To determine whether infiltrating monocytes, neutrophils, or resident macrophages contribute to the early inflammatory response to monosodium urate monohydrate (MSU) crystals in vivo.
To determine whether infiltrating monocytes, neutrophils, or resident macrophages contribute to the early inflammatory response to monosodium urate monohydrate (MSU) crystals in vivo.
MSU crystal–induced inflammation was monitored using a peritoneal model of acute gout. The production of proinflammatory cytokines (interleukin-1β [IL-1β], tumor necrosis factor α [TNFα], IL-6) by resident macrophages, infiltrating monocytes, and neutrophils during the onset of gout was determined by flow cytometry. Infiltrating and resident peritoneal cells were cultured with MSU crystals ex vivo, and proinflammatory cytokine production was determined by multiplex cytokine array. Activated macrophages on the visceral epithelial lining of the peritoneum were identified by immunofluorescence histochemistry. The inflammatory immune response to MSU crystals was then compared with the inflammatory response in mice depleted of resident macrophages by pretreatment with clodronate liposomes.
The production of cytokines in vivo preceded the influx of Gr-1intermediate7/4+ monocytes. Monocytes and neutrophils recruited during the inflammatory phase of the response to MSU crystals failed to produce proinflammatory cytokines either in vivo, or ex vivo following restimulation with MSU crystals. Stimulation of the naive peritoneal resident cell population with MSU crystals ex vivo resulted in positive staining of resident macrophages for the proinflammatory cytokines IL-1β, TNFα, and IL-6. Depletion of the resident macrophage population resulted in a significant decrease in both MSU crystal–induced neutrophil infiltration and proinflammatory cytokine production in vivo despite the presence of infiltrating monocytes.
These data indicate that resident macrophages, rather than infiltrating monocytes or neutrophils, are important for initiating and driving the early proinflammatory phase of acute gout.
Gout is an inflammatory arthritis induced by the precipitation of monosodium urate monohydrate (MSU) crystals in articular joints and periarticular tissues, where it presents as a severe acute inflammation that spontaneously resolves after 7–10 days (1, 2). The early inflammatory phase of acute gout is associated with the production of proinflammatory cytokines (interleukin-6 [IL-6], tumor necrosis factor α [TNFα], IL-1β) and the infiltration of leukocytes, including neutrophils and monocytes (3–6).
Previous in vitro and ex vivo studies indicate that the differentiation state of mononuclear phagocytes plays a key role in the type of cytokines produced in response to MSU crystals and therefore in the initiation and progression of an attack of gout. In those studies, monocyte-like cells were reported to produce proinflammatory cytokines such as TNFα and IL-1β, whereas macrophage-like cells produced the antiinflammatory cytokine transforming growth factor β (TGFβ) (7–9). A similar pattern of proinflammatory to antiinflammatory cytokine production has been reported following differentiation of human CD14+ blood monocytes into macrophages (7). In addition, unpurified leukocyte infiltrates isolated from blisters during the course of an inflammatory response to cantharidin also exhibit cytokine “switching” following stimulation with MSU crystals in vitro (7). Together these data have led to the hypothesis that in an acute gout attack infiltrating monocytes drive the inflammatory response, while macrophages with a differentiated phenotype mediate the resolution of inflammation. As a result, previous gout research has mostly ignored the potential of resident macrophages to induce inflammation.
Recent literature now indicates a possible proinflammatory role of macrophages in gouty inflammation. Macrophages treated with MSU crystals have been reported to induce proinflammatory cytokines such as IL-1β (10, 11), whereby signaling through the IL-1 receptor is required for activation of nonhemopoietic cells to induce neutrophilia (10), a key characteristic of gouty inflammation (12). Other proinflammatory activities such as the production of TNFα, monocyte chemotactic protein 1, IL-18, and inducible nitric oxide synthase and the up-regulation of triggering receptor expressed on myeloid cells 1 (TREM-1) on macrophages provide evidence of the potential involvement of this cell type in gout (10, 11, 13–15). This would be consistent with findings in other acute inflammatory conditions that rely heavily on macrophages in the onset of inflammation (16, 17).
