To determine the mechanisms involved in inflammatory responses to octacalcium phosphate (OCP) crystals in vivo.
To determine the mechanisms involved in inflammatory responses to octacalcium phosphate (OCP) crystals in vivo.
OCP crystal–induced inflammation was monitored using a peritoneal model of inflammation in mice with different deficiencies affecting interleukin-1 (IL-1) secretion (IL-1α–/–, IL-1β–/–, ASC–/–, and NLRP3–/– mice) or in mice pretreated with IL-1 inhibitors (anakinra [recombinant IL-1 receptor antagonist] and anti–IL-1β). The production of IL-1α, IL-1β, and myeloid-related protein 8 (MRP-8)–MRP-14 complex was determined by enzyme-linked immunosorbent assay. Peritoneal neutrophil recruitment and cell viability were determined by flow cytometry. Depletion of mast cells or resident macrophages was performed by pretreatment with compound 48/80 or clodronate liposomes, respectively.
OCP crystals induced peritoneal inflammation, as demonstrated by neutrophil recruitment and up-modulation of IL-1α, IL-1β, and MRP-8–MRP-14 complex, to levels comparable with those induced by monosodium urate monohydrate crystals. This OCP crystal–induced inflammation was both IL-1α– and IL-1β–dependent, as shown by the inhibitory effects of anakinra and anti–IL-1β antibody treatment. Accordingly, OCP crystal stimulation resulted in milder inflammation in IL-1α–/– and IL-1β–/– mice. Interestingly, ASC–/– and NLRP3–/– mice did not show any alteration in their inflammation status in response to OCP crystals. Depletion of the resident macrophage population resulted in a significant decrease in crystal-induced neutrophil infiltration and proinflammatory cytokine production in vivo, whereas mast cell depletion had no effect. Finally, OCP crystals induced apoptosis/necrosis of peritoneal cells in vivo.
These data indicate that macrophages, rather than mast cells, are important for initiating and driving OCP crystal–induced inflammation. Additionally, OCP crystals induce IL-1–dependent peritoneal inflammation without requiring the NLRP3 inflammasome.
Basic calcium phosphate (BCP) crystals including hydroxyapatite, carbonated apatite, tricalcium phosphate, and octacalcium phosphate (OCP) have long been associated with rheumatic syndromes. BCP crystal deposition occurs most frequently in soft tissue, muscle, and articular sites and can manifest with acute inflammation and tissue degradation. Indeed, the presence of BCP crystals in synovial fluid is more common in patients with more severe osteoarthritis (OA) (1). Furthermore, it has recently been reported that BCP crystal deposition in knee and hip cartilage is associated with end-stage OA (2, 3). In the study concerning hip OA (3), the amount of calcification (predominantly BCP crystals) correlated with clinical symptoms and histologic OA grade. The role of inflammation itself in OA disease progression is still uncertain. BCP crystals have also been associated with destructive arthropathies such as the Milwaukee shoulder syndrome (4). However, the mechanisms that underlie the inflammatory reaction induced by BCP crystals remain unclear.
In vitro, BCP crystals induce fibroblast proliferation, protooncogene stimulation, production of inflammatory cytokines (interleukin-1 [IL-1] and tumor necrosis factor β), metalloproteinase production and activation, cyclooxygenase 1 (COX-1), COX-2, and prostaglandin E2 production (5, 6), and chondrocyte production of nitric oxide and apoptosis (7, 8). In vivo, BCP crystals have been reported to be proinflammatory, inducing neutrophil influx in the rat air pouch model (9). Recently, a role for IL-1β has been demonstrated in monosodium urate monohydrate (MSU) crystal– and calcium pyrophosphate dihydrate (CPPD) crystal–induced inflammation; MSU and CPPD crystals are associated with acute gout and pseudogout, respectively. It is not known whether BCP crystals induce inflammation through this pathway.
IL-1β is a potent inflammatory cytokine, the production of which is tightly controlled at the level of gene expression, proteolytic processing, and secretion (10). Thus, proIL-1β protein (protein of 35 kd molar mass) is converted to active IL-1β (protein of 17 kd molar mass) mainly by caspase 1, but other leukocyte proteinases such as proteinase 3, elastase, chymase, and granzyme A may also be involved during inflammation (11–16). The activity of caspase 1 is regulated by the inflammasome, an intracellular multicomponent complex that is assembled following cellular stimuli from pathogens and danger signals (17). Several inflammasome complexes have been described, and activation of the inflammasome has been linked to infectious and autoinflammatory diseases (for review, see ref.17).
