nox2-dependent Reactive Oxygen Generation J. Biol. Chem., September 9, 2005; 280(36): 31859 - 31869. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Crandall, Y. Ma, D. M. Dunn, R. S. Sundsbak, J. F. Zachary, P. Olofsson, R. Holmdahl, J. H. Weis, R. B. Weiss, C. Teuscher, et al. Bb2Bb3 Regulation of Murine Lyme Arthritis Is Distinct from Ncf1 and Independent of the Phagocyte Nicotinamide Adenine Dinucleotide Phosphate Oxidase Am. J. Pathol., September 1, 2005; 167(3): 775 - 785. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Y. Kassim, X. Fu, W. C. Liles, S. D. Shapiro, W. C. Parks, and J. W. Heinecke NADPH Oxidase Restrains the Matrix Metalloproteinase Activity of Macrophages J. Biol. Chem., August 26, 2005; 280(34): 30201 - 30205. [Abstract] [Full Text] [PDF]
|
|
PCR cycles). Data represent the mean mRNA expression of 3 intercrossed p47phox-/- and 3 WT mice.