Both macrophages and monocytes appear to produce the key proinflammatory cytokines IL-1β, IL-6, and TNFα following in vitro stimulation with MSU crystals; however, the relevance of these findings with respect to the initiation and progression of inflammation in vivo has not been established. Therefore, studies that distinguish between the different phenotypes of mononuclear phagocyte populations over the course of the inflammatory response in vivo are needed to provide a more accurate depiction of the function of these cells in the context of gouty arthritis.
Although neutrophil infiltration is a hallmark feature of gout, little is known about the ability of these cells to subsequently respond to MSU crystals at the site of inflammation and produce IL-1β, IL-6, or TNFα during the early phase of inflammation in vivo. Using a murine peritoneal model of acute gout, we have investigated the production of the gout-associated proinflammatory cytokines IL-1β, IL-6, and TNFα by infiltrating neutrophils, monocytes, and resident macrophages during the initiation and early phase of MSU crystal–induced inflammation. Our results showed that resident macrophages, rather than infiltrating monocytes or neutrophils, were primarily responsible for the production of the proinflammatory cytokines IL-1β and IL-6 and the infiltration of neutrophils in vivo during the onset of MSU crystal–induced inflammation. Contrary to the current dogma, infiltrating monocytes did not produce proinflammatory cytokines in response to MSU crystals in the initial stages of acute inflammation.
Male C57BL/6 mice were bred and housed in a conventional animal facility at the Malaghan Institute of Medical Research, Wellington, New Zealand. All animals used for the experiments were age 8–10 weeks. All experimental procedures were approved by the Victoria University Animal Ethics Committee in accordance with their guidelines for the care of animals.
Uric acid, lipopolysaccharide (LPS), and saponin were obtained from Sigma (Auckland, New Zealand). Cytokine Bead Array kits, chamber slides, GolgiStop, and phycoerythrin-conjugated anti-mouse IL-6 and allophycocyanin-conjugated anti–Gr-1 monoclonal antibodies were obtained from BD Biosciences (North Ryde, New South Wales, Australia). Anti-mouse IL-1β, biotin-conjugated anti-mouse IL-1β, anti-mouse TNFα, anti-mouse IL-6, and biotin-conjugated mouse anti-rat IgG monoclonal antibodies were obtained from eBioscience (San Diego, CA). Anti-F4/80 and fluorescein isothiocyanate (FITC)–conjugated 7/4 antibodies were obtained from Serotec (Oxford, UK). Bio-Plex multiplex arrays were purchased from Bio-Rad (Hercules, CA). The mouse IL-1β enzyme-linked immunosorbent assay (ELISA) kit was obtained from R&D Systems (Minneapolis, MN). The Limulus amebocyte cell lysate assay kit was obtained from Associates of Cape Cod (East Falmouth, MA). Heparin was obtained from Mayne Pharma (Melbourne, Victoria, Australia). Low cell binding plates were purchased from Nunc (Rochester, NY). Diff-Quik was obtained from Dade Behring Diagnostics (Newark, DE). AnalySIS Life Science extended focal imaging software was obtained from Olympus (Auckland, New Zealand). All other products were obtained from Invitrogen (Auckland, New Zealand) unless otherwise stated.
MSU crystals were prepared by crystallization of a supersaturated solution of uric acid under mildly basic conditions. Briefly, 250 mg uric acid was added to 45 ml of double-distilled water containing 300 μl of 5M NaOH, and the solution was boiled until the uric acid was dissolved. The solution was passed through a 0.2-μM filter, and 1 ml of 5M NaCl was added to the hot solution, which was then stored at 26°C. After 7 days the resulting MSU crystals were washed with ethanol and acetone. The resulting triclinic, needle-shaped MSU crystals were 5–25 μm in length and were birefringent to polarized light. All MSU crystals were determined to be endotoxin free (<0.01 EU/10 mg) by Limulus amebocyte cell lysate assay.
C57BL/6 mice were administered an intraperitoneal (IP) injection of a 3-mg slurry of MSU crystals in 0.5 ml phosphate buffered saline (PBS). At different time points mice were euthanized by CO2 administration, and the peritoneal exudate cells were harvested by lavage with 3 ml PBS containing 25 units/ml heparin and 10% fetal bovine serum (FBS). Cells were retrieved from the lavage fluid and analyzed by flow cytometry and by Diff-Quik staining of cytospin samples. Lavage fluid was retained for cytokine assay.