NLRP3 is thus far the best characterized inflammasome and is formed by the adaptor protein ASC, caspase 1, and NLRP3 (18). NLRP3 gain-of-function mutations are responsible for one of the hereditary autoinflammatory syndromes, cryopyrin-associated periodic syndrome, that responds dramatically to IL-1 inhibition (19). Similarly, NLRP3 is needed for monocyte IL-1β production upon stimulation with MSU and CPPD crystals (20), and studies have demonstrated the clinical efficacy of IL-1β blockade in both acute gout and pseudogout attacks (21–23). However, the situation in vivo could be different, as has been suggested by studies in which a role of the inflammasome was not demonstrated in murine models of arthritis that are well known to be IL-1β dependent (15). Possible explanations for this discrepancy include the contribution of multiple cell types to the inflammatory state in vivo which is not the case in vitro, and the possibility that crystals exert other effects on tissues to provoke an inflammatory response independent of IL-1β production. Finally, crystals may interact with host proteins in vivo to modify their phlogistic effects, as has been demonstrated in MSU crystal–induced inflammation (24, 25). This prompted us to assess in vivo the inflammatory effect of OCP crystals using the murine peritonitis model, and to dissect the mechanisms involved in IL-1β production. Furthermore, the contribution of the NLRP3 inflammasome in OCP crystal–induced inflammation and peritonitis was investigated.
C57BL/6J mice were purchased from Harlan. IL-1α–/– and IL-1β–/– mice were a gift from Dr. Yoichiro Iwakura (University of Tokyo, Tokyo, Japan) (26). ASC–/– mice (27) and NLRP3–/– mice (20) were backcrossed into the C57BL/6J background for at least 9 generations and were compared with wild-type (WT) littermates in this study. Mice were bred under conventional, non–specific pathogen–free conditions. Mice ages 8–12 weeks were used for experiments. Institutional approval was obtained for these experiments.
Sterile, pyrogen-free MSU and BCP crystals were synthesized as previously described (9, 20). Crystals were suspended in sterile phosphate buffered saline (PBS) and dispersed by brief sonication. All crystals were determined to be endotoxin free (<0.01 endotoxin units/10 mg of crystal) by Limulus amebocyte cell lysate assay.
Mice were injected intraperitoneally (IP) with 1 mg of MSU or OCP crystals in 0.5 ml sterile PBS. To analyze the involvement of IL-1, mice were injected IP with either 10 μg of neutralizing rabbit polyclonal anti–IL-1β antibody (in 0.5 ml PBS) (Novartis) or 200 μg of anakinra (recombinant IL-1 receptor antagonist [IL-1Ra]) (Kineret; Amgen) 30 minutes prior to crystal administration. An equal volume of sterile PBS was injected into control mice. To test the involvement of neutrophil proteinases, mice were treated with 1 mg of the neutrophil elastase inhibitor methoxysuccinyl-alanyl-alanyl-prolyl-valine-chloromethylketone (AAPV; Calbiochem) 1 hour before crystal administration. (AAPV was initially resuspended in DMSO at 10 mg/ml and diluted in PBS at 1 mg/ml for injections.) Control mice were injected with the vehicle alone. After 6 hours, blood was collected, mice were euthanized by CO2 administration, and peritoneal exudate cells were subsequently harvested by performing lavage with 3 ml of PBS. Total numbers of viable peritoneal exudate cells were determined by trypan blue exclusion. Lavage fluids were centrifuged at 450g for 10 minutes. Supernatants were used for analysis of cytokines and myeloid-related protein 8 (MRP-8)–MRP-14 complex. Cells were subjected to cytospin staining and flow cytometric analysis. Neutrophil numbers in the peritoneal exudate cells were determined by multiplying the total cell numbers by the percentage of lymphocyte antigen 6 complex, locus G (Ly-6G)–positive CD11b+ cells in individual mice.
