Innate immunity conferred by toll-like receptors 2 and 4 and myeloid differentiation factor 88 expression is pivotal to monosodium urate monohydrate crystal–induced inflammation

Authors


Abstract

Objective

In gout, incompletely defined molecular factors alter recognition of dormant articular and bursal monosodium urate monohydrate (MSU) crystal deposits, thereby inducing self-limiting bouts of characteristically severe neutrophilic inflammation. To define primary determinants of cellular recognition, uptake, and inflammatory responses to MSU crystals, we conducted a study to test the role of Toll-like receptor 2 (TLR-2), TLR-4, and the cytosolic TLR adapter protein myeloid differentiation factor 88 (MyD88), which are centrally involved in innate immune recognition of microbial pathogens.

Methods

We isolated bone marrow–derived macrophages (BMDMs) in TLR-2−/−, TLR-4−/−, MyD88−/−, and congenic wild-type mice, and assessed phagocytosis and cytokine expression in response to endotoxin-free MSU crystals under serum-free conditions. MSU crystals also were injected into mouse synovium-like subcutaneous air pouches.

Results

TLR-2−/−, TLR-4−/−, and MyD88−/− BMDMs demonstrated impaired uptake of MSU crystals in vitro. MSU crystal–induced production of interleukin-1β (IL-1β), tumor necrosis factor α, keratinocyte-derived cytokine/growth-related oncogene α, and transforming growth factor β1 also were significantly suppressed in TLR-2−/− and TLR-4−/− BMDMs and were blunted in MyD88−/− BMDMs in vitro. Neutrophil influx and local induction of IL-1β in subcutaneous air pouches were suppressed 6 hours after injection of MSU crystals in TLR-2−/− and TLR-4−/− mice and were attenuated in MyD88−/− mice.

Conclusion

The murine host requires TLR-2, TLR-4, and MyD88 for macrophage activation and development of full-blown neutrophilic, air pouch inflammation in response to MSU crystals. Our findings implicate innate immune cellular recognition of naked MSU crystals by specific TLRs as a major factor in determining the inflammatory potential of MSU crystal deposits and the course of gouty arthritis.

The deposition of monosodium urate monohydrate (MSU) crystals in articular joints and bursal tissues can be asymptomatic in gout or can erupt via incompletely defined factors into acute, episodic, self-limiting joint inflammation largely dependent on marked neutrophil influx (1–3). Because neutrophils are absent from normal joint fluid, the interaction of MSU crystals with resident cells in the joint is believed to be the primary factor stimulating neutrophil–endothelial adhesion in the synovial microvasculature (4), acute neutrophil ingress, and paroxysms of gouty inflammation (1, 4, 5).

Cells that encounter MSU crystals express a broad array of inflammatory mediators that contribute to acute gouty inflammation, including cyclooxygenase 2, tumor necrosis factor α (TNFα), interleukin-1 (IL-1), and IL-6 (5–9). Neutrophil chemotactic CXCR2-binding chemokines, including keratinocyte-derived cytokine (KC)/growth-related oncogene α (GROα)/CXCL1 and IL-8/CXCL4 (10), appear to be absolutely essential for the neutrophil-dependent inflammation triggered by the MSU crystals, as demonstrated previously in CXCR2−/− mice (11). Release of the neutrophil-expressed calgranulin heterodimer S100A8/A9, one of the most abundant protein constituents of the neutrophil cytoplasm (12), appears to substantially amplify neutrophil recruitment in gouty inflammation (13). Importantly, the state of macrophage differentiation is a major factor in the uptake of MSU crystals and sequelae, including the capacity to express transforming growth factor β1 (TGFβ1), a native suppressor of experimental gouty inflammation (14–16).