C57BL/6 mice were administered an IP injection of 4% thioglycolate (0.5 ml). After 4 hours the mice were euthanized by CO2 administration, and the peritoneal exudate cells were harvested as described above for MSU crystal–induced peritonitis.
Mice were administered an IP injection of MSU crystals (3 mg in 0.5 ml PBS). After 4 hours the mice were euthanized by CO2 administration, and the peritoneal exudate cells were harvested by lavage with 3 ml PBS containing 25 units/ml heparin. Cells were retrieved from the lavage fluid and stained for F4/80, Gr-1, and 7/4. Neutrophils (F4/80–Gr-1high7/4+) were isolated by fluorescence-activated cell sorting (FACS) using a BD FACSVantage Diva (Becton Dickinson, San Jose, CA). Neutrophils were determined to be >99% pure by flow cytometry and Diff-Quik analysis of cytocentrifuged samples. Neutrophil viability was confirmed to be >99% by trypan blue exclusion.
Harvested peritoneal exudate cells were cultured in RPMI 1640 and 10% FBS at 1 × 106 cells/ml in 96-well plates and treated with PBS, 200 μg/ml MSU crystals, or 100 ng/ml LPS for 16 hours at 37°C. Supernatants were then harvested for cytokine analysis.
Cytokine levels in harvested lavage PBS and from culture supernatants were assayed by Cytokine Bead Array and analyzed on a FACSCalibur flow cytometer (BD Biosciences), by Bio-Plex multiplex array and analyzed on a Bio-Plex flow cytometer, or by ELISA.
Total peritoneal exudate cells from mice treated with MSU crystals for 4 hours were harvested by peritoneal lavage and quickly suspended at 106 cells/ml in RPMI 1640 containing 10% FBS, PenStrep, Glutamax, and 1:1,500 GolgiStop and transferred into 24-well low cell binding plates. Positive controls were stimulated with 1 μg/ml LPS, after which all cells were incubated for a further 4 hours at 37°C to allow for intracellular accumulation of cytokines. Cells were permeabilized with 0.1% saponin and stained for intracellular IL-6. The cells were washed and resuspended in FACS buffer, then stained for the surface markers F4/80, Gr-1, and 7/4. Cells were analyzed by flow cytometry using a FACSCalibur flow cytometer. Cytospin samples of exudate cells were also prepared by Diff-Quik staining to complement differential cell counts determined by flow cytometry.
Resident peritoneal cells from naive mice were suspended at 106 cells/ml in RPMI 1640 containing 10% FBS, PenStrep, and Glutamax and placed into 8-well chamber slides. After incubation at 37°C for 1 hour, 1:1,500 GolgiStop was added. Cells were stimulated with 200 μg/ml MSU crystals, 1 μg/ml LPS, or PBS and incubated for 4 hours, then treated with zinc fixative for 30 minutes. Endogenous biotin was blocked using a biotin blocking kit (Invitrogen) in accordance with the manufacturer's instructions. Nonspecific IgG binding sites were then blocked using 5% mouse serum or 5% FBS, followed by incubation overnight with anti-mouse IL-6, anti-mouse TNFα, biotinylated anti-mouse IL-1β, or the appropriate isotype control. Cells treated with anti-mouse IL-6 or TNFα were then incubated with a biotinylated mouse anti-rat IgG antibody. Excess antibody was removed by washing with PBS, and the samples were stained with Alexa Fluor 555–conjugated streptavidin, FITC-conjugated anti-F4/80, and Hoechst 55542, mounted in Vectashield antifade (Vector, Burlingame, CA), and examined using a BX51 fluorescence microscope (Olympus, Tokyo, Japan).
Mice were treated with an IP injection of 3 mg MSU crystals as described above. After 4 hours the skin was removed from the abdomen, the peritoneum was cleared of leukocytes by lavage with 3 ml PBS, and the ventral tissue covering the peritoneal cavity was excised and placed into cold RPMI 1640. The tissue was rinsed in PBS and fixed in acetone for 10 minutes at −20°C. The visceral lining of the peritoneum was harvested from the fixed tissue by carefully stripping the lining off as a single piece. The harvested visceral tissue was cleaned of fascia, and nonspecific IgG binding sites were blocked with a solution containing 10% FBS. Endogenous biotin was blocked using a kit. The tissue was stained for the surface markers 7/4 and F4/80 and mounted onto slides with Vectashield antifade. Slides were analyzed by fluorescence microscopy, and images were processed using AnalySIS Life Science extended focal imaging software.
