To identify macrophage-independent sources of transforming growth factor β1 (TGFβ1) production during monosodium urate monohydrate (MSU) crystal–induced inflammation and to determine how TGFβ1 alters MSU crystal–recruited neutrophil functions.
C57BL/6J mice were injected intraperitoneally with MSU crystals with or without TGFβ1-neutralizing antibody. MSU crystal–recruited peritoneal and blood neutrophils were purified and cultured ex vivo. Peritoneal neutrophils were treated with the caspase inhibitor Q-VD-OPh, anti-TGFβ1 antibody, or fluorochrome-labeled apoptotic neutrophils. Neutrophils were analyzed for expression of annexin V, caspase 3, and TGFβ1 by flow cytometry or fluorescence microscopy, for superoxide production using the redox-sensitive dye water-soluble tetrazolium 1, and for TGFβ1 and interleukin-1β (IL-1β) production by enzyme-linked immunosorbent assay.
Eighteen hours after MSU crystal administration in vivo, TGFβ1 levels were elevated in peritoneal lavage fluids, and a significant number of peritoneal neutrophils were TGFβ1+. Purified blood or peritoneal neutrophils cultured ex vivo showed TGFβ1+ neutrophils coexpressing the apoptosis marker caspase 3 and increased TGFβ1 production, both of which dropped following inhibition of apoptosis. Live neutrophils that had phagocytosed apoptotic neutrophils showed greatest TGFβ1 expression. Superoxide production by purified MSU crystal–recruited neutrophils ex vivo was enhanced by anti-TGFβ1 antibody treatment. Neutrophils purified from the peritoneum of MSU crystal–challenged mice treated with anti-TGFβ1 antibody produced elevated levels of superoxide, but neutrophil IL-1β production was unaffected.
Neutrophil cannibalism and TGFβ1 production have the potential to make a significant contribution to the controlled resolution of neutrophil-driven inflammatory diseases such as gout.
An attack of gout is triggered by the crystallization of monosodium urate monohydrate (MSU) in the joints and connective tissues and is characterized by rapid inflammatory cell infiltration, localized swelling, reddening of the skin, and debilitating pain. Interleukin-1β (IL-1β) is the pivotal cytokine involved in the autoinflammatory cascade in response to MSU crystals. Produced initially by resident cells, there is evidence that infiltrating neutrophils also have the capacity to produce IL-1β and could therefore augment local IL-1β levels to drive inflammation.
Stimulation of neutrophils also triggers the production of toxic reactive oxygen species (ROS) via activation of NADPH oxidase (1), where their primary function is to kill pathogens (2). Although beneficial to host immune defenses, this respiratory burst must be tightly regulated, since inappropriate ROS production, such as that associated with MSU crystal–induced autoinflammation in gout, can be cytotoxic, resulting in cell and tissue damage (3).
In patients having a gout attack, spontaneous resolution of inflammation is associated with elevated levels of transforming growth factor β1 (TGFβ1) in the synovial fluid of the affected joint (4). The exogenous administration of TGFβ1 has also been shown to significantly attenuate MSU crystal–induced cellular recruitment in vivo (5). Furthermore, TGFβ1 gene polymorphisms in humans have been linked with longer inflammatory attacks and increased tophus formation in the joints (6). These findings identify TGFβ1, a key mediator commonly involved in resolution and tissue repair (7, 8), as an important cytokine in MSU crystal–induced inflammation and resolution.
The efficient clearance of apoptotic cells is an important step in the resolution of inflammation (9), including MSU crystal–induced inflammation (10). As neutrophils undergo spontaneous apoptosis they are phagocytosed by macrophages, a mechanism that triggers macrophage TGFβ1 production (7). However, little is known about alternative sources of TGFβ1 during the resolution phase of MSU crystal–induced inflammation and how this affects respiratory burst and IL-1β production in neutrophils.
