NLRP3 inflammasome–mediated neutrophil recruitment and hypernociception depend on leukotriene B4 in a murine model of gout

Authors


Abstract

Objective

Deposition of monosodium urate monohydrate (MSU) crystals in the joints promotes an intense inflammatory response and joint dysfunction. This study evaluated the role of the NLRP3 inflammasome and 5-lipoxygenase (5-LOX)–derived leukotriene B4 (LTB4) in driving tissue inflammation and hypernociception in a murine model of gout.

Methods

Gout was induced by injecting MSU crystals into the joints of mice. Wild-type mice and mice deficient in NLRP3, ASC, caspase 1, interleukin-1β (IL-1β), IL-1 receptor type I (IL-1RI), IL-18R, myeloid differentiation factor 88 (MyD88), or 5-LOX were used. Evaluations were performed to assess neutrophil influx, LTB4 activity, cytokine (IL-1β, CXCL1) production (by enzyme-linked immunosorbent assay), synovial microvasculature cell adhesion (by intravital microscopy), and hypernociception. Cleaved caspase 1 and production of reactive oxygen species (ROS) were analyzed in macrophages by Western blotting and fluorometric assay, respectively.

Results

Injection of MSU crystals into the knee joints of mice induced neutrophil influx and neutrophil-dependent hypernociception. MSU crystal–induced neutrophil influx was CXCR2-dependent and relied on the induction of CXCL1 in an NLRP3/ASC/caspase 1/IL-1β/MyD88–dependent manner. LTB4 was produced rapidly after injection of MSU crystals, and this was necessary for caspase 1–dependent IL-1β production and consequent release of CXCR2-acting chemokines in vivo. In vitro, macrophages produced LTB4 after MSU crystal injection, and LTB4 was relevant in the MSU crystal–induced maturation of IL-1β. Mechanistically, LTB4 drove MSU crystal–induced production of ROS and ROS-dependent activation of the NLRP3 inflammasome.

Conclusion

These results reveal the role of the NLRP3 inflammasome in mediating MSU crystal–induced inflammation and dysfunction of the joints, and highlight a previously unrecognized role of LTB4 in driving NLRP3 inflammasome activation in response to MSU crystals, both in vitro and in vivo.

Gout is caused by deposition of urate crystals, a byproduct of purine degradation, in the joints and kidneys (1). The prevalence of gout and of elevated levels of free uric acid in the serum of patients (hyperuricemia) has increased in the last century, and it is estimated that more than 1% of adult men in Western countries have gout (2). Even though the condition can be managed in many patients by controlling the levels of uric acid in plasma, it is very incapacitating and painful when joints are affected by deposition of urate crystals. Recurrent gout attacks may lead to incapacitation and renal dysfunction (2, 3).

It is unclear how monosodium urate monohydrate (MSU) crystals engage surface immune receptors to trigger phagocytosis and signaling in cells. Although it is possible that proteins from the serum, such as immunoglobulins, bind crystals and facilitate their recognition by immune cells (4, 5), it has recently been demonstrated that direct receptor-independent binding of crystals to dendritic cells can occur and is sufficient to trigger signaling cascades (6). NLRP3 and its adaptor protein, ASC, mediate caspase 1–dependent processing of certain cytokines, especially interleukin-1β (IL-1β) (7, 8). There is now clear evidence that MSU crystals trigger activation of the NLRP3/ASC/caspase 1 inflammasome, an effect that culminates in the production of IL-1β (9–11). Indeed, mice deficient in the components of the NLRP3 inflammasome have impaired neutrophil recruitment in response to injection of MSU crystals (9, 11). However, it is unclear what triggers NLRP3 inflammasome activation after the injection of MSU crystals, and the sequence of events initiated by inflammasome-derived IL-1β that culminates in leukocyte influx and functional joint impairment in vivo has not been determined.

In the present study, we investigated the sequence of events leading to neutrophil influx in a murine model of gout, induced by the injection of MSU crystals into the joints of mice, and we came to the unexpected finding that leukotriene B4 (LTB4) was relevant in the assembly of the NLRP3 inflammasome after stimulation with MSU crystals. Injection of MSU crystals into the knee joints of the mice caused neutrophilic inflammation and hypernociception. Neutrophil influx was dependent on CXCR2, and CXCR2 ligand production was driven by NLRP3 inflammasome–derived IL-1β production. Inhibition of 5-lipoxygenase (5-LOX)–derived LTB4 prevented the neutrophil influx, but this could not be attributed to a direct neutrophil-recruiting ability of LTB4. Indeed, the presence of LTB4 was necessary for the optimal production of IL-1β in vivo, and for the optimal production of reactive oxygen species (ROS) and NLRP3 inflammasome activation in vitro.

MATERIALS AND METHODS

Animals.