Clodronate liposomes were prepared as previously described (18). Mice were treated by IP injection of 200 μl of either clodronate liposomes or PBS. Three days later, macrophage depletion in liposome-treated mice was confirmed by flow cytometry, following which mice were immediately challenged with an IP injection of 3 mg of MSU crystals.
To identify the cell populations in the peritoneal lavage fluid of naive and MSU crystal–treated mice over time, cells were stained for the macrophage differentiation marker F4/80 and the myeloid differentiation antigens Gr-1 and 7/4 (Figure 1). Neutrophils were identified as Gr-1high7/4+, and monocytes were identified as Gr-1intermediate7/4+ cells expressing low levels of F4/80. Resident macrophages were identified as F4/80high cells staining negative for Gr-1.
To establish the acute inflammatory profile of peritoneal inflammation in response to MSU crystals, we monitored leukocyte infiltration and the production of the proinflammatory cytokines IL-1β, IL-6, and TNFα over 72 hours. Infiltration of both monocytes and neutrophils into the peritoneum was observed after 4 hours and peaked at 16 hours after injection of MSU crystals (Figure 2A). Neutrophil infiltration was notably faster and greater than that of monocytes. As shown in Figures 2B–D, IL-6, IL-1β, and TNFα levels were elevated in the peritoneum of MSU crystal–treated mice within 2 hours of administration of MSU crystals, peaking at 4 hours for all 3 cytokines. The inflammatory response was self limiting, with leukocyte cell numbers and cytokine levels returning to normal within 48 hours. Based on peak cytokine production combined with early leukocyte infiltration, we chose to focus on the first 4–8 hours to determine which cell types (neutrophils, monocytes, and/or macrophages), were driving the early inflammatory response to MSU crystals.
To investigate the potential contribution of infiltrating monocytes and neutrophils to early inflammation following exposure to MSU crystals, we collected peritoneal lavage fluid from mice 4 hours after treatment with MSU crystals and examined the capacity of infiltrating Gr-1intermediate7/4+ monocytes and Gr-1high7/4+ neutrophils to produce the proinflammatory cytokine IL-6 4 hours after MSU crystal treatment (Figure 3A). Surprisingly, the infiltrating monocyte population did not produce IL-6 in response to MSU crystals, although monocytes were still able to produce inflammatory cytokines in response to LPS stimulation. Infiltrating neutrophils did not produce IL-6 in response to either MSU crystals or LPS (Figure 3A). In fact, none of the cells isolated from the peritoneal lavage fluid appeared to be producing large amounts of IL-6 (Figure 3A).
To confirm that the cells in the peritoneal exudate from MSU crystal–treated mice were unable to produce proinflammatory cytokines in response to MSU crystals, they were restimulated with MSU crystals ex vivo, and the supernatants were analyzed for the presence of IL-6, IL-1β, and TNFα. None of these cytokines was elevated following MSU crystal treatment (Figure 3B), indicating that infiltrating cells were not the source of cytokine production illustrated in Figure 2B. Serum components in cell culture have been shown to inhibit MSU crystal–induced neutrophil superoxide production (19) and could therefore affect neutrophil cytokine production. However, isolated neutrophils restimulated with MSU crystals also failed to produce IL-6, TNFα, or IL-1β under low serum conditions (data available online at http://www.malaghan.org.nz/research/arthritis/arthritis-publications/), confirming that the lack of cytokine production was not serum dependent.
Monocytes recruited following IP administration of thioglycolate also failed to produce significant levels of inflammatory cytokines in response to restimulation with MSU crystals (data available online at http://www.malaghan.org.nz/research/arthritis/arthritis-publications/), indicating that the lack of inflammatory response to MSU crystals may be a common feature of newly recruited monocytes in acute inflammation. However, peritoneal cells from naive mice did produce significant amounts of inflammatory cytokines following exposure to MSU crystals ex vivo (Figure 3C), indicating that one or more of the resident cell populations was responsible for cytokine production following administration of MSU crystals.
