Plasmin Induces In Vivo Monocyte Recruitment through Protease-Activated Receptor-1–, MEK/ERK-, and CCR2-Mediated Signaling

The plasminogen (Plg)/plasmin (Pla) proteolytic system is activated by the conversion of the zymogen Plg into the serine protease Pla mainly through the proteolytic activity of urokinase-type Plg activator and tissue-type Plg activator (1–4). The primary function in vivo of Pla is to degrade fibrin clots and regulate vascular patency (1, 2). However, the identification of Plg/Pla receptors in many cell types (5) along with the ability of Pla to degrade extracellular matrix showed the involvement of this protease in cell migration. Indeed, Pla-induced cell migration constitutes a key event in both physiological (embryogenesis, wound healing, and angiogenesis) and pathological processes (growth and spread of tumors and inflammatory diseases) (2–4, 6, 7).

The recruitment of leukocytes to the site of inflammatory insult characterizes the exudative phase of inflammation and requires the participation of several chemoattractant soluble factors, especially chemokines (8). After that, integrins promote firm adhesion of activated leukocytes to endothelial cells by binding to the cell adhesion molecules exposed on the endothelial surface. Subsequently, leukocytes crawl on endothelium, cross the basement membrane, and transmigrate into the interstitium (9). It has been demonstrated that the Plg system contributes to the process of cell migration in vivo. Plg knockout (Plg⁻/⁻) mice present deficient monocyte and lymphocyte recruitment in a model of peritonitis induced by thioglycollate (10, 11). Additionally, several studies have shown the role of distinct Plg receptors during leukocyte migration in vivo and in vitro (12–15). However, the role of chemokines and the mechanism underlying the signaling pathways that govern Plg-induced leukocyte migration in vivo have not been completely elucidated.

Plasmin stimulates expression of cytokines (TNF-α, IL-6, IL-1α/β, CD40), chemokines (MCP-1/CCL2), tissue factors, and the release of lipid mediators and chemotaxis in purified monocytes (16–19). Some of these effects are followed by the activation of the transcription factors NF-κB and AP-1 (16–19). Pla also triggers multiple signaling pathways such as JAK/STAT, p38 MAPK, and ERK1/2 (18–20) and it has been shown to activate protease-activated receptor-1 (PAR-1) in fibroblasts (21, 22). We have previously demonstrated that the proinflammatory cytokine IFN-α increased the expression of the Plg receptor α-enolase on the surface of human monocytes with a consequent increase of Pla generation and proteolysis (23). Furthermore, IFN-α and Plg-induced expression of α-enolase was dependent on the MEK/ERK pathway (23, 24).

In this study, we investigated the ability of Pla to induce cell migration in vitro and into the pleural cavity of mice and evaluated the underlying mechanisms. We demonstrated that Pla stimulates mononuclear cell recruitment in vivo through a PAR-1–dependent activation of MEK/ERK and NF-κB signaling, with the consequent release of CCL2. Additionally, Pla-induced mononuclear recruitment was dependent on CCR2 activation.

All procedures described here had prior approval from the Animal Ethics Committee of Universidade Federal de Minas Gerais (CETEA/UFMG protocol no. 19/2011). Male BALB/c mice (8–10 wk of age) obtained from the Bioscience Unit of Instituto de Ciências Biológicas were housed under standard conditions and had free access to commercial food and water. The CCR2-deficient mice (CCR2^−/−) (25) and C57BL/6 control mice were bred in the animal facility of the Immunopharmacology Laboratory.

Mouse embryonic fibroblasts (MEFs) and mouse leukemia monocyte macrophage (RAW 264.7) cell lines were cultured in DMEM supplemented with 10% (v/v) heat-inactivated FBS and antibiotics in 5% CO₂ at 37°C. After reaching 70–80% confluence, cells were serum starved in DMEM supplemented with 1% FBS for 24 h. Rabbit anti–p-ERK1/2 and anti–p-p65, mouse anti–p-IκB-α, and the MEK/ERK inhibitor U0126 were from Cell Signaling Technology (Beverly, MA). Mouse anti–β-actin Ab, purified human Pla (no. P1867, lyophilized powder, ≥2.0 U/mg), tranexamic acid (TXA), leupeptin (Leup), and human serum albumin (HSA) were from Sigma-Aldrich (St. Louis, MO). Three different batches of human Pla were used through the experiments (SLBB0579V, 091M1354V, and SLBD6287V). Murine Pla (lyophilized powder, >10 U/mg), batch no. B140215.01, was from Biopur (Reinach, Switzerland). PAR-1 inhibitors SCH79797 and RWJ56110, as well as the CCR2 antagonist RS504393, were from Tocris Bioscience (Ellisville, MO). PAR-1 agonist RP19979 was from GenScript (Piscataway, NJ). Selumetinib was from Selleck Chemicals (Houston, TX). Secondary anti-rabbit and anti-mouse peroxidase conjugate Abs were from Santa Cruz Biotechnology (Santa Cruz, CA).