Eight-to-ten-week-old male C57BL/6J (wild-type [WT]) mice were purchased from the Centro de Bioterismo of the Universidade Federal de Minas Gerais (UFMG) in Brazil. Gene-deficient mice (NLRP3−/−, ASC−/−, caspase 1−/−, IL-1β−/−, IL-1 receptor type I–deficient [IL-1RI−/−], myeloid differentiation factor 88–deficient [MyD88−/−], and IL-18R−/−) (9, 12–17) and their controls were bred and obtained from the Transgenose Institute at the Centre National de Recherche Scientifique (CNRS) in Orléans, France. All animals were maintained with filtered water and food ad libitum and kept in a controlled environment. Experiments received prior approval by the animal ethics committee of the UFMG and CNRS.

Reagents.

Uric acid, lipopolysaccharide (LPS) from Escherichia coli, fucoidin, and N-acetyl-L-cysteine were obtained from Sigma. Boric acid was from Promega, and uricase/fasturtec was from Sanofi-Synthelabo. IL-1 receptor antagonist (IL-1Ra) was supplied by Biogen. The inhibitors DF2162, CP105,696, and MK886 were provided by Dompe Pharma, Pfizer, and Calbiochem, respectively. LTB4 was from Cayman Chemical. The MSU crystals were prepared as described previously (18).

Gout model.

Mice were placed under anesthesia (150:10 mg/kg ketamine:xylazine injected intraperitoneally [IP]) and were injected with MSU crystals (3–300 μg; 10 μl) into the tibiofemoral knee joint. The amount of endotoxin present in the MSU crystals injected was <5 pg (at the higher dose), as assessed using a Chromogenic Limulus amebocyte lysate assay (Endosafe). Inflammation parameters were evaluated at different time points after injection of the MSU crystals (1, 3, 6, 15, and 24 hours after injection), and mechanical hypernociception was also evaluated at various time points after injection.

Groups of mice were culled for cervical dislocation, and the articular cavity was washed with phosphate buffered saline (PBS)–3% bovine serum albumin (2 × 5 μl) for cell counts. Periarticular tissue was removed from the joints for evaluation of cytokine and chemokine production and LTB4 and myeloperoxidase (MPO) activity. The number of leukocytes from the articular cavity was determined in a Neubauer chamber, after staining the tissue with Turk's solution. Differential counts were performed in Cytospin (Shandon III) preparations by evaluating the percentage of each leukocyte on a slide stained with May-Grünwald-Giemsa. In some experiments, LTB4 was injected into the knee joint (50 ng/cavity), and samples were processed as described above.

Evaluation of hypernociception.

Evaluation of hypernociception was performed as described previously (19), using an electronic pressure meter (INSIGHT Instruments). The flexion-elicited withdrawal threshold was used to infer behavioral responses associated with pain. Results are expressed as the change in withdrawal threshold, calculated by subtracting zero-time mean measurements from time-interval mean measurements (in grams).

Cytokine, chemokine, and MPO determination.

Periarticular tissue was collected and homogenized in PBS containing antiproteases (20). Samples were processed and the supernatant was evaluated for cytokine and chemokine concentrations, in accordance with the manufacturer's instructions (R&D Systems) (20). MPO activity (a quantitative measurement of neutrophil sequestration) in periarticular tissue homogenates, standardized to the number of neutrophils obtained from the peritoneal cavity of casein-injected mice, was assayed as described previously (20).

Enzymatic immunoassay for LTB4..

The tissue surrounding the joints was obtained at various time points after MSU crystal injection, and then homogenized in 500 μl buffer containing protease inhibitors (21). In cell culture experiments, supernatants of peritoneal macrophages were collected at 1, 3, and 6 hours after stimulation with MSU crystals (300 μg/ml). LTB4 levels were determined by enzyme-linked immunosorbent assay (ELISA) using a commercial kit (Biotrak; Amersham Pharmacia Biotech) in accordance with the manufacturer's instructions.

Intravital microscopy.

Intravital microscopy was performed in the synovial microcirculation of the mouse knee, as described previously (20). Leukocytes were stained with rhodamine 6G (0.5 mg/kg body weight) and observed using a microscope (Nikon Eclipse 50i, 20× objective lens) outfitted with a fluorescent light source (epi-illumination at 510–560 nm, using a 590-nm emission filter). Leukocytes were considered adherent to the venular endothelium if they remained stationary for a minimum of 30 seconds. Rolling leukocytes were defined as white cells moving at a velocity lower than that of erythrocytes within a given vessel.

In vitro studies.

Preparation of mouse macrophages.