The naked MSU crystal has a highly negatively charged, reactive surface that nonspecifically binds more than 25 serum proteins (17) and also binds plasma membrane proteins, including integrins (18, 19). Significantly, MSU crystals activate both the classic and alternative complement pathways in vitro (20–23). In studies of C6-deficient rabbits, we recently discovered that local assembly of the C5b–9 membrane attack complex played a major role in both intraarticular IL-8 expression and acute neutrophilic inflammation in experimental MSU crystal–induced arthritis of the knee joint (24). These findings implicated the complement arm of innate immunity as one of the factors that controls acute gouty inflammation. Furthermore, we observed that MSU crystals activate the chondrocyte, a cell of mesenchymal origin, in a manner that requires canonical signal transduction via Toll-like receptor 2 (TLR-2) (25). TLRs play a vital role in host defense by initiating the innate immune response against pathogens (26). Here, we tested the hypothesis that innate immune–mediated cellular recognition of the naked MSU crystal through certain TLRs is critical in the capacity of MSU crystals to launch acute gouty inflammation.

There are more than 10 defined members in the TLR group (26). TLRs are type I transmembrane receptors characterized by the presence of extracellular leucine-rich repeat motifs that recognize pathogen-associated molecular patterns (26). Most TLRs have a cytoplasmic Toll/IL-1 receptor (IL-1R) domain, which is required for the activation of downstream signaling pathways that lead to the activation of NF-κB (26), a transcription factor activated rapidly by MSU crystals and centrally involved in MSU crystal–induced cell activation (27, 28). In response to pathogen-associated molecular patterns, the canonical TLR signaling pathway recruits the cytosolic TLR adapter protein myeloid differentiation factor 88 (MyD88), IL-1R–activated kinase, and TNF receptor–associated factor 6 to activate IKKs, and the process culminates in NF-κB–mediated expression of proinflammatory messenger RNA (26).

Ligation of certain TLRs also activates signal transduction pathways, leading to phagocytosis, killing, and clearance of pathogens by leukocytes (29). Recently, we observed that TLR-2–mediated and MyD88-dependent signaling in chondrocytes played critical roles in NF-κB activation and nitric oxide generation in response to MSU crystals (25). However, the direct exposure of MSU crystals to chondrocytes is generally limited in the joint, unlike the case with fibroblast-like and macrophage-like synovium lining cells (1). Hence, in this study, we sought to determine the roles of phagocyte expression of TLR-2, MyD88, and TLR-4 (30) in inflammatory responses to MSU crystals by macrophages in vitro, and in the synovium-like mouse subcutaneous air pouch (11, 31) in vivo.

MATERIALS AND METHODS

Reagents.

All chemical reagents were obtained from Sigma-Aldrich (St. Louis, MO), unless indicated otherwise. Triclinic MSU crystals were prepared under pyrogen-free conditions, using uric acid pretreated for 2 hours at 200°C prior to crystallization (10). The crystals were suspended at 25 mg/ml in sterile, endotoxin-free phosphate buffered saline (PBS), and verified to be free of detectable lipopolysaccharide contamination (<0.025 endotoxin units/ml) by the Limulus amebocyte cell lysate assay (BioWhittaker, Walkersville, MD). MSU crystals (14C labeled) were prepared as described above, using as a starting material trace 14C uric acid (Perkin Elmer, Boston, MA) added to 1 gm of cold uric acid. The specific activity of the 14C-labeled MSU crystals was 1.4 μCi/mg.

Isolation and culture of murine macrophages.

All animal experiments were performed humanely under institutionally approved protocols. TLR-2–knockout (TLR-2−/−), TLR-4−/−, and MyD88−/− mice on a C57BL/6 background (kindly provided by Dr. Shizuo Akira, University of Osaka, Japan) were maintained under specific pathogen–free conditions and genotyped by polymerase chain reaction, as previously described (32–34). Macrophages were prepared from bone marrow obtained from 8–10-week-old homozygous TLR-2−/−, TLR-4−/−, MyD88−/−, and congenic wild-type (WT) control mice on the C57BL/6 background.