The absence of cytokine-producing cells from the peritoneal lavage fluid of MSU crystal–treated mice led us to investigate changes in total cell numbers during the course of the inflammatory response to MSU crystals. We observed an initial drop in the total number of peritoneal cells within the first 2 hours following administration of MSU crystals (Figure 4A) and before the infiltration of neutrophils and monocytes (Figure 2A). Between 60% and 70% of cells in the peritoneum of naive mice are resident macrophages; therefore, any significant drop in cell number is likely to be associated with a decrease in this population. Flow cytometric analysis of the F4/80highGr-1–7/4− resident macrophage population showed complete disappearance of this population from the peritoneal lavage fluid after 2 hours (Figure 4B).
Cell adhesion is a common feature of macrophage activation, and the absence of macrophages shortly after administration of inflammatory stimuli has been reported as the “macrophage disappearance reaction” in other models of acute inflammation (20–22). To confirm that macrophage adherence was occurring in vivo following administration of MSU crystals, we isolated visceral peritoneal membranes from mice 4 hours after MSU crystal administration and stained for the presence of F4/80high7/4− resident macrophages, using immunofluorescence. Following MSU crystal treatment, we observed on the peritoneal membrane adherent resident macrophages that were not present on tissue from untreated mice (Figure 5A), indicating that macrophage activation and adherence had occurred in response to treatment with MSU crystals. In vitro cultures of the naive peritoneal cell population identified the F4/80+ resident macrophages as producing not only IL-6 but also TNFα and IL-1β in response to exposure to MSU crystals (Figure 5B), indicating that resident macrophages were the likely source of inflammatory cytokines following MSU crystal treatment in vivo.
Next, we sought to determine the importance of MSU crystal–induced activation of resident macrophages to the inflammatory response in vivo, by depleting the resident macrophage population in the peritoneum prior to MSU crystal administration. Resident peritoneal macrophages were depleted using clodronate-loaded liposomes, with liposome pretreatment causing 80% depletion of F4/80+ cells (Figure 6A and data not shown). Following MSU crystal treatment, monocyte infiltration was the same in both macrophage-depleted mice and nondepleted mice (Figure 6A). However, in macrophage-depleted mice there was a significant reduction in neutrophil infiltration (Figure 6B). The production of IL-6 and IL-1β was also significantly reduced in macrophage-depleted mice, although TNFα levels remained the same (Figure 6C). These results confirmed that resident macrophages play an essential role in the production of the key proinflammatory cytokines IL-6 and IL-1β and in neutrophil recruitment in MSU crystal–induced inflammation.
Our results show that resident tissue macrophages, which have a highly differentiated phenotype, play a significant role in triggering acute MSU crystal–induced inflammation, whereby resident macrophages are a key source of proinflammatory cytokines including IL-1β and IL-6. In addition, the resident macrophage population appears to play an essential role in the initiation of neutrophil infiltration, as illustrated by the significant reduction in neutrophil infiltration following in vivo depletion of this population.
Once recruited to the site of inflammation, neither neutrophils nor monocytes appear to contribute to the proinflammatory cytokine production observed during the early stages of the acute inflammatory response to MSU crystals. The monocyte results were surprising in light of previous studies showing MSU crystal–induced proinflammatory cytokine production by monocytes treated with MSU crystals in vitro and by leukocyte infiltrates from cantharidin-induced blisters treated with MSU crystals ex vivo (7). However, in addition to using a different inflammatory stimulus for recruitment of cells, the cantharidin study investigated later-stage cellular infiltrates (16 hours compared with 4–8 hours) that likely differ in cellular composition and phenotype compared with the initiation phase of MSU crystal–induced inflammation. In fact, phenotypic differences between in vivo MSU crystal–recruited monocytes and isolated monocytes and monocytic cell lines may also explain why our results differ from earlier in vitro data.