Mice were injected IP with 3% thioglycolate solution, and macrophages were collected by peritoneal lavage 4 days later. Cells were plated at a density of 1 × 106 cells in 96-well plates, and nonadherent cells were removed after 2 hours. Cells were cultured in RPMI medium complemented with 10% fetal calf serum, penicillin/streptomycin, and L-glutamine, and kept in a humidified incubator at 37°C with 5% CO2 overnight. LPS (10 ng/ml, overnight) was used to prime the cells. The cells were then stimulated with MSU crystals (300 μg/ml) or LTB4 (10−9 to 10−7M) for 6 hours. CP105,696 (10 μM) was added to the cell cultures 40 minutes before administration of the MSU crystals. Supernatants were collected for protein measurements by ELISA (R&D Systems) and Western blotting.

Western blot assay.

The cleaved form of caspase 1 was detected in supernatants of peritoneal macrophages using Western blotting with rat anti–caspase 1 p20 monoclonal antibody clone 4B4 (Genentech), as previously described (9).

ROS assay.

Peritoneal macrophages were cultured in 96-well plates (1 × 106 cells/well) using RPMI medium without phenol red (Gibco), supplemented with 10% heat-inactivated fetal calf serum (Gibco) for 24 hours in a humidified incubator at 37°C with 5% CO2. Cells were then loaded for 30 minutes with the ROS-specific fluorescent probe H2DCFDA (2′,7′-dichlorofluorescein diacetate, 20 μM final concentration; Sigma-Aldrich), washed twice with preheated medium, and exposed to MSU crystals (300 μg/ml). In the treated group, cells were incubated with 10 μM CP105,696 for 40 minutes before the addition of the MSU crystals. Fluorescence was assessed in 10-minute intervals over 1 hour, using a spectrofluorometer (Synergy 2; BioTek) with a fluorescein isothiocyanate filter (excitation 485 nm, emission 538 nm).

Statistical analysis.

Results were tested for normality and are shown as the mean ± SEM. Differences between groups were evaluated using analysis of variance followed by Student-Newman-Keuls post hoc test, using normal or normalized (by log transformation) data. The level of significance was set at P values less than 0.05.

RESULTS

Crucial role of the NLRP3 inflammasome complex and IL-1β in inflammatory and functional responses in murine articular gout.

Injection of 100 μg of MSU crystals was optimal to induce inflammatory and functional changes in the knee joints of mice (details available from the corresponding author upon request). Levels of the neutrophil-active chemokines CXCL1 and CXCL2 and IL-1β were optimally produced with a dose of 100 μg of MSU crystals, and this was associated with neutrophil influx in the joint and around the joint, as assessed by MPO assay (results not shown).

Consistent with the findings from the MPO assay, histopathologic analysis showed neutrophil infiltration in the synovial fluid and in the soft tissues surrounding the joint, as well as marked hyperplasia of the synovial epithelium (details available from the corresponding author upon request). Neutrophil accumulation in the joint was already detectable at 3 hours and persisted up to 15 hours after administration of MSU crystals (Figure 1A). Neutrophil accumulation in the periarticular tissue was only detectable at significant levels at 15 hours after stimulation (Figure 1B). Levels of CXCL1 peaked at 3 hours after administration of MSU crystals, and dropped thereafter (Figure 1C), whereas levels of CXCL2 were stable from 3 hours to 15 hours after administration of the crystals (Figure 1D). Periarticular levels of IL-1β were already detectable at 3 hours, peaked at 6 hours, and were still high at 15 hours after intraarticular injection of the crystals (Figure 1E).

Figure 1.

Kinetics of the inflammatory response induced by injection of monosodium urate monohydrate (MSU) crystals into the knee joints of mice. Intraarticular injection of MSU crystals (100 μg/knee joint) induced the accumulation of neutrophils into the articular cavity (A) and in the periarticular tissue as assessed by myeloperoxidase (MPO) assay (B), at 3, 6, 15, and 24 hours after injection. In addition, the neutrophil-active chemokines CXCL1 (C) and CXCL2 (D) and the cytokine interleukin-1β (IL-1β) (E) were produced in the periarticular tissue after MSU crystal injection. Tissue inflammation was accompanied by joint dysfunction, as noted by increased nociceptive responses to mechanical stimulation after MSU crystal injection, assessed as the change in withdrawal threshold (F). Bars show the mean ± SEM results in 6–8 mice per group. = P < 0.01 versus animals injected with vehicle (V).

Pain is the most common clinical finding, after acute gout, in humans and is a major cause of disability associated with gout. We have recently described a new method that allows for concomitant evaluation of articular hypernociception and inflammation (19). When we applied this method in our murine model of gout, mechanical hypernociception, an index of pain and measurement of articular dysfunction, followed a pattern that was similar to that of IL-1β, with levels of hypernociception peaking at 6 hours and being maintained at a raised level at 15 hours after crystal injection (Figure 1F). All parameters of inflammation and hypernociception had returned to background levels within 24 hours after administration of MSU crystals (Figures 1A–F).