The aim of this study was to identify macrophage-independent sources of TGFβ1 during the resolution phase of acute inflammation and to determine the regulatory effect of TGFβ1 on ROS and IL-1β production by MSU crystal–recruited neutrophils.
MATERIALS AND METHODS
C57BL/6J male 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 ages 8–10 weeks. The experiments were approved by the Victoria University Animal Ethics Committee and carried out in accordance with the Committee's guidelines for the care of animals.
Mouse anti-TGFβ1 monoclonal antibody (clone 2Ar2) used in ex vivo assays was from Abcam. The general caspase inhibitor Q-VD-OPh and the mouse TGFβ1 and IL-1β enzyme-linked immunosorbent assay (ELISA) kits were from R&D Systems. Multiplex bead array was obtained from BioRad. Anti-mouse lymphocyte antigen 6G (Ly-6G), anti-mouse Gr-1, and anti-mouse CD11b monoclonal antibodies and fluorescein isothiocyanate (FITC)–annexin V, phycoerythrin–rabbit anti–active caspase 3, GolgiStop, and tissue plasticware were obtained from BD Biosciences. Low cell binding 96-well plates were from Nunc. Red blood cell (RBC) lysis solution was obtained from Qiagen Science. Alexa Fluor 555–streptavidin, Alexa Fluor 488–goat anti-mouse IgG1 (γ1 chain), and cell culture reagents including media, phosphate buffered saline (PBS), penicillin–streptomycin, Glutamax, bovine serum albumin, and ProLong Gold anti-fade with DAPI were obtained from Invitrogen. Lympholyte-M cell separation centrifugation medium was obtained from Cedarlane. The TGFβ1-neutralizing antibody used in vivo was from R&D Systems.
Preparation of MSU crystals.
MSU crystals were prepared and characterized as previously described (11). Briefly, 250 mg uric acid was boiled in 45 ml of 30 mM NaOH/double-distilled water until completely dissolved. The uric acid solution was filtered sterile (0.2-μm filter) and stored at 26°C for 7 days to allow crystal formation. The resulting MSU crystals were washed and air-dried under sterile conditions. The MSU crystals were needle shaped (5–20 μm in length) and showed optical birefringence. The MSU crystals used were endotoxin free as determined by Limulus amebocyte cell lysate assay (<0.01 endotoxin units/10 mg).
In vivo model of MSU crystal–induced inflammation.
Mice were injected intraperitoneally with 3 mg MSU crystals suspended in 0.5 ml PBS. At different time points mice were euthanized by CO2 asphyxiation, blood was collected by cardiac puncture, and the peritoneal cavity was lavaged with 3 ml PBS containing 25 units/ml heparin. For TGFβ1 neutralization in vivo, mice were treated with 1 mg of the anti-TGFβ1 antibody 30 minutes prior to MSU crystal administration.
Peritoneal exudate cells from MSU crystal–treated mice were washed twice with PBS, and neutrophils were purified by sedimentation using a Lympholyte-M solution in accordance with the manufacturer's instructions. Blood was collected from MSU crystal–treated mice, and RBCs were lysed using RBC lysis solution. Blood neutrophils were purified using a density gradient (Lympholyte-M solution) in accordance with the manufacturer's instructions. Neutrophil purity was 85–90% as determined by flow cytometry (CD11b+Ly-6G+Gr-1high). Purified neutrophils (1 × 106/ml in serum-free RPMI 1640) were then plated onto 96-well plates (200 ml per well) for subsequent ex vivo assays.
Ex vivo neutrophil assays.
Purified blood and peritoneal neutrophils were incubated ex vivo for up to 24 hours in the presence or absence of MSU crystal restimulation (200 μg/ml). Neutrophils were harvested at different time points and analyzed by flow cytometry for the expression of the surface markers CD11b, Ly-6G, Gr-1, annexin V, and intracellular TGFβ1 and caspase 3. Supernatants were collected at different time points, and cytokine levels were determined by ELISA or multiplex bead array in accordance with the manufacturers' instructions. In assays investigating intracellular expression of TGFβ1, neutrophils were cultured with GolgiStop (1:1,000) to block cytokine secretion. To investigate the effect of apoptosis and TGFβ1 on neutrophil function, neutrophil cultures were treated either with the general caspase inhibitor Q-VD-OPh (20 μM) or with the anti-TGFβ1 antibody (35 μg/ml).