Cells were serum starved as described above and then incubated with Pla at the times shown. When indicated, cells were preincubated for 60 min with the following concentrations of inhibitor prior to and throughout Pla treatment: U0126, 15 μM (MEK); Leup, 25 μg/ml (serine protease); and TXA, 0.01 M (lysine binding sites [LBS]) as previously performed (20).

For the cell migration assay, 5 × 10⁵ MEFs or 1 × 10⁶ RAW 264.7 cells were seeded into six-well plates and grown in DMEM containing 10% FBS to a nearly confluent cell monolayer. The cells were serum starved in DMEM containing 1% FBS for 24 h. The monolayers were carefully scratched using a 200-μl pipette tip. Cellular debris was removed by washing with PBS, and then the cells were incubated with Pla (2 μg/ml) for the indicated times or were pretreated with U0126 30 min before the scratch. Leup and TXA were incubated with Pla for 60 min at 37°C before addition to scratched monolayers. After treatment, cells were fixed in formaldehyde 10% solution for 10 min and washed three times with PBS. The cells that migrated into the scratch were observed and counted by using an Olympus IX70 microscope (×200 magnification). To obtain representative figures as in Fig. 1A and 1B, phase-contrast images were captured.

Cell migration induced by Pla was also analyzed by chemotaxis in triplicate using tissue culture–treated 24-well Transwell plates (Corning, Corning, NY) with polycarbonate membranes of pore size of 5.0 μm. RAW 264.7 cells were suspended in DMEM without FBS at a concentration of 5 × 10⁶ cells/ml, and 100 μl was added to the upper compartment of each well. Chemoattractants or medium was added to the lower compartments, and cells were allowed to migrate at 37°C for 4 h. Polycarbonate membranes were fixed and then stained with hematoxylin (Accustain; Sigma-Aldrich). MCP-1 (50 ng/ml) was used as standard chemoattractant. HSA used at the same concentration as Pla (2 μg/ml) was applied as human control protein. The cells were preincubated with U0126 (15 μM), selumetinib (15 μM), RS504393 (10 μM), or Pla preincubated with the inhibitors Leu (25 μg/ml) and TXA (0.1 M) 60 min before the experiment. Images were taken at ×400 magnification and then cell counts were performed.

Mice received an intrapleural (i.pl.) administration of Pla (2 μg/cavity), Plg (2 μg/cavity), HSA (2 μg/cavity), or PBS, as described (26, 27). Some groups received a pretreatment with specific inhibitors (U0126, 2 mg/kg or 60 μg/cavity i.pl.; Leup, 100 μg/cavity i.pl.; SCH79797, 5 mg/kg i.p.; RWJ56110, 1 mg/kg i.p.; and RS504393, 2 mg/kg i.pl.) 1 h before Pla. The PAR-1 agonist RP19979 was given at the dose of 50 μg/cavity i.pl. Doses of drugs were chosen according to previous data reported in the literature (11, 28, 29) or from in vivo dose-response experiments (for U0126, RP19979, and RWJ56110, data are not shown). Cells in the pleural cavity were harvested at different times after Pla injection, 24 h after Pla with inhibitors, or 48 h after RP19979 injection by washing the cavity with 2 ml PBS. Total cell counts were performed in a modified Neubauer chamber using Turk’s stain. Differential cell counts were performed on cytocentrifuge preparations (Shandon III) stained with May–Grünwald–Giemsa using standard morphological criteria to identify cell types. The results are presented as the number of cells per cavity.