Peritoneal exudate cells were resuspended in fluorescence-activated cell sorting (FACS) buffer (5% fetal calf serum [FCS] plus 5 mM EDTA in PBS) and incubated with conjugated monoclonal antibodies (mAb). The mAb used were phycoerythrin-conjugated anti–Ly-6G (clone RB6-8C5), fluorescein isothiocyanate (FITC)–conjugated anti-CD11b (clone M1/70), and allophycocyanin-conjugated anti-F4/80 (clone BM8) (all from eBioscience). Peritoneal cells (1 × 106) were incubated with appropriate conjugated antibodies for 30 minutes at 4°C in the dark. Stained cells were subsequently washed twice in FACS buffer and fixed in BD CellFIX solution (BD Biosciences). All data acquisition was performed on a FACSCalibur flow cytometer (BD Biosciences) using CellQuest software (BD Biosciences). Data analysis was performed using FlowJo software (Tree Star).
Cytokine levels in harvested lavage fluid supernatant were analyzed by IL-1α ELISA (BioLegend) and IL-1β ELISA (eBioscience) according to the manufacturers' instructions. Concentrations of MRP-8–MRP-14 complex in serum and peritoneal supernatants were determined by ELISA as previously described (28).
Macrophages were isolated from the peritoneal cavity of C57BL/6J mice as described previously (29). Briefly, naive mice were given IP injections of 4% sterile thioglycolate (0.5 ml). After 4 days, peritoneal cells collected by lavage were seeded at 1 × 106/ml in RPMI 1640 medium supplemented with 10% calf serum and antibiotics for 4 hours to allow the macrophages to adhere to the plates. Nonadherent cells were subsequently removed and adherent macrophages were used for experiments.
Bone marrow–derived mast cells were generated from bone marrow of C57BL/6J mice as described previously (30). Briefly, naive mice were killed and intact femurs and tibias were harvested. Sterile RPMI 1640 medium was repeatedly flushed through the bone shaft using a syringe with a 25-G needle. After lysis of red blood cells, cells were washed and cultured at a concentration of 1 × 106/ml in RPMI 1640 supplemented with 10% FCS, 100 units/ml penicillin, and 100 μg/ml streptomycin. Recombinant mouse IL-3 (5 ng/ml; R&D Systems) was added weekly to the cultures. Nonadherent cells were transferred to fresh medium at least once per week. Cells were used after 8 weeks of culture, when a mast cell purity of >95% was achieved as assessed by toluidine blue staining and FACS analysis of CD117 (c-Kit) expression using FITC-conjugated anti-CD117 mAb (clone 2B8; eBioscience).
Treatment with compound 48/80 (Sigma) was based on slight modifications to a previously reported protocol (31). Briefly, mice were treated IP with compound 48/80 (2 daily injections of 10 μg) for 3 days before crystal administration. Control mice received sterile PBS. Mast cell depletion was confirmed by identification of toluidine blue–stained cells in the peritoneal fluid, following which mice were immediately challenged with an IP dose of 1 mg of OCP crystals.
Clodronate liposomes were kindly provided by Dr. Nico van Rooijen (VU University Medical Center, Amsterdam, The Netherlands) and were prepared as previously described (32). Mice were given an IP injection of 200 μl of clodronate liposomes. Control mice received liposomes containing PBS. Three days later, macrophage depletion in liposome-treated mice was confirmed by flow cytometry, following which mice were immediately challenged with an IP dose of 1 mg of OCP crystals.
Cell viability was assessed by flow cytometric analysis using the FITC Annexin V Apoptosis Detection Kit I (BD Biosciences) according to the manufacturer's instructions. Briefly, 6 hours after IP injection of 1 mg of OCP crystals, peritoneal cells were recovered and stained with FITC-conjugated annexin V and propidium iodide (PI). Viable, early apoptotic, and late apoptotic and/or necrotic cells were identified as annexin V negative/PI negative, annexin V positive/PI negative, and annexin V positive/PI positive, respectively.
All values are expressed as the mean ± SEM. Variation between data sets was evaluated using Student's t-test or one-way analysis of variance where appropriate. P values less than 0.05 were considered significant. Data were analyzed using GraphPad Prism software.