Briefly, bone marrow cells were isolated from the femurs and tibias of the mice by flushing the medullary cavity with PBS containing 2% fetal calf serum (FCS). After 1 wash in the same solution, cells were seeded in tissue culture dishes in low-glucose Dulbecco's modified Eagle's medium supplemented with 10% FCS, 100 μg/ml of streptomycin, 100 IU/ml of penicillin, and 40 ng/ml of recombinant granulocyte–macrophage colony-stimulating factor (BioSource International, Camarillo, CA) (35) at 37°C for 7–9 days. Macrophages were then assessed by flow cytometry using a FACScan (Becton Dickinson Biosciences, San Jose, CA) by staining with allophycocyanin (APC)–conjugated anti-F4/80 (Caltag, Burlingame, CA), a marker preferentially expressed by mature macrophages (36). Double staining was performed using APC-conjugated anti-F4/80 and phycoerythrin-conjugated anti–TLR-2 or anti–TLR-4 (eBioscience, San Diego, CA) for TLR-2 or TLR-4 expression, or APC-anti-F4/80 and MyD88 polyclonal primary antibodies (eBioscience, San Diego, CA) and fluorescein isothiocyanate–conjugated goat anti-rabbit IgG (Jackson ImmunoResearch, West Grove, PA) as secondary antibodies for MyD88 expression. For double-staining studies, cells were permeabilized in Cytofix/Cytoperm (Becton Dickinson), following the manufacturer's protocol.

Assays of phagocytosis and cytokine production.

BMDMs of individual genotypes were treated with MSU crystals (0.5 mg/ml) for 2 hours at 4°C or 37°C and then washed 3 times with cold PBS containing 5 mM EDTA and harvested in the same buffer. The proportion of macrophages taking up MSU crystals was assessed by flow cytometry based on increased side scatter properties (36). The amount of MSU crystals ingested by the macrophages was determined under the same conditions using 14C-labeled MSU crystals by measuring 14C radioactivity in each sample of cells after they were washed 3 times in cold PBS containing 5 mM EDTA.

The generation of IL-1β, TNFα, KC/GROα, and TGFβ1 was evaluated by DuoSet enzyme-linked immunosorbent assay (ELISA; R&D Systems, Minneapolis, MN), following the manufacturer's protocol, by testing conditioned media collected from BMDMs of each genotype (5 × 105/well) stimulated with MSU crystals (0.5 mg/ml) for 24 hours.

Studies of synovium-like subcutaneous air pouches.

Subcutaneous pouches were generated by the injection of sterile, filtered air to create an accessible space that developed a synovium-like membrane within 7 days, as previously described (10). Briefly, anesthetized 10–12-week-old WT, TLR-2−/−, TLR-4−/−, and MyD88−/− mice were injected with 5 ml of sterile air into the subcutaneous tissue of the back, followed by a second injection of 3 ml of sterile air into the pouch 3 days later. MSU crystals (3 mg), in 1 ml of sterile, endotoxin-free PBS, were injected into the pouch 7 days after the first injection of air. Mice were killed and pouch fluids were harvested at specific time points by injecting 5 ml of PBS containing 5 mM EDTA, and cells infiltrating the air pouch were counted manually using a hemocytometer. Smears of cells from the air pouches were prepared by centrifugation of either 50 μl of the pouch exudates or 105 of cells in cytofunnels (ThermoShandon, Pittsburgh, PA) in a Cytospin 4 centrifuge (ThermoShandon) at 110g for 2 minutes. Differential leukocyte subpopulation counts were measured by Wright-Giemsa staining of cytospin slides. IL-1β in supernatants of air pouch exudates was measured by ELISA, as described above.

For histologic analysis of the air pouches, frozen sagittal sections of the air pouches were fixed in 80% ethanol and then stained with hematoxylin for 2–5 minutes. Sections were washed 3 times with water and incubated in ammonium hydroxide for 30 seconds, and then incubated with acid alcohol (0.5 % HCl in 70% ethanol) for 30 seconds. Sections were then washed 3 times again and counterstained with eosin for 3–5 minutes.

Statistical analysis.

Data are expressed as the mean ± SD. Statistical analyses were performed using Student's 2-tailed t-test. P values less than 0.05 were considered significant.

RESULTS

Roles of TLR-2 and MyD88 in macrophage responsiveness to MSU crystals in vitro.

We first tested the hypothesis that both TLR-2 and MyD88 mediated the capacity of MSU crystals to induce macrophage expression of selected cytokines previously implicated in regulating gouty inflammation and known to be directly and rapidly induced in cells via simple exposure to MSU crystals (IL-1β, TNFα, KC/GROα, and TGFβ1, as cited above). To do so, we generated BMDMs from TLR-2−/−, MyD88−/−, and congenic WT mice. The differentiation state of isolated BMDMs was confirmed by flow cytometry using F4/80 as a mature, macrophage-specific marker (36). In all experiments and for all genotypes studied, ≥85% of the isolated BMDMs were F4/80 positive. In F4/80-positive WT BMDMs, ≥95% expressed TLR-2, ≥85% expressed TLR-4, and ≥90% expressed MyD88, as assessed by flow cytometry of permeabilized cells.