In order to confirm that the observed effects were indeed due to the crystals and not secondary to any contaminant in the MSU crystal preparations, we pretreated mice with uricase (0.4 mg/kg IP) 40 minutes before injection of the MSU crystals into the knee joint. Uricase abolished the inflammatory and functional changes induced by MSU crystals (polymorphonuclear neutrophil influx into the joint, mean ± SEM 12 ± 5 × 104 neutrophils/cavity with MSU alone versus 0.6 ± 0.6 × 104 neutrophils/cavity with MSU + uricase; IL-1β levels, mean ± SEM 560 ± 60 pg IL-1β/100 mg tissue with MSU alone versus 13 ± 1 pg IL-1β/100 mg tissue with MSU + uricase; each n = 3). As mentioned above, detection of LPS contamination was low in the MSU crystal preparations, and injection of MSU crystals in Toll-like receptor 2 (TLR-2)/TLR-4 double-knockout mice induced an inflammatory response that was similar to that observed in WT mice (results not shown).

It has been previously demonstrated that the NLRP3 inflammasome is a relevant factor in the inflammatory responses induced by injection of MSU crystals in mice (9, 11). Using mice deficient in NLRP3, ASC, or caspase 1, we confirmed that MSU crystal–induced articular inflammation was dependent on the NLRP3 inflammasome, as seen by the abrogation of neutrophil influx (Figure 2A) and reduced production of CXCL1 and IL-1β (Figures 2B and C) in the gene-deficient mice. In addition, hypernociceptive responses were also found to be dependent on the NLRP3 inflammasome, as shown by the marked decrease in MSU crystal–induced hypernociception in NLRP3−/−, ASC−/−, and caspase 1−/− mice (Figure 2D).

Figure 2.

Essential role of the NLRP3 inflammasome and IL-1β signaling in MSU crystal–induced articular inflammation and dysfunction. Mice were injected with MSU crystals (100 μg/joint) and parameters of inflammation and hypernociception were assessed 15 hours later. Inflammatory and functional responses elicited by injection of MSU crystals were decreased in mice deficient in NLRP3, ASC, or caspase 1 (Casp-1) (A–D) and in mice deficient in IL-1β, IL-1 receptor type I (IL-1RI), and myeloid differentiation factor 88 (MyD88) (E–H), as observed by a decrease in neutrophil accumulation in the articular cavity (A and E), a decrease in the levels of CXCL1 (B and F) and IL-1β (C and G) in the periarticular tissue, and reduced mechanical hypernociception (D and H). Bars show the mean ± SEM results in 6–8 mice per group. = P < 0.01 versus animals injected with vehicle; # = P < 0.01 versus wild-type (WT) mice injected with MSU crystals. See Figure 1 for other definitions.

Previous studies have shown that IL-1β production is necessary for MSU crystal–induced inflammation (22–24). The activation of the NLRP3 inflammasome is associated with caspase 1–dependent processing of genes of the IL-1 cytokine family, including IL-1β and IL-18 (9). Experiments in mice lacking IL-18R suggest that IL-18 production and activation of the receptor were not important for the phenotype observed in mice lacking components of the NLRP3 inflammasome system (results not shown). In contrast, MSU crystal–induced inflammatory and hypernociceptive responses were greatly decreased in the mice lacking IL-1β, IL-1RI, or MyD88 (Figures 2E and H). Treatment with IL-1Ra also decreased the inflammatory and hypernociceptive responses induced by MSU crystals (results not shown). Taken together, these results show that activation of the NLRP3 inflammasome, release of IL-1β, activation of IL-1R, and downstream signaling via MyD88 are essential processes in the induction of chemokine production and leukocyte recruitment.

Mechanisms of neutrophil influx and relevance to hypernociception.

The interaction of MSU crystals with resident cells, especially macrophages (25, 26), in the cavity and surrounding tissues is believed to be the primary factor for triggering MSU crystal–induced inflammation, leading to acute neutrophil ingress and paroxysms of gouty inflammation episodes. Neutrophil influx into the synovium and joint fluid is the pathologic hallmark of acute gout, and these cells are thought to contribute to the pathogenesis of gout (27–29). As seen in Figure 3, treatment with an allosteric inhibitor of CXCR2 (DF2162), the main receptor for the neutrophil-related chemokines CXCL1 and CXCL2, greatly decreased the neutrophil influx into the joint (Figure 3A) and surrounding tissue (Figure 3B).

Figure 3.