Cell lines (MEFs and RAW) and inflammatory cells harvested from the pleural cavity were washed with PBS and whole-cell extracts were prepared as described (24, 26, 27). Protein amounts were quantified with the Bradford assay reagent from Bio-Rad. Extracts (40 μg) were separated by electrophoresis on a denaturing 10% polyacrylamide-SDS gel and electrotransferred to nitrocellulose membranes, as described (26). Membranes were blocked overnight at 4°C with PBS containing 5% (w/v) nonfat dry milk and 0.1% Tween 20, washed three times with PBS containing 0.1% Tween 20, and then incubated with specifics primary Abs (anti–p-ERK1/2, anti–p-IκB-α, anti–p-p65/RelA, or anti–β-actin) using a dilution of 1:1000 in PBS containing 5% (w/v) BSA and 0.1% Tween 20. After washing, membranes were incubated with appropriated HRP-conjugated secondary Ab (1:3000). Immunoreactive bands were visualized by using an ECL detection system, as described by the manufacturer (GE Healthcare, Piscataway, NJ).

The levels of the chemokines MCP-1/CCL2, KC/CXCL1, and MIP-2/CXCL2 and cytokines IL-1β, TNF-α, and IL-6 were measured in frozen supernatants obtained from pleural cavity washes after different time points of Pla or 6 h after Pla plus U0126 challenge by ELISA using commercially available Abs according to the procedures supplied by the manufacturer (R&D Systems, Minneapolis, MN).

Forty-eight hours after Plg, Pla, or PBS injection, the pleural cells were recovered by washing the cavity with PBS. The populations of monocytes/macrophages (CD45⁺/F4/80⁺/Gr-1^+/−), neutrophils (CD45⁺/ F4/80⁻/Gr-1⁺), and T lymphocytes (CD45⁺/CD3⁺) were analyzed by staining 2 × 10⁵ cells for 30 min on ice with fluorescent mAbs against F4/80 (PE-Cy7, eBioscience, San Diego, CA), Gr-1 (PE, BioLegend, San Diego, CA), CD45 (PE-Cy5, BD Biosciences, San Jose, CA), and biotinylated CD3. Stained cells were acquired in a BD FACSCanto II (BD Biosciences) and analyzed using FlowJo software (Tree Star, Ashland, OR). The cells were first selected based on size and granularity and CD45 to separate lymphocyte and macrophage/neutrophil populations. From the CD45 gate in the macrophage/neutrophil population, F4/80⁺ cells were selected versus Gr-1 as presented in representative dot plots (Fig. 3E). Percentages of cells in the total acquired events were plotted in the graph (Fig. 3D).

All results are presented as the means ± SEM. Data were analyzed by one-way ANOVA, and differences between groups were assessed using the Student–Newman–Keuls posttest. When only two groups were evaluated the Student t test was used. A p value <0.05 was considered significant. Calculations were performed using the Prism 4.0 software program for Windows (GraphPad Software, San Diego, CA).

Initial experiments evaluated whether Pla could induce in vitro migration of fibroblasts and a macrophage cell line (RAW 264.7) by using a cell scratch assay and the role of the MEK/ERK pathway in such process. Our results show that Pla promoted migration of fibroblasts (Fig. 1A) and macrophages (Fig. 1B) in a time-dependent manner. Because we have previously demonstrated that the MEK/ERK pathway is involved in Plg-induced gene expression in fibroblasts and monocytes (20, 23, 24), and it is known to be involved in cell motility (30), we wondered whether such pathway could underlie the Pla-induced migration in these cells. Interestingly, cell migration was associated with increased phosphorylation of ERK1/2 in both cell types (Fig. 1E, 1F), and the pretreatment of fibroblasts (Fig. 1C, 1E) or macrophages (Fig. 1D, 1F) with U0126, an MEK/ERK inhibitor, abolished the Pla-induced migration. Although Pla is a known in vitro inducer of cell migration of both cells types used in the present study, fibroblasts (22), and monocytes/macrophages (17, 31), we applied a different migration assay to validate the data obtained from the scratch assay, because it could be difficult to differentiate between migration and proliferation. By using a chemotaxis assay we were also able to observe that U0126 efficiently decreased macrophage migration (Fig. 2D). Similarly, selumetinib, a highly selective MEK/ERK inhibitor, also decreased Pla-induced RAW cell migration (Fig. 2E). Because Pla used in this study was purified from human serum, we also tested murine Pla and HSA