To avoid potential masking effects of both serum protein opsonization of the crystals (1) and crystal-induced complement activation (24), we treated BMDMs with endotoxin-free MSU crystals at a concentration of 0.5 mg/ml under entirely serum-free conditions, based on our previous studies (27). At 24 hours, MSU crystals induced the production of several cytokines, including IL-1β, TNFα, KC/GROα, and TGFβ1 in WT BMDMs (Figure 1A), as determined by ELISA of conditioned media. However, production of each of these cytokines was partially, but substantially, reduced in TLR-2−/− BMDMs compared with WT BMDMs (Figure 1B). Blunting of production of each of these cytokines in response to MSU crystals was observed in MyD88−/− BMDMs (Figure 1).

Figure 1.

Effects of Toll-like receptor 2 (TLR-2) and myeloid differentiation factor 88 (MyD88) deficiency on cytokine production in bone marrow–derived macrophages (BMDMs) in response to monosodium urate monohydrate (MSU) crystals in vitro. BMDMs prepared from wild-type (WT), TLR-2−/−, and MyD88−/− mice were incubated with endotoxin-free MSU crystals (0.5 mg/ml) for 24 hours under serum-free conditions, as described in Materials and Methods. Levels of interleukin-1β (IL-1β), tumor necrosis factor α (TNFα), keratinocyte-derived cytokine (KC)/growth-related oncogene α (GROα), and transforming growth factor β1 (TGFβ1) in conditioned media were measured by enzyme-linked immunosorbent assay. A, Cytokine production in WT BMDMs. B, Percentage of cytokine production in TLR-2−/− and MyD88−/− BMDMs relative to WT BMDMs. Values are the mean and SEM of ≥5 individual experiments, using cells from ≥5 different mice of each genotype. ∗ = P < 0.05 versus WT mice.

Mediation of MSU crystal phagocytosis in the macrophage by TLR-2 and MyD88.

Ingestion of MSU crystals promotes cell activation in phagocytes (12, 37, 38), and TLR-2 signaling is known to mediate a phagocytic program (29) and to physically cooperate with other receptors in promoting phagocytosis of specific microbial pathogens (39). Therefore, we hypothesized that there was impaired uptake of MSU crystals in TLR-2−/− cells. First, we validated that our primary assay system discerned crystal uptake rather than nonspecific crystal–cell association by treating murine BMDMs with MSU crystals for 2 hours at 4°C versus 37°C, again under serum-free conditions. The proportion of macrophages containing MSU crystals was measured by flow cytometry, based on an increase in side scatter profile, as previously described (36). As seen in Figure 2A, there was an absence of detectable uptake of MSU crystals by either WT or TLR-2−/− BMDMs at 4°C. At 37°C, the proportion of TLR-2−/− BMDMs that took up MSU crystals was reduced by ∼50% relative to WT BMDMs.

Figure 2.

Decreased uptake of MSU crystals by TLR-2−/− and MyD88−/− BMDMs in vitro. BMDMs were incubated with either unlabeled or 14C-labeled MSU crystals (0.5 mg/ml) under the conditions indicated. A and C, The percentage of BMDMs that took up the unlabeled MSU crystals was analyzed by flow cytometry based on the increase in the side scatter profile. B and D, The amount of 14C-labeled MSU crystals ingested by BMDMs was determined by measuring the radioactivity of 14C associated with washed cells. In A and B, TLR-2−/− BMDMs were incubated with unlabeled MSU crystals for 2 hours at 4°C or 37°C, or with 14C-labeled MSU crystals for 2 hours at 37°C. In C and D, MyD88−/− BMDMs were incubated with unlabeled or 14C-labeled MSU crystals for 15 minutes and 2 hours at 37°C. Values in B and D are the mean and SEM of 3 different experiments on 3 different mice of each genotype. ∗ = P < 0.05 versus WT mice. SSC-H = side scatter height; FSC-H = forward scatter height; R1 = region 1; R2 = region 2 (see Figure 1 for other definitions).