To study the inflammatory response to OCP crystals in vivo, we used an established model of crystal-induced neutrophil infiltration into the peritoneal cavity. In previous experiments we determined that IP administration of OCP crystals induces a dose-dependent accumulation of neutrophils at the site of crystal deposition, with a plateau effect observed between 0.5 mg and 1 mg (data not shown). Therefore, for subsequent in vivo experiments, we used a dose of 1 mg of OCP crystals/mouse. MSU crystals injected at the same dose were used as a positive control. We initially assessed the in vivo kinetics of neutrophils and macrophages during OCP crystal–induced peritonitis. Compared with baseline levels (PBS-injected negative controls), absolute numbers of neutrophils (Ly-6G+CD11b+) in the peritoneal lavage fluid increased gradually following IP injection of OCP crystals, reaching a plateau by 6 hours (Figure 1A). In contrast, absolute numbers of macrophages (Ly-6G–CD11b+) decreased substantially from baseline values 3 hours after IP injection of OCP crystals (Figure 1B). After 6 hours, the numbers of macrophages within the peritoneal lavage fluid began to increase slowly, although this increase was not statistically significant. Since 6 hours after OCP crystal administration appears to be a good time point for showing the acute reaction of both neutrophils and macrophages in response to OCP crystal injection into the peritoneal cavity, the 6-hour time point was used as the end point for peritoneal lavage fluid harvest for subsequent in vivo experiments.
We next compared the inflammatory response to OCP and MSU crystals in vivo. Upon IP administration, OCP crystals were able to induce Ly-6G+CD11b+ neutrophil influx at levels comparable with those achieved following injection of MSU crystals (Figure 1C). As expected, the absolute numbers of peritoneal Ly-6G–CD11b+ macrophages decreased significantly after crystal injection (mean ± SEM 706,000 ± 128,000 in PBS-injected mice, 396,000 ± 230,000 in OCP crystal–injected mice, and 121,000 ± 38,000 in MSU crystal–injected mice) (data not shown). The recruitment of neutrophils was associated with elevated levels of both IL-1α and IL-1β in the peritoneal lavage fluid of crystal-injected mice compared with levels in PBS-injected controls (Figures 1D and E, respectively). These results also correlated with a significant increase in the level of MRP-8–MRP-14 complex, which is considered a reliable marker of inflammation (33, 34), both in peritoneal lavage fluid (Figure 1F) and in serum samples (Figure 1G). Taken together, these observations demonstrate that in vivo OCP crystal stimulation is able to induce strong inflammation both locally and systemically.
We next assessed whether OCP crystal–triggered inflammation acts via IL-1–dependent pathways. Naive mice were treated with anakinra (recombinant IL-1Ra), which binds tightly to IL-1R type I, blocking the activity of either IL-1α or IL-1β (35). Anakinra-treated mice displayed a significant reduction in both OCP crystal– and MSU crystal–induced neutrophil recruitment (Figure 2A). Decreased OCP crystal– and MSU crystal–induced neutrophil infiltration into the peritoneal cavity was also observed in mice treated with IL-1β–neutralizing antibodies. Collectively, these results demonstrate that IL-1R activation and IL-1 production play essential roles in OCP crystal–triggered inflammation.
Since we showed that both IL-1α and IL-1β levels were increased in peritoneal lavage fluid samples from crystal-injected mice (Figures 1D and E, respectively) and that anti–IL-1β antibodies and recombinant IL-1Ra were efficient in blocking neutrophil recruitment, we next assessed the relative contributions of IL-1α and IL-1β in OCP crystal–induced inflammation. Mice deficient in either IL-1α or IL-1β were injected with OCP crystals, and peritoneal neutrophil recruitment was assessed 6 hours following crystal administration (Figure 2B). In the absence of IL-1α, there was a striking and significant decrease in neutrophil recruitment upon OCP crystal injection. A similar decrease was observed in IL-1β–deficient mice, although this was not statistically significant. As expected, IL-1α–/– and IL-1β–/– mice indeed did not produce IL-1α and IL-1β, respectively, after OCP crystal administration, and the absence of one of these cytokines affected the production of the other (Figures 2C and D). These results indicate that IL-1α and IL-1β as independent cytokines are important in crystal-induced neutrophil recruitment. However, these 2 soluble factors when functioning separately may not be solely responsible for facilitating crystal-induced neutrophil accumulation in the peritoneal cavity.