Neutrophil phagocytosis assay.
MSU crystal–recruited peritoneal neutrophils were isolated and purified as above. The cells were labeled with PerCP–Ly-6G, then incubated ex vivo (18 hours in RPMI 1640) to induce apoptosis. The apoptotic neutrophils (PerCP–Ly-6G positive) were then coincubated in RPMI 1640 with freshly isolated MSU crystal–recruited live peritoneal neutrophils labeled with allophycocyanin (APC)–Ly-6G (apoptotic cell:live cell ratio of 5:1) in the presence of GolgiStop (1:1,000). After 8 hours the cells were harvested and analyzed by flow cytometry for the expression of APC–Ly-6G, PerCP–Ly-6G, and intracellular TGFβ1. Live neutrophils that had phagocytosed apoptotic neutrophils were identified as APC–Ly-6G–positive/PerCP–Ly-6G–positive cells.
Neutrophil superoxide assay.
Neutrophil superoxide production was measured using the colorimetric dye water-soluble tetrazolium 1 (WST-1) as previously described (12). Purified MSU crystal–recruited neutrophils (1 × 106/ml) were plated out in phenol red–free RPMI 1640 containing WST-1 (250 mg/ml) and incubated with anti-TGFβ1 antibody (35 μg/ml) or MSU crystals (200 μg/ml) or with a combination of both agents at 37°C for 1 hour. The absorbance at wavelength 450 nm was then measured using a Versamax spectrophotometer (Molecular Devices).
For fluorescence staining of cell surface marker expression, cells were washed and resuspended in wash buffer (0.1% bovine serum albumin [BSA], 0.01% NaN3 in PBS, pH 7.4), stained with fluorochrome-conjugated antibodies to the surface markers Gr-1, Ly-6G, and CD11b, and then washed and resuspended in wash buffer. For intracellular caspase 3 and TGFβ1 expression, cells were permeabilized using saponin buffer (PBS, 0.1% saponin azide, 1% fetal bovine serum [FBS], 0.1% NaN3).
To measure expression of the apoptosis surface marker, cells were washed in annexin V binding buffer (PBS, 10 mM HEPES, 140 mM NaCl, 2.5 mM CaCl2, pH 7.4) and stained with FITC–annexin V. Cells were then analyzed by flow cytometry using a FACSCalibur (Becton Dickinson).
Purified neutrophils were spun onto microscope slides (500 revolutions per minute for 6 minutes). Cells were washed and then fixed with ice-cold methanol:acetone (50:50) solution for 10 minutes. Nonspecific binding was blocked by incubation with block buffer (PBS, 0.1% saponin, 0.1% BSA, 0.01% NaN3, 10% FBS) for 10 minutes at 37°C. Cells were then washed and stained with the surface marker anti-mouse Ly-6G at 37°C for 2 hours. Following 3 wash steps, the Alexa Fluor 555–streptavidin was added, and cells were incubated for 1 hour at 4°C. Saponin buffer was used to permeabilize the cells, which were then stained for intracellular TGFβ1 using a mouse anti-TGFβ1 monoclonal antibody for 2 hours at 37°C. Cells were washed with saponin buffer and treated for 1 hour with the secondary antibody Alexa Fluor 488–goat anti-mouse IgG1. Excess antibody was removed by washing with PBS, and cells were fixed with 4% formalin for 10 minutes at 22°C. One drop of ProLong Gold anti-fade with DAPI was applied, and the slides were mounted with a cover slip. Slides were examined using an Olympus BX51 fluorescence microscope.