Important role of neutrophils in MSU crystal–induced articular inflammation and dysfunction. Mice were injected with MSU crystals (100 μg/joint) and parameters of inflammation and hypernociception were assessed 15 hours later. The role of neutrophils was investigated using the allosteric antagonist of CXCR2, DF2162 (DF) (15 mg/kg orally, 40 minutes before MSU crystal injection), or the selectin inhibitor fucoidin (FUC) (20 mg/kg intravenously, 15 minutes before MSU crystal injection). Both treatments greatly decreased the number of neutrophils accumulated in the intraarticular cavity (A) and the periarticular tissue (B) after injection of MSU crystals. Treatment with DF2162 also decreased the number of cells adherent to synovial microvessels of the knee joint, as assessed by intravital microscopy (C), and treatment with DF2162 or fucoidin reduced mechanical hypernociception after MSU crystal injection (D). Bars show the mean ± SEM results in 6–8 mice per group. = P < 0.01 versus animals injected with vehicle alone as control; # = P < 0.01 versus wild-type vehicle-treated (V) mice injected with MSU crystals. See Figure 1 for other definitions.

Intravital microscopy studies showed that the major role of CXCR2 in the system was to mediate recruitment of neutrophils, akin to the findings in previous studies in antigen-immunized and -challenged animals (20). Indeed, administration of the CXCR2 antagonist DF2162 greatly decreased the adhesion of leukocytes (Figure 3C), but not the rolling of leukocytes (results not shown), to synovial microvessels. In this system, >90% of rolling and adherent cells are neutrophils, as was previously shown by the decrease in events 6 hours after MSU crystal injection, following the administration of anti–Gr-1/Ly6G/C (RB6-8C5 at 0.5 mg IP, administered 12 hours before MSU injection) (30), a neutrophil-depleting antibody (results not shown). Therefore, neutrophil influx induced by MSU crystal injection is mediated by CXCR2.

Previous studies from our group have suggested that influx of neutrophils is relevant in the induction of inflammatory hypernociception by several stimuli (20, 31). As seen in Figure 3, fucoidin, a selectin inhibitor, greatly decreased the neutrophil influx into the joints (Figure 3A) and surrounding tissue (Figure 3B) after MSU crystal injection. More importantly, treatment with either fucoidin or the CXCR2 inhibitor significantly blocked MSU crystal–induced hypernociception (Figure 3D). Therefore, NLRP3 inflammasome–initiated, IL-1β–driven production of CXCR2-active chemokines drives neutrophil influx and neutrophil-dependent hypernociception after MSU crystal injection into the knee joints of mice.

Regulation of the MSU crystal–induced inflammatory response by leukotrienes.

LTB4 is a very active lipid mediator derived from the action of 5-LOX on arachidonic acid (32). Our previous studies have shown that LTB4 can act synergistically with chemokines to facilitate recruitment of leukocytes in vivo (33). As seen in Figure 4, 5-LOX deficiency in mice or treatment of mice with MK886, a 5-LOX activation inhibitor compound (FLAP inhibitor), greatly decreased the accumulation of neutrophils (Figure 4A), hypernociception (Figure 4B), and the levels of IL-1β and CXCL1 (Figures 4C and D) in the periarticular tissue in response to injection of MSU crystals.

Figure 4.

Essential role of 5-lipoxygenase (5-LOX) in MSU crystal–induced articular inflammation and dysfunction. Mice were injected with MSU crystals (100 μg/joint). Intravital microscopy was conducted to assess cell adhesion at 6 hours after injection, and all other parameters were evaluated at 15 hours. The role of 5-LOX was investigated using 5-LOX–deficient mice or the 5-LOX inhibitor MK886 (MK) (3 mg/kg orally, 40 minutes before MSU crystal injection). In animals injected with MSU crystals, neutrophil accumulation was decreased in the intraarticular cavity of 5-LOX–deficient or MK886-treated mice (A), and this was associated with reduced mechanical hypernociception (B) and reduced levels of IL-1β (C) and CXCL1 (D) in the periarticular tissue. Treatment with MK886 reduced leukocyte adhesion when given before MSU crystal injection (E) but not when given just before the intravital microscopy procedure (F). Bars show the mean ± SEM results in 6–8 mice per group. = P < 0.01 versus animals injected with vehicle alone as control (C); # = P < 0.01 versus wild-type, non–inhibitor-treated mice injected with MSU crystals. See Figure 1 for other definitions.

In order to investigate whether, akin to CXCR2 antagonists, 5-LOX is involved in neutrophil adhesion and consequent migration, we performed intravital experiments to assess the mouse synovial microvasculature after injection of MSU crystals. As seen in Figure 4E, treatment of mice with MK886, given before the injection of MSU crystals, decreased neutrophil adhesion to the microvasculature, which is consistent with the inhibitory effects of MK886 on neutrophil accumulation in the knee joint. However, MK886 did not affect neutrophil adhesion when given just prior to the intravital procedure (Figure 4F), suggesting that the major role of 5-LOX–derived products in the system was not to directly mediate neutrophil recruitment, but rather to act indirectly via control of the production of IL-1β and consequent release of chemokines.