To ascertain the effects of TLR-2 deficiency on the amount of crystals taken up by the BMDM population as a whole, we treated cells with 14C-labeled MSU crystals. The amount of 14C-labeled MSU crystals ingested by BMDMs of TLR-2−/− mice was decreased by ∼70% compared with WT BMDMs (Figure 2B). Similarly, the proportion of MyD88−/− BMDMs that took up MSU crystals was reduced by ∼50% compared with WT BMDMs, with inhibitory effects of MyD88 deficiency observed as early as 15 minutes after stimulation with MSU crystals (Figure 2C). Moreover, the amount of 14C-labeled MSU crystals ingested by MyD88−/− BMDMs was decreased by ∼70% at 15 minutes, and by ∼80% 2 hours after stimulation relative to WT BMDMs.

Effects of TLR-2 and MyD88 deficiency on MSU crystal–induced inflammation in the mouse synovium-like subcutaneous air pouch.

Next, we tested the roles of TLR-2 and MyD88 in MSU crystal–induced inflammation in vivo. To do so, we injected endotoxin-free MSU crystals into subcutaneous air pouches of TLR-2−/−, MyD88−/−, and WT mice, a model system characterized by generation of a synovium-like lining cell layer containing fibroblastic and phagocytic cells (31). In order to screen for inhibitory effects on a submaximal inflammatory response, we chose a dose of MSU crystals (3 mg in 1 ml of PBS) that was 70% lower than that used in our previous studies of MSU crystal–induced air pouch inflammation (11). At 0, 6, and 24 hours post–MSU crystal injection, mice were killed, and we measured both IL-1β production and leukocyte influx in the air pouch exudates. IL-1β expression (Figure 3C) and leukocyte ingress (Figure 3A) were robust in WT mice 6 hours after MSU crystals were injected into the air pouch, with neutrophils accounting for the majority of infiltrated leukocytes (Figure 3B). Both cytokine induction and neutrophilic inflammation were self-limiting by 24 hours postinjection (Figure 3). Six hours after MSU crystal injection, IL-1β expression and the total number of infiltrated leukocytes and neutrophils were partially, but significantly, suppressed in TLR-2−/− mice relative to WT mice (Figure 3). Furthermore, there was a virtual absence of IL-β induction as well as leukocyte ingress in the air pouch of MyD88−/− mice in response to MSU crystals (Figure 3).

Figure 3.

Suppressed infiltration of leukocytes and induction of IL-1β in response to MSU crystals in subcutaneous air pouches of TLR-2−/− and MyD88−/− mice. Subcutaneous air pouches were created on the backs of mice of the indicated genotypes via injections of sterile air, as described in Materials and Methods, and 7 days after initial generation of the air pouches, a 1-ml suspension of 3 mg MSU crystals in phosphate buffered saline (PBS) was injected into the air pouches. Mice were killed at the indicated times, and the air pouch exudates were harvested by washing with 5 ml of PBS containing 5 mM EDTA. A and B, The leukocytes in the pouch exudates were counted using a hemocytometer, and the fraction of neutrophils was determined using Wright-Giemsa staining. C, Supernatants of air pouch exudates were collected by centrifugation and IL-1β production was measured by enzyme-linked immunosorbent assay. Under these conditions, injection of PBS control alone was associated with a background of only 0.15 × 106 leukocytes per air pouch in WT mice, 0.13 × 106 leukocytes per air pouch in TLR-2−/− mice, and 0.05 × 106 leukocytes per air pouch in MyD88−/− mice throughout the time course. Values shown are the mean ± SEM of 8 WT mice, 9 TLR-2−/− mice, and 8 MyD88−/− mice. ∗ = P < 0.05 versus WT mice. See Figure 1 for other definitions.