Statistical analysis was carried out using Student's paired 2-tailed t-test and a one-way or two-way analysis of variance with Bonferroni correction for multiple comparisons using GraphPad Prism software, version 5.0. Every condition was compared with the control condition. P values less than 0.05 were considered significant.
Neutrophils as a source of TGFβ1.
Elevated levels of TGFβ1 have been found in the synovial fluid from patients having gout attacks (4, 13), linking the production of TGFβ1 to the resolution phase of gouty arthritis (14). In a previous study our group showed that MSU crystal–induced proinflammatory cytokine levels return to background levels 18 hours after the initiation of acute inflammation in mice (11), indicating a switch toward resolution. To confirm that TGFβ1 levels were also elevated 18 hours after MSU crystal administration in vivo, peritoneal lavage fluid was collected from MSU crystal–challenged mice 18 hours after crystal administration and analyzed for TGFβ1 by ELISA. Consistent with the clinical inflammatory profile, the TGFβ1 levels in the lavage fluid from MSU crystal–treated mice were higher than those in the lavage fluid from naive mice (Figure 1A).
Neutrophils are present in large numbers in the peritoneum 18 hours following MSU crystal–induced inflammation (11). To determine whether MSU crystal–recruited neutrophils could be contributing to the observed TGFβ1 production in vivo, neutrophils in the blood and peritoneum of MSU crystal–treated mice were stained for the neutrophil surface marker Ly-6G and intracellular TGFβ1 and then analyzed by flow cytometry and fluorescence microscopy. TGFβ1+ neutrophils were identified in both the freshly isolated blood and peritoneal lavage as well as when neutrophils were cultured ex vivo (Figures 1B–D). However, quantitative analysis by flow cytometry showed that the percentage of TGFβ1+ neutrophils was greater in the peritoneum compared to the blood and increased again following ex vivo culture (Figure 1E). These data identified activated neutrophils as a potential source of local TGFβ1 production in vivo.
TGFβ1 production by ex vivo–cultured peritoneal neutrophils.
To determine whether TGFβ1+ neutrophils were secreting TGFβ1, supernatants from ex vivo cultures of MSU crystal–induced peritoneal neutrophils were analyzed by ELISA. Consistent with the observed increase in TGFβ1+ cells ex vivo, cultured neutrophils secreted significantly higher levels of TGFβ1 compared to freshly isolated peritoneal neutrophils (Figure 2A).
Neutrophils are known to have a very short lifespan (15, 16); therefore, freshly isolated and ex vivo–cultured peritoneal neutrophils from MSU crystal–treated mice were analyzed by flow cytometry for expression of the apoptosis marker annexin V. As shown in Figure 2A, there was a significant increase in the percentage of apoptotic peritoneal neutrophils following ex vivo culture compared to that in uncultured peritoneal cells.
To investigate a possible link between neutrophil TGFβ1 production and apoptosis, purified neutrophils from MSU crystal–treated mice were cultured ex vivo. At different time points, the cells were analyzed by flow cytometry to identify apoptotic and TGFβ1+ cells, and the culture supernatants were analyzed by ELISA to measure TGFβ1 release. As shown in Figure 2B, cultured neutrophils exhibited an increase in both the percentage of TGFβ1+ neutrophils and the levels of secreted TGFβ1 over time. Neutrophil TGFβ1 generation was also associated with an up-regulation of the apoptosis markers annexin V and caspase 3 (Figure 2C), indicating that increased TGFβ1 production was associated with increased neutrophil apoptosis.
TGFβ1 production by neutrophils requires apoptosis.
To further examine the phenomenon of TGFβ1 production by neutrophils, MSU crystal–recruited peritoneal neutrophils and naive circulating blood neutrophils were cultured ex vivo, and the percentage of TGFβ1+caspase 3+ cells was determined. Analysis of MSU crystal–recruited neutrophils showed that only caspase 3+ neutrophils expressed TGFβ1 (Figure 3A). A similar pattern of increased neutrophil TGFβ1 expression and secretion was observed for cultured apoptotic blood neutrophils (Figures 3B and C), indicating that apoptosis-driven neutrophil TGFβ1 production was not dependent on MSU crystal activation in vivo.