Since we have shown that IL-1β plays a prominent role in OCP crystal–induced inflammation, we next investigated the contribution of the inflammasome, the multiprotein complex able to convert proIL-1β into biologically active IL-1β, to OCP crystal–induced inflammation. We investigated the roles of the adaptor protein ASC and of the inflammasome sensor NLRP3. ASC–/– and NLRP3–/– mice were injected IP with OCP crystals. Compared with WT mice, these deficient mice did not show any alteration in their inflammatory response, as demonstrated by similar numbers of neutrophils in the peritoneal cavity (Figure 3A) along with comparable levels of MRP-8–MRP-14 complex (Figure 3B) and IL-1β (Figure 3C) in the peritoneal lavage fluid. In addition, anakinra had similar inhibitory effects in WT, ASC–/–, and NLRP3–/– mice (Figure 3D). Taken together, these results indicate that these OCP crystal–associated inflammatory responses are independent of the classic NLRP3 inflammasome and suggest that other inflammasomes may play a role, or, alternatively, that OCP crystal–induced inflammation involves a caspase 1–independent IL-1β–processing mechanism. In this context we have tested the involvement of neutrophil elastase, a proteinase able to convert proIL-1β into biologically active IL-1β (13, 36). Indeed, when mice were treated prophylactically with AAPV, an inhibitor of neutrophil elastase, we found a nonsignificant trend toward a decrease (a 30% reduction) in both neutrophil recruitment and IL-1β levels in peritoneal fluid (data not shown).
We first tested whether mast cells were able to respond to OCP crystals by releasing IL-1α and IL-1β, thereby contributing to the onset of OCP crystal–induced inflammation. We therefore generated bone marrow–derived mast cells of >95% purity. OCP crystal stimulation of mast cells, either previously primed with lipopolysaccharide (LPS) or not, resulted in a massive release of IL-1α and IL-1β into the supernatant (Figures 4A and B, respectively). To study in vivo the effect of this OCP crystal–associated induction of IL-1 by mast cells, we next induced OCP crystal–mediated peritonitis in mast cell–depleted mice. Mice locally injected with compound 48/80 had >95% depletion of peritoneal mast cells (Figure 4C). We observed comparable levels of neutrophil recruitment between OCP crystal–injected mice pretreated with compound 48/80 and those pretreated with PBS (Figure 4D). These results suggest that resident mast cells do not play an important role in the cellular response induced following crystal administration.
Peritoneal macrophages could also be likely candidates that respond to OCP crystals by releasing IL-1α and IL-1β and thereby triggering OCP crystal–induced inflammation. When purified peritoneal macrophages that were previously primed with LPS (37) were stimulated with OCP crystals, cells were able to release both IL-1α and IL-1β in supernatants (Figures 5A and B). We next analyzed the inflammatory response to OCP crystals in mice depleted of resident macrophages by pretreatment with clodronate liposomes (32). An IP injection of clodronate liposomes 3 days prior to peritoneal lavage fluid harvest resulted in >90% depletion of resident macrophages (Figure 5C). Control mice received liposomes containing PBS. We found that macrophage-depleted mice were not able to recruit neutrophils upon IP administration of OCP crystals, whereas in non–macrophage-depleted mice the vast majority of peritoneal cells were neutrophils (Figures 5C and D). As expected, IL-1β levels in peritoneal lavage fluid were significantly decreased in mice pretreated with clodronate liposomes (Figure 5E).
We have found that OCP crystals have a deleterious effect on the viability of murine bone marrow–derived macrophages cultured in vitro (∼50% cell death after 6-hour incubation with 500 μg/ml OCP crystals) (37). In order to assess the relevance of such a phenomenon following in vivo OCP crystal administration, peritoneal cells were costained with annexin V and PI. We found that ∼50% of cells recovered 6 hours after IP injection of OCP crystals were dead (either in late apoptosis or in necrosis). In contrast, the vast majority of cells from PBS-injected mice were viable (Figures 6A and B).
OCP crystals, members of the BCP crystal family, elicit joint as well as periarticular inflammation and may have a pathogenic role in OA. The mechanisms of inflammation due to BCP crystals have not been extensively studied. Here we report that in the murine peritonitis model, OCP crystals cause a strong recruitment of neutrophils that is comparable with levels achieved with MSU crystals, whose proinflammatory properties have been well documented (20, 38, 39). These findings prompted us to investigate the molecular mechanisms responsible for neutrophil accumulation in an experimental model of OCP crystal–induced peritonitis.