In comparisons of TLR-2−/− and WT mice, we analyzed the histologic features of the air pouch following MSU crystal injection. The resting air pouches possessed a thin synovium-like lining and a thicker subcutaneous layer of vascularized fibrous and adipose tissue beneath the synovium-like lining (Figure 4). Six hours after the injection of MSU crystals, the architecture of the WT synovium-like lining layer became disrupted, and there was marked swelling of the tissues beneath the air pouch lining layer and massive infiltration of leukocytes throughout the air pouch lining and tissue immediately surrounding it (Figure 4). By comparison, at the same time point following MSU crystal injection in the TLR-2−/− air pouch, there was markedly less swelling of the synovium-like layer (Figure 4). There also was less intense infiltration of leukocytes in the TLR-2−/− air pouch, a result that correlated well with gross analyses of Wright-Giemsa–stained smears of cells in the air pouch exudates (Figure 4) and with the aforementioned counts of infiltrated leukocytes (Figures 3A and B).

Figure 4.

Comparison of histologic features of MSU crystal–induced inflammation in synovium-like subcutaneous air pouches in WT and TLR-2−/− mice. Top, frozen sections of air pouches stained with hematoxylin and eosin. Bottom, Wright-Giemsa–stained smears of cells from the same air pouches, following centrifugation of 50 μl of air pouch exudates at 110g for 2 minutes. Results shown are representative of 3 different experiments on 3 different mice of each genotype for each condition and time point shown. See Figure 1 for definitions.

Mediation of MSU crystal–induced macrophage activation and inflammation by TLR-4.

Results to this point indicated that TLR-2 deficiency partially impaired MSU crystal–induced inflammatory responses, as opposed to the nearly complete attenuation of these same responses in macrophages and air pouches of mice deficient in MyD88. Hence, we hypothesized that 1 or more TLRs other than TLR-2 also contributed significantly to the triggering of MSU crystal–induced inflammation. We elected to assess TLR-4, which, like TLR-2, is constitutively expressed in macrophages (30) and can modulate recognition and phagocytosis of microbial pathogens (29, 40). TLR-4−/− BMDMs showed marked impairment of production of TNFα, KC/GROα, and TGFβ1 in response to MSU crystals (Figure 5A) as well as impaired uptake of MSU crystals (Figures 5B and C). Finally, the capacity of MSU crystals to induce IL-1β expression, leukocyte ingress, and neutrophilic inflammation in the air pouch was markedly suppressed in TLR-4−/− mice (Figures 6A and B).

Figure 5.

Effects of TLR-4 deficiency on MSU crystal uptake and MSU crystal–induced cytokine expression in BMDMs in vitro. BMDMs prepared from WT and TLR-4−/− mice were incubated with MSU crystals (0.5 mg/ml) for 24 hours to determine TNFα, KC/GROα, and TGFβ1 expression by enzyme-linked immunosorbent assay (A), or for 2 hours to assess crystal uptake by flow cytometry (B) and levels of cell-associated 14C-labeled MSU crystals (C). Data shown in A, which are presented as a percentage of cytokine production relative to that in WT mice, were pooled from 5 different experiments on 5 WT and TLR-4−/− mice. Data shown in B and C are representative of 3 different experiments on 3 different mice of each genotype. Values in A and C are the mean and SEM. ∗ = P < 0.05 versus WT mice. SSC-H = side scatter height; FSC-H = forward scatter height; R1 = region 1; R2 = region 2 (see Figure 1 for other definitions).

Figure 6.

Suppressed infiltration of leukocytes and induction of IL-1β in response to MSU crystals in subcutaneous air pouches of TLR-4−/− mice. A suspension of 3 mg of MSU crystals in 1 ml of phosphate buffered saline (PBS) was injected into the air pouches, and MSU crystal–induced leukocyte infiltration (A), the number of neutrophils in the exudates (B), and IL-1β induction in the air pouch in vivo (C) (measured by enzyme-linked immunosorbent assay in supernatants of air pouch exudates collected by centrifugation) were determined. Under these conditions, injection of PBS control alone was associated with a background of only 0.15 × 106 leukocytes per air pouch in WT mice and 0.12 × 106 leukocytes per air pouch in TLR-4−/− mice. Values are the mean ± SEM of 8 WT mice and 9 TLR-4−/− mice. ∗ = P < 0.05 versus WT mice. See Figure 1 for other definitions.