The process of neutrophil apoptosis can be regulated via a number of caspases, including caspases 1, 3, 4, and 8 (17–19). To confirm the relationship between apoptosis and TGFβ1, MSU crystal–recruited peritoneal neutrophils were cultured ex vivo in the presence of the general caspase inhibitor Q-VD-OPh (Figure 4A). The blockade of neutrophil apoptosis resulted in fewer TGFβ1+ neutrophils and corresponded to a decrease in the levels of secreted TGFβ1 (Figure 4B), providing evidence that neutrophil apoptosis was required for neutrophil TGFβ1 production.
Neutrophil TGFβ1 production is triggered by phagocytosis.
Studies show that the process of recognition and clearance of apoptotic neutrophils by macrophages is an important mechanism involved in the generation of macrophage TGFβ1 and the resolution of acute inflammation (7, 9, 20). To determine if neutrophil phagocytosis of apoptotic cells was triggering neutrophil TGFβ1 production, freshly isolated MSU crystal–recruited peritoneal neutrophils were labeled with the neutrophil surface marker APC–Ly-6G and coincubated for 8 hours with apoptotic peritoneal neutrophils labeled with PerCP–Ly-6G. Flow cytometry analysis of the cell culture showed that TGFβ1 expression was greatest in live cells that had phagocytosed apoptotic cells (APC–Ly-6G positive/PerCP–Ly-6G positive) (Figure 4C). TGFβ1 levels in supernatants from cocultures of live and apoptotic neutrophils were significantly higher than in supernatants from single cell cultures of live or apoptotic neutrophils alone (Figure 4D). This indicated that the clearance of apoptotic cells by live neutrophils was a primary trigger for inducing TGFβ1 production.
TGFβ1 regulates IL-1β production and respiratory burst by neutrophils ex vivo.
It has been reported previously that self clearance of apoptotic neutrophils can suppress proinflammatory responses to external stimuli, including cytokine production and the production of ROS (21, 22). IL-1β is a primary signal for the inflammatory cascade, and ROS production by neutrophils can contribute directly to cell and tissue damage (3, 23); therefore, we investigated whether neutrophil TGFβ1 production could be moderating superoxide and IL-1β production in in vivo–activated and ex vivo–restimulated neutrophils. Purified neutrophils from MSU crystal–treated mice were treated with the TGFβ1-neutralizing antibody ex vivo and tested for the ability to produce superoxide and proinflammatory cytokines. In the presence of the anti-TGFβ1 antibody, MSU crystal–recruited peritoneal neutrophils, with and without MSU crystal restimulation ex vivo, exhibited an increase in the production of both superoxide and IL-1β (Figures 5A and B).
Neutrophils treated with anti-TGFβ1 antibody did not produce tumor necrosis factor α, IL-6, or the neutrophil chemokine CXCL1, either alone or after MSU crystal restimulation ex vivo (data not shown). Taken together, these results indicated that neutrophil phagocytosis and TGFβ1 production were important for regulating neutrophil respiratory burst and IL-1β production.
TGFβ1 suppresses neutrophil respiratory burst in vivo.
Finally, we investigated the effect of neutralizing TGFβ1 on MSU crystal–induced neutrophil superoxide and IL-1β production by neutrophils in vivo. As shown in Figures 5C and D, neutrophils isolated 18 hours after MSU crystal administration produced more superoxide and less IL-1β than neutrophils isolated 8 hours after MSU crystal administration. TGFβ1 neutralization did not alter either superoxide or IL-1β production by neutrophils isolated 8 hours after MSU crystal administration. However, TGFβ1 neutralization caused a significant increase in superoxide production by neutrophils 18 hours following MSU crystal administration, but IL-1β production was unaffected (Figures 5C and D).