IL-1β levels remained at baseline levels at 1 hour after MSU crystal injection (results not shown) and peaked at 6 hours after induction with MSU crystals (Figure 1E). Interestingly, LTB4 levels peaked as early as 1 hour after administration of MSU crystals and declined thereafter to near-baseline levels at 6 hours (Figure 5A), suggesting that production of LTB4 preceded production of IL-1β. Treatment with an LTB4 receptor (BLT1) antagonist, CP105,696, significantly decreased the accumulation of neutrophils and hypernociception in the knee joint (Figures 5B and C). Interestingly, experiments in 5-LOX–deficient mice or in mice treated with MK886 or CP105,696 showed that there was also marked inhibition of MSU crystal–induced production of IL-1β (Figures 4C and 5D) and CXCL1 (Figures 4D and 5E). Taken together, these results support the notion that LTB4 is the main 5-LOX metabolic factor associated with the response to MSU crystals, and that LTB4 is important in the cascade of events leading to IL-β and CXCL1 production, and the consequent CXCR2-mediated neutrophil migration.

Figure 5.

Production and essential role of leukotriene B4 (LTB4) in MSU crystal–induced articular inflammation and dysfunction. Levels of LTB4 increased at 1, 3, and 6 hours after injection in the periarticular tissue of mice injected with MSU crystals (100 μg/joint), compared to vehicle-injected mice (A). In animals injected with MSU crystals, treatment with the LTB4 antagonist CP105,696 (CP), compared to treatment with saline as vehicle control, decreased neutrophil accumulation in the intraarticular cavity (B), and this was associated with reduced mechanical hypernociception (C) and reduced production of IL-1β (D) and CXCL1 (E). Bars show the mean ± SEM results in 6–8 mice per group. = P < 0.01 versus animals injected with vehicle alone as control; # = P < 0.01 versus saline (vehicle)–treated mice injected with MSU crystals. See Figure 1 for other definitions.

As IL-1β production was shown to be dependent on activation of the NLRP3 inflammasome, we hypothesized that 5-LOX–derived LTB4 could play a role in facilitating caspase 1 activation after stimulation with MSU crystals. It must be noted, however, that injection of LTB4 into the knee joints of mice induced neutrophil influx and IL-1β production that was similar between WT, ASC−/−, and caspase 1−/− mice (details available from the corresponding author upon request). Because LTB4 was given in pharmacologic doses and administered as a single injection in these experiments, we reasoned that different compartments could be activated, and that LTB4 could still be playing an endogenous role in mediating MSU crystal–induced NLRP3 inflammasome activation and inflammation.

Promotion of caspase 1 activation and IL-1β production by LTB4.

Addition of MSU crystals to thioglycolate-elicited peritoneal macrophages induced the production of LTB4 in the supernatant (Figure 6A). Significant levels of LTB4 were already detected in the supernatants of nonprimed macrophages, but the levels increased markedly when macrophages were primed with LPS (Figure 6A), a stimulus used for induction of proIL-1β in vitro. In LPS-primed, thioglycolate-elicited peritoneal macrophages, LTB4 was able to induce the production of IL-1β (Figure 6B). Indeed, the levels of IL-1β were ∼50% of that produced in response to MSU crystals.

Figure 6.

Activation of the inflammasome and IL-1β production via the production of reactive oxygen species (ROS) by leukotriene B4 (LTB4). Macrophages stimulated with MSU crystals (300 μg/ml), compared to vehicle controls, released LTB4, and this was enhanced by previous priming with lipopolysaccharide (LPS) (10 ng/ml, overnight) (A). Activation of macrophages with LTB4 (10−8M) or MSU crystals (300 μg/ml), compared to vehicle-injected controls, induced IL-1β release (B). LTB4-stimulated macrophages induced cleavage of caspase 1, and macrophages pretreated with the antioxidant N-acetyl-L-cysteine (NAC) (25 mM) reduced the caspase 1 cleavage induced by MSU crystals or LTB4 (10−8M) (C). Treatment of LPS-primed macrophages with the LTB4 receptor antagonist CP105,696 (CP) (10 μM) decreased MSU crystal–induced IL-1β release (D). Nonprimed macrophages stimulated with MSU crystals or LTB4 induced production of reactive oxygen species (ROS), which was rapid in onset, and the MSU crystal–stimulated ROS production was significantly reduced by treatment with CP105,696 (E). Representative results from the latter experiments are shown as the area under the curve (AUC) (F). Bars show the mean ± SEM results in at least 4 animals per group, and experiments were repeated at least twice. Western blots show representative results from 1 of at least 3 repeated experiments. = P < 0.01 versus nonstimulated (vehicle-injected) control macrophages; # = P < 0.01 versus nonprimed macrophages or versus non–antagonist-treated macrophages stimulated with MSU crystals or LTB4 alone. See Figure 1 for other definitions.