The recent emergence of the inflammasome as a caspase 1 activator and its role in hereditary autoinflammatory syndromes and in gout raises the question of whether this complex is equally relevant in IL-1β production in BCP crystal– and, especially, OCP crystal–induced inflammatory situations. To address the role of the inflammasome in OCP crystal–induced pathogenesis, we assessed the effect of genetic deletion of different elements of the inflammasome on neutrophil recruitment into the peritoneal cavity upon OCP crystal administration. We found no effect of either ASC or NLRP3 deficiency on neutrophil recruitment, and similar amounts of MRP were detected in the peritoneal cavity. The absence of an obvious influence of NLRP3 and ASC on OCP crystal–induced inflammation unequivocally rules out a role for the NLRP3 inflammasome in OCP-induced inflammatory responses in vivo. These findings contrast with observations that MSU crystals (20), as well as inorganic particles such as asbestos fibers, silica particles, and alum crystals, activate the NLRP3 inflammasome to produce IL-1β (40–42).
The presence of IL-1β in the peritoneal exudate and the attenuation of inflammation by IL-1β blockade and IL-1β deficiency suggest that there is an NLRP3 inflammasome–independent mechanism. We and others have recently demonstrated the existence of inflammasome-independent pathways of IL-1β processing in IL-1β–mediated diseases such as antigen-induced arthritis (43), collagen-induced arthritis (44), K/BxN serum transfer–induced arthritis (14), acute arthritis (15), and experimental models of infections (for review, see ref.11). In these disease models, enzymes distinct from caspase 1 are able to process proIL-1β (discussed in refs.11 and16). Neutrophil-, macrophage-, and mast cell–derived serine proteinases such as proteinase 3, elastase, cathepsin G, and chymase have been reported to be able to convert proIL-1β into the 21-kd active form (10, 12, 13). A crucial role of chymase and elastase in proIL-1β activation was recently shown in the K/BxN arthritis model using specific proteinase inhibitors (14). Consistent with this latter finding, we tested the effects of AAPV in OCP crystal–induced peritonitis and found a 30% reduction in neutrophil recruitment and peritoneal IL-1β (this difference did not reach significance [data not shown]), suggesting that a part of proIL-1β processing in OCP crystal–induced peritonitis is due to neutrophil elastase.
OCP crystal–induced inflammation in the peritonitis model depends on both IL-1α and IL-1β. The first line of evidence for this was the finding that both IL-1α and IL-1β were released into the peritoneal exudate after crystal injection. A role for both cytokines was highlighted by the inhibitory effects of individual deficiency of IL-1α and IL-1β. Interestingly, their effects on inflammation seem to be linked, since IL-1α deficiency had an effect on peritoneal IL-1β levels and vice versa. Such a reciprocal regulation of IL-1α over the production of IL-1β has been previously reported in another experimental model (26). IL-1α is biologically active in its precursor form and can be found on the surface of several cells, particularly monocytes, where it is referred to as membrane IL-1α (10). Cleavage of the precursor by calpain, a membrane-associated calcium-activated cysteine proteinase, releases mature IL-1α. It may also be released from dying cells (45).
We demonstrated that OCP crystals induced necrosis of ∼50% of peritoneal cells, principally represented by infiltrating neutrophils. In this context, it is likely that at least part of the peritoneal fluid IL-1α that we measured following OCP crystal injection was released from dying cells. Interestingly, it has been shown that IL-1α released from necrotic cells triggers CXCL1/cytokine-induced neutrophil chemoattractant (KC) secretion and recruitment of neutrophils via IL-1R/myeloid differentiation factor 88 signaling on neighboring mesothelial cells (46). Therefore, we can anticipate that the inflammatory properties of OCP crystals include their ability to induce cellular necrosis. The subsequent passive release of IL-1α from dying cells would in turn facilitate chemokine production (eventually CXCL1) and neutrophil recruitment to the inflamed site. Validation of such a mechanism would require crystal administration in mice with a targeted mutation in CXCL1/KC (47).