DISCUSSION

In this study, we demonstrated that host expression of TLR-2, TLR-4, and their shared adapter protein MyD88 was a major determinant of the capacity of endotoxin-free MSU crystals to turn on the macrophage in vitro and to trigger acute neutrophilic inflammation in vivo. To avoid confounding effects of serum opsonins on the in vitro results, we limited this study to an evaluation of the effects of uncoated MSU crystals under serum-free conditions, including an analysis of cultured macrophage uptake of MSU crystals and cytokine release. But it was notable that the decreases in inflammation induced by injected MSU crystals in vivo in the air pouches of mice deficient in TLR-2, TLR-4, and MyD88 paralleled the decreases in cytokine production observed in MSU crystal–stimulated macrophages under serum-free conditions in vitro.

The inert MSU crystal has the capacity to avidly bind at least 20 different plasma proteins (1, 14, 15). Although such MSU crystal binding interactions are nonspecific via hydrogen and electrostatic bonds, MSU does associate preferentially with selected proteins in complex mixtures, as illustrated by apoB and complement pathway protein binding in crystals exposed to human plasma (41, 42). Clearly, naked MSU crystals can directly promote inflammation by activating complement and the contact coagulation system (1, 24). But our results support a model for triggering gouty inflammation in which the capacity of the MSU crystal to directly activate resident cells in the joint is pivotal (1).

MSU crystals bind IgG, which can promote recognition by phagocytes and enhanced cellular responses to the crystals (1, 43). However, the predominant effect of coating MSU crystals with plasma or serum is marked inhibition of cell activation via apoB binding to the crystal surface, which physically suppresses MSU crystal–cell binding (44). Synovial MSU crystals deposited in microscopic tophi have been observed to be tightly packed in a core contained by a protein-rich wall that includes fibrinogen (45). As such, our results here, and the known association of acute gouty attacks with rapid rises and falls in the levels of serum urate, suggest that gouty inflammation is triggered either by the de novo formation of uncoated MSU crystals in the joint or by the release from synovial tophi of MSU crystals liberated from coating proteins by factors including partial crystal dissolution.

MSU crystals preferentially bind to certain cell membrane proteins, such as integrins and the Fc receptor CD16 (18, 19). In this context, both macrophages and synovium lining cells express TLR-2 and TLR-4 (30, 46, 47). Although expression of both TLR-2 and TLR-4 is relatively low in normal synovium, TLR-2 and TLR-4 expression is subject to regulation, as illustrated by cytokine-inductive effects in vitro and up-regulated expression in RA synovium in vivo (46, 47). We speculate that TLR-2 and TLR-4 or closely interacting proteins are among the plasma membrane proteins directly engaged by MSU crystals. Articular chondrocytes have been particularly advantageous in testing this notion, because we observed that normal articular chondrocytes constitutively expressed TLR-2 but not several other TLRs, including TLR-4, in vitro (25). Moreover, direct up-regulation and down-regulation of MSU crystal–induced chondrocyte nitric oxide production was inducible by specific “gain-of-function” of TLR-2 expression versus “loss-of-function” of TLR-2–dependent signaling (25). It will be interesting to determine if TLRs other than TLR-2 and TLR-4 expressed by macrophages and synovium lining cells (48) also mediate MSU crystal–induced inflammation. We speculate that MSU crystals also mediate the activation of infiltrating neutrophils via TLRs to amplify synovitis once the acute gouty inflammatory process has been initiated.

The markedly reduced capacity of both TLR-2–deficient and TLR-4–deficient macrophages to ingest and respond to MSU crystals is compelling, because deficiencies of TLR-2 and TLR-4 are associated with selective rather than generalized phagocytosis defects (29), mediated partly by divergent modes of TLR-dependent and TLR-independent phagosome maturation (40). For example, in TLR-2−/− macrophages, phagocytosis of inert latex beads is intact (30, 40), as is ingestion of zymosan particles, although TLR-2−/− macrophages demonstrate decreased activation by zymosan (49, 50). Furthermore, macrophages from TLR-2/TLR-4 double-knockout mice and MyD88−/− mice demonstrate no differences in phagocytosis of apoptotic cells relative to WT cells under conditions in which phagocytosis of bacteria is impaired (40). Although our results suggest that TLR-2 and TLR-4 recognize the inert MSU crystal surface as a pathogen-associated molecular pattern to directly promote phagocytosis, the current study did not unequivocally prove this. One alternative scenario is that TLR-2 ligands promote phagocytosis through induction of a function-specific gene expression program that includes up-regulation of scavenger receptor A expression, an activity shared by ligands of certain other TLRs (29). Given this, it is possible that deficient TLR-2– and/or TLR-4–mediated induction of other plasma membrane proteins that recognize the inert MSU crystal may have contributed to our findings.