The lack of proinflammatory response to MSU crystals does not arise as a result of an inability to produce proinflammatory cytokines, since MSU crystal–elicited monocytes readily produce IL-6 in response to LPS. The difference in responsiveness of infiltrating monocytes to MSU crystals and LPS suggests that the proinflammatory response to MSU crystals, unlike that to LPS, does not occur via Toll-like receptor 4 (TLR-4). This is consistent with earlier work showing that knocking out the gene encoding TLR-4 does not ablate the inflammatory response to MSU crystals (10). In addition, thioglycolate-elicited cells are also unresponsive to MSU crystal restimulation, showing that desensitization to MSU is unlikely to be the reason for the absence of cytokine production. In fact, the lack of the proinflammatory response to MSU crystal restimulation may be a common phenotype of early infiltrating monocytes regardless of how they are elicited.
Our study pinpoints the resident macrophage population as a primary source of the proinflammatory cytokines IL-1β and IL-6 in the early response to MSU crystal exposure. Although the resident macrophages produce TNFα in response to exposure to MSU crystals, they do not appear to be an essential source of TNFα in vivo. Our findings are consistent with a key involvement of resident macrophages in the initiation of other forms of acute inflammation (16, 17, 23).
Although earlier in vitro studies have indicated that differentiated macrophages produce “antiinflammatory” cytokines such as TGFβ following exposure to MSU crystals, more recent reports have shown that MSU crystals can activate differentiated macrophages, leading to the activation of the NALP3 inflammasome, the production of IL-1β, the up-regulation of TREM-1, and the production of cytokine-induced neutrophil chemoattractant (also called KC or CXCL1) (10, 11, 13). These later findings support our in vivo results showing that the resident macrophage population initiates inflammation in response to MSU crystals. In our model it appears that MSU crystal–activated resident macrophages adhere to the surrounding tissues, produce proinflammatory cytokines, and initiate recruitment of neutrophils from the blood during the early inflammatory response.
The proinflammatory response of the resident macrophage population to MSU crystals does not preclude differentiated macrophages from playing a role in the resolution of inflammation at a later point in time. The plasticity of the macrophage phenotype is well known, and macrophage switching from a “proinflammatory” to an “antiinflammatory” phenotype over time remains a viable mechanism of action for resolution of acute inflammation. However, TGFβ-producing macrophages have yet to be identified in vivo.
In part, the lack of proinflammatory cytokine production by infiltrating monocytes in response to MSU crystals could contribute to the self-limiting nature of the disease. Classically, infiltrating monocytes would augment and further prolong or amplify an inflammatory response by producing increased amounts of proinflammatory cytokines. Since recruited monocytes did not exhibit a proinflammatory response to MSU crystals in our model, it is also possible that augmentation of the early inflammatory response to MSU crystals by infiltrating monocytes does not occur in gout.
Mononuclear phagocytes are highly diverse in phenotype and function, making them difficult to model faithfully in vitro. Although care needs to be taken when extrapolating findings from the peritoneal model of inflammation to joint inflammation in gout, the unmanipulated peritoneal environment allows for the complex interplay between resident and infiltrating immune cells, which is not provided by in vitro studies. Therefore, the in vivo model provides greater insight into the true effector phenotypes in the context of the inflammation that occurs in acute gout. To this end, this approach has shown that infiltrating monocytes elicited by MSU crystals in vivo behave differently from monocytes derived by other methods.
The immune response to MSU crystals is multifaceted in nature, making it difficult to account for all aspects of the response in one study. Although macrophages play a pivotal role, they obviously do not control the entire inflammatory response to MSU crystals. For example, macrophage activation and neutrophil infiltration do not appear to be a requirement for monocyte infiltration in our model, highlighting another aspect of the inflammatory response to MSU crystals that warrants further investigation. Taken together, our findings bring us a step closer to clearly identifying which immune cells are likely to be critically involved in different aspects of early inflammation in gout.
In summary, our work identifies resident macrophages, not infiltrating monocytes or neutrophils, as triggering and driving early MSU crystal–induced inflammation in the context of IL-1β and IL-6 production and the initiation of neutrophil infiltration. Based on these findings, macrophages are also likely to be key cells involved in initiating and driving inflammation in gouty arthritis, suggesting the need to revisit the current understanding of induction of acute gout.
Dr. Harper 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 design. Martin, Harper.
Acquisition of data. Martin, Walton.
Analysis and interpretation of data. Martin, Walton, Harper.
Manuscript preparation. Martin, Harper.
Statistical analysis. Martin.