It has recently been shown that in addition to passive release of danger signals such as uric acid or ATP, necrotic cells drive inflammatory cell infiltration in vivo and induce the production of IL-1β in an NLRP3 inflammasome–dependent manner (42, 48). Our data do not support the influence of such a mechanism, since neutrophil recruitment and IL-1β production were not altered in the absence of the NLRP3 inflammasome. Such a discrepancy could be explained by the dose of necrotic cells required to activate the NLRP3 inflammasome. To trigger a sterile inflammatory response through NLRP3, 107 necrotic cells were injected IP (48). In our experimental setting, cells were progressively dying and we recovered ∼10-fold fewer necrotic cells 6 hours following crystal administration. Thus, below a certain threshold of necrosis, the NLRP3 inflammasome is probably not activated.
Mast cells and resident macrophages have been implicated in MSU crystal–induced peritonitis (36, 38), and we investigated the contribution of these cells in OCP crystal–induced peritonitis. Depletion studies using either compound 48/80 or clodronate liposomes showed that only peritoneal macrophage depletion led to a reduction of neutrophil recruitment and IL-1β levels, demonstrating that resident macrophages play an essential role in the production of IL-1β and in neutrophil recruitment in OCP crystal–induced inflammation, even though in in vitro studies, both cell types secreted IL-1β when stimulated with OCP crystals. We recently found that OCP crystals stimulated IL-1β secretion in murine bone marrow–derived macrophages and peritoneal macrophages through an NLRP3 inflammasome–dependent pathway in vitro (37). The difference, in terms of NLRP3 dependency, between the in vivo and in vitro results may be due to OCP crystal–induced cell death and the release of non–caspase 1 proteinases already discussed, or to factors that can modulate crystal interactions with cells, such as protein coating of crystals in vivo (24, 25).
An interesting observation during crystal-induced peritonitis was that peritoneal macrophages decreased significantly in the peritoneal lavage fluid following IP administration of both OCP and MSU crystals. Similar disappearance of macrophages has been previously observed in harvested peritoneal lavage fluid following IP injection with MSU crystals (49). This disappearance of macrophages shortly following administration of inflammatory stimuli has been termed “macrophage disappearance reaction” and has been highlighted in other models of acute inflammation (50–52).
Finally, we observed a significant release of MRP-8–MRP-14 complex (S100A8/A9, calprotectin) during OCP crystal–induced peritonitis, which may further amplify the local inflammatory response. MRPs are secreted by monocytes and neutrophils following cellular activation or necrosis (for review, see ref.53) and participate in a positive feedback loop of neutrophil recruitment by up-regulating integrin expression and mediating chemotaxis (54, 55). Not only peritoneal but also serum levels of MRP-8–MRP-14 complex were increased upon OCP crystal injection. Similar results were reported in a murine air pouch model of MSU crystal–induced inflammation that was inhibited by anti-MRP antibodies (56). This suggests that in BCP crystal–related diseases such as the Milwaukee shoulder syndrome, MRP levels will be increased and might play a role in pathogenesis.
In conclusion, we have demonstrated that in an in vivo model, OCP crystals induce inflammation via IL-1α and IL-1β, independent of the NLRP3 inflammasome, and this process is linked to cell death induced by crystals (Figure 6C). Furthermore, our data highlight that macrophages play a crucial role in this inflammatory process but mast cells do not. These mechanisms can account for the acute inflammatory reaction seen in acute periarthritis and arthritis due to OCP crystals. In OA, in which BCP crystals are found in the cartilage as well as in the joint fluid, these mechanisms do not seem to predominate, since inflammation is not a prominent feature. These results have implications in the search for effective therapies for BCP crystal–associated diseases.
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. Busso 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. Narayan, Pazar, Ea, Kolly, Bagnoud, Chobaz, Lioté, So, Busso.
Acquisition of data. Narayan, Pazar, Ea, Kolly, Bagnoud, Chobaz, Vogl, Holzinger.
Analysis and interpretation of data. Narayan, Pazar, Ea, Kolly, Bagnoud, Chobaz, Lioté, Vogl, Holzinger, So, Busso.
We are grateful to Dr. Nico van Rooijen (VU University Medical Center, Amsterdam, The Netherlands) for kindly providing us with clodronate liposomes. We thank Dr. Yoichiro Iwakura (University of Tokyo, Tokyo, Japan) for providing us with IL-1α–/– and IL-1β–/– mice and Professor Jurg Tschopp (University of Lausanne, Lausanne, Switzerland) for providing us with NLRP3–/– and ASC–/– mice.