Corroborating these findings, we detected the cleaved form of caspase 1 in the supernatant of LTB4-stimulated macrophages, demonstrating an activation of the inflammasome complex by LTB4 (Figure 6C). Treatment of the macrophages with N-acetyl-L-cysteine significantly decreased the levels of cleaved caspase 1 after stimulation with LTB4 (Figure 6C), suggesting that caspase 1 cleavage was ROS-dependent. Furthermore, treatment with CP105,696 decreased the production of IL-1β, as seen by ELISA (Figure 6D). Taken together, these results show that LTB4 contributes to MSU crystal–induced NLRP3 inflammasome assembly and the consequent production of IL-1β in vitro.

A series of experiments was performed to investigate the mechanism by which LTB4 production could lead to IL-1β production. It has been suggested that ROS generated in spatial and temporal proximity to the inflammasome may activate NLRP3 and foster inflammasome assembly (34, 35). The addition of MSU crystals to macrophages induced rapid production of ROS, which was of a magnitude similar to that induced by LTB4 (Figures 6E and F). Importantly, treatment of cells with CP105,696 significantly decreased the MSU crystal–induced production of ROS (Figures 6E and F). Taken together, these findings suggest that the capacity of LTB4 to induce production of ROS underlies its role in mediating NLRP3 inflammasome activation.

DISCUSSION

Deposition of MSU crystals causes gout and renal disease and may also play a role in immune regulation (3). Herein, we evaluated the cascade of events leading to localized inflammation and hypernociception after injection of MSU crystals into the tibiofemoral joints of mice. The major findings of the present study can be summarized as follows. First, consistent with the findings from previous studies (9, 18), MSU crystal–induced inflammation relies on the activation of the NLRP3 inflammasome, release of IL-1β, activation of IL-1R, and activation of MyD88. This is, however, the first time that functional changes in the joint, i.e., hypernociception, have been shown to be dependent on the same pathways as those for joint inflammation.

Second, the influx of neutrophils in the joint after injection of MSU crystals is dependent on CXCR2, and this makes a major contribution to articular hypernociception. The production of the CXCR2-active chemokine CXCL1 is driven by the NLRP3 inflammasome and IL-1β.

Third, 5-LOX–derived LTB4 plays a crucial role in mediating neutrophil recruitment and hypernociception. Although 5-LOX is not necessary for the direct migration of neutrophils, it was found to control the release of IL-1β and the consequent release of chemokines. In vitro, LTB4 induces and is necessary for MSU crystal–induced production of IL-1β by LPS-stimulated macrophages.

Fourth, mechanistically, LTB4 drives production of ROS and ROS-dependent caspase 1 cleavage and IL-1β production by LPS-primed MSU crystal–activated macrophages. Thus, taken together, the findings from the present study unravel a previously unrecognized role of 5-LOX–derived LTB4 in driving NLRP3 inflammasome–associated inflammation in vivo.

There are conflicting reports in the literature regarding the participation of the NLRP3 inflammasome in MSU crystal–induced inflammation. Although some groups have demonstrated that the NLRP3 inflammasome is relevant in the induction of inflammatory responses by IP or subcutaneous injection of MSU crystals into the knee joints of mice (9, 11), Joosten and colleagues demonstrated that injection of purified MSU crystals into the knee cavity did not promote joint inflammation (36). In our experiments, injection of MSU crystals into the joints of mice induced leukocyte recruitment that was NLRP3/ASC/caspase 1/IL-1β/IL-1RI/MyD88–dependent. Moreover, we have shown, for the first time, that hypernociception, an index of inflammatory pain in experimental animals, induced by MSU crystals also depends on the same molecular pathway.

Neutrophil influx into the synovium and joint fluid is the pathologic hallmark of acute gout, and these cells are thought to contribute to the pathogenesis of gout (27–29). In the present study, the absence of components of the NLRP3 inflammasome system or IL-1β signaling was accompanied by a reduction in neutrophil recruitment, which is consistent with that observed in other studies in the peritoneal cavity (9). Indeed, our study showed that administration of a CXCR2 inhibitor prevented MSU crystal–induced influx of neutrophils into the knee joints of mice by preventing their adhesion to the synovial microvasculature. Moreover, our study clearly showed that production of CXCR2-active chemokines (CXCL1 and CXCL2) was driven by NLRP3 inflammasome/IL-1β signaling.

Previous studies have shown that neutrophils may be the main effectors of inflammatory pain in different models of inflammation (20, 31, 37). In our study, prevention of neutrophil influx decreased joint hypernociception, demonstrating a major contribution of these cells in driving joint dysfunction after MSU crystal injection. Therefore, NLRP3 inflammasome–initiated, IL-1β–driven production of CXCR2-active chemokines drives neutrophil influx and neutrophil-dependent hypernociception after MSU crystal injection into the knee joints of mice.