The elimination of apoptotic cells and TGFβ1 production is a characteristic feature of the spontaneous resolution of acute inflammation and is traditionally attributed to macrophage function (20). Using a model of MSU crystal–induced inflammation, we now show that neutrophils also contribute to the resolution of inflammation via phagocytosis-driven TGFβ1 production.
Neutrophils are short-lived cells (15, 16) that play an integral part in innate immune responses. As professional phagocytes, neutrophils are well recognized for their ability to take up and kill microbes (2). However, recent literature reports that neutrophils also have the capacity to effect “noninflammatory” phagocytosis of apoptotic neutrophils in vitro and in vivo (21, 22) and may therefore contribute directly to the resolution of inflammation. Our current findings indicate that this mechanism of self clearance also triggers TGFβ1 production. In macrophages, phagocytosis of apoptotic cells is not essential for the induction of TGFβ1 production (24, 25). In contrast, phagocytosis of apoptotic cells appears to be important for the up-regulation of neutrophil TGFβ1, illustrated by the fact that neutrophil cannibalism specifically augments intracellular TGFβ1 accumulation and production.
The production of TGFβ1 is a hallmark feature of resolution in MSU crystal–induced inflammation and gout. Due to the large number of neutrophils recruited during the inflammatory response to MSU crystals, our data suggest that the production of TGFβ1 by cannibalizing neutrophils may in fact be a principal source of TGFβ1 during a gout attack. It is also interesting to note that in the absence of inflammation, blood neutrophils also generate TGFβ1 upon phagocytosis of apoptotic cells. Like macrophages, this TGFβ1 production is most likely a general function linked to the neutrophil's ability to recognize and engulf apoptotic cells rather than a disease-specific response per se. As such, neutrophil self clearance and TGFβ1 production may be particularly important in maintaining homeostasis in the blood of humans, where neutrophils are found in high numbers in the circulation.
A pivotal step in the proinflammatory response to MSU crystals is NLRP3 inflammasome activation and release of active IL-1β (26, 27). Recently, it has been proposed that this activation is triggered by ROS production (26, 28). In the present study we found that both neutrophil respiratory burst and IL-1β production were regulated by neutrophil TGFβ1 production in vitro. Although these findings could represent indirect evidence of a link between ROS production and IL-1β release, our in vivo data showed that neutralization of TGFβ1 increases ROS production without altering IL-1β production. Therefore, TGFβ1 production may be only one of a number of independent mechanisms capable of regulating the generation of active IL-1β in vivo. Such redundancy would remove reliance on any single event to switch off the IL-1β–dependent proinflammatory cascade, thereby reducing susceptibility to uncontrolled, chronic inflammation. In this context, TGFβ1 may be more important for regulating extracellular, rather than intracellular, ROS production by neutrophils during MSU crystal–induced inflammation in vivo.
Neutrophils are professional phagocytes that express many of the receptors associated with the recognition and clearance of apoptotic cells. Among these, transglutaminase 2 (TG2) is known to significantly affect neutrophil phenotype and function (29). TG2 expression is reported to be important in the clearance of apoptotic cells by macrophages in MSU crystal–induced inflammation (10). Combined with our findings and those of others (21, 22), TG2 expression on neutrophils could also contribute to apoptotic cell clearance via neutrophil cannibalism, ultimately leading to the production of TGFβ1.
In summary (Figure 6), we show that the process of neutrophil cannibalism triggers neutrophil TGFβ1 production, which in turn suppresses neutrophil respiratory burst and acts as a nonessential moderator of IL-1β production. In this way neutrophils have the capacity to actively drive the spontaneous resolution of inflammation in gout and other neutrophilic inflammatory conditions.
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. 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 conception and design. Steiger, Harper.
Acquisition of data. Steiger.
Analysis and interpretation of data. Steiger, Harper.
Xiao Liu (Malaghan Institute of Medical Research, Wellington, New Zealand) contributed toward the conception of the study.