We limited our analysis of cultured cells to unfractionated bone marrow macrophage preparations. The state of differentiation of macrophages clearly mediates the ability of these cells to not only take up MSU crystals (14), but to also respond in a proinflammatory or antiinflammatory manner to the crystals (14, 15). Hence, the observed effects of host TLR-2 and TLR-4 expression on the capacity of macrophages to ingest and react to MSU crystals were possibly at least partially mediated by indirect effects on macrophage differentiation. Previously described effects of expression of certain TLRs on bacterial phagocytosis by macrophages mediated by regulation of signal transduction and gene expression (29) might provide an analogy to the situation for MSU crystal–macrophage interactions. However, further elucidation of direct and indirect effects of TLR-2 and TLR-4 on true “recognition” of the MSU crystal by the macrophage will clearly require specific TLR “loss-of-function” and “gain-of-function” studies in mature macrophages.

Striking suppression of MSU crystal uptake in TLR-2−/− and TLR-4−/− macrophages was mirrored in across-the-board suppression of the capacity of MSU crystals to induce expression by macrophages of proinflammatory cytokines and of TGFβ1 in the same cells in this study. Importantly, fully differentiated macrophages that clear MSU crystals express TGFβ1, thereby promoting resolution of acute MSU crystal–induced inflammation (4, 15). Therefore, our results suggest that under some conditions, TLR-2 and TLR-4 expression might not only promote the triggering of acute gout, but also contribute to the spontaneous self-limitation so characteristic of gouty inflammation (1, 4).

Limitations of this study include the fact that that we did not assess whether MSU crystals regulate inflammation indirectly through TLR and MyD88 signaling via induced cellular release of endogenous ligands of TLR-2 and TLR-4, such as Hsp70 and HMGB-1 (51, 52). MyD88 is one of several cytosolic adapter proteins for TLR-2 and TLR-4 (26), and MyD88 also transduces IL-1R–mediated responses (53). These are points to be considered because we did not test to determine whether the blunting effects of MyD88 deficiency on MSU crystal–induced macrophage activation and inflammation were mediated partly by impaired IL-1 signaling. Moreover, we have not yet evaluated the potential roles of the extracellular TLR-2 and/or TLR-4 adapter molecules CD14 and myeloid differentiation protein 2 (26, 39) in responsiveness to MSU crystals.

In conclusion, this study has established that host expression of TLR-2, TLR-4, and their intracellular adapter protein MyD88 is a major mediator of MSU crystal–induced inflammation. Our results suggest that acute gouty inflammation is triggered and regulated in intensity at least in part by cellular recognition of the naked MSU crystal as a function of TLR-dependent innate immunity. Significantly, TLR-2 signaling, culminating in NF-κB activation, critically transduces chondrocyte nitric oxide generation in response to MSU and calcium pyrophosphate dihydrate (CPPD) crystals (25). Since the acute inflammatory responses to CPPD crystals resemble those of MSU crystals, we speculate that TLR-2, TLR-4, and MyD88 could also mediate CPPD crystal–induced acute inflammation. Finally, TLR-2 and TLR-4 sequence variants have been linked to altered microbial carriage and phenotypic response of the host to infection (54, 55). Hence, it will be of interest to determine if inherited or acquired alterations in the structure and function of TLR-2 and TLR-4 contribute to variability in the clinical phenotype of gout in humans with hyperuricemia (1).

Acknowledgements

We gratefully acknowledge Drs. Peter Tobias and Richard Ulevitch (The Scripps Research Institute, La Jolla, CA) for helpful comments relating to the design and execution of these studies. We thank Monika Polewski (VA Medical Center, San Diego) for expert technical assistance with the air pouch model studies.

Ancillary