Neutrophils are known to contribute to the production of IL-1β under some inflammatory conditions (38, 39), suggesting that neutrophil influx may also contribute to the production of IL-1β at the late stages of MSU crystal–induced inflammation. However, it is important to mention that other mediators and cell types not analyzed in the present study could contribute to MSU crystal–induced hypernociception. For example, a recent study demonstrated that particulate materials, including silica crystals and aluminum salt, are able to induce prostaglandin E2 production independent of the NLRP3 inflammasome (40).

Despite the accepted finding that MSU crystals induce IL-1β maturation in an NLRP3 inflammasome–dependent manner, less is known of the molecular mechanisms by which MSU crystals trigger the assembly and activation of the NLRP3 inflammasome. Studies have suggested that the NLRP3 inflammasome can sense particulate structures directly, via recognition by leucine-repeat domains (11), or indirectly with participation of signaling from key intermediates, such as ROS (34, 35). Phagocytosis of silica crystals and aluminum salts promotes NLRP3 inflammasome activation via lysosomal destabilization, with an important participation of cathepsin B release (41). Lysosomal damage may cause release of cathepsin B, which is then able to trigger ROS production in neurons (42). Consistent with the notion of a possible central role of ROS in driving NLRP3 inflammasome activation, a few studies have shown that NLRP3 agonists, including particulate activators, induce ROS production, and blockade of ROS by chemical scavengers can suppress NLRP3 inflammasome activation (35, 43–45).

Our study showed a critical role of 5-LOX and LTB4 in MSU crystal–induced inflammation and hypernociception. Previous studies have already shown that MSU crystals may induce production of LTB4 (46–48). However, our study is the first to show that these molecules do not appear to drive neutrophil influx directly, akin to CXCR2-active chemokines, but act indirectly via control of the production of IL-1β and consequent release of chemokines. Indeed, MSU crystals induced LTB4 production both in vivo and in vitro, and genetic or pharmacologic abrogation of 5-LOX, the crucial enzyme responsible for the production of LTB4, or treatment with an LTB4 antagonist decreased the MSU crystal–induced production of IL-1β in vivo or the LPS-primed production of thioglycolate-elicited peritoneal macrophages in vitro. In contrast, exogenous LTB4 induced recruitment of neutrophils and IL-1β production that was similar between WT, ASC−/−, and caspase 1−/− mice. Therefore, although we show that LTB4 is necessary for MSU crystal–induced triggering of the NLRP3 inflammasome, the results would suggest that administration of pharmacologic quantities of LTB4 can indeed induce neutrophil influx in a caspase 1–independent manner.

IL-1β can be generated from neutrophils in a caspase 1–independent manner (38). In the case of MSU crystals, generation of IL-1β is caspase 1–dependent and driven by LTB4, probably in resident cells. Injection of LTB4 exogenously may increase the levels of these mediators in different compartments, and may result in kinetic patterns that are different from those seen after injection of MSU crystals, likely accounting for the differences observed.

LTB4 is known to be a powerful inducer of ROS generation by several cell types, including macrophages (32, 49). In our experiments, treatment with a ROS scavenger (N-acetyl-L-cysteine) decreased caspase 1 activation in macrophages stimulated with LTB4. LTB4 did induce ROS production by macrophages, and blockade of the LTB4 receptor significantly decreased ROS production in response to MSU crystals. Therefore, interaction of MSU crystals with the lipid membrane may induce early production of LTB4, by a mechanism yet to be defined, which, in turn, is able to induce NLRP3 inflammasome assembly and production of IL-1β in a ROS-dependent manner. IL-1β, acting on its receptor, IL-1R, then drives chemokine production, which, in turn, activates CXCR2 and induces neutrophil recruitment.

Migrated neutrophils play a relevant role in the functional impairment of the joints (hypernociception). Blockade of 5-LOX or LTB4 prevents all of the major events evaluated, including IL-1β production, neutrophil migration, and functional joint impairment. Therefore, such strategies may be useful in the treatment of conditions, such as gout, that are associated with MSU crystal deposition.

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. Teixeira 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. Amaral, Ryffel, Souza, Teixeira.

Acquisition of data. Amaral, Costa, Tavares, Sachs, Coelho, Fagundes, Soriani, Silveira, L. D. Cunha, Peres, T. M. Cunha.

Analysis and interpretation of data. Amaral, Zamboni, Quesniaux, F. Q. Cunha, Souza, Teixeira.

Acknowledgements

We thank Ilma Marçal, Dora Aparecida Alves Rodrigues, and Gilvânia Ferreira da Silva Santos (Universidade Federal de Minas Gerais, Brazil) for their technical assistance, and we are grateful to Sergio Dalmora (Universidade Federal de Santa Maria, Brazil) for providing the LPS measurements.

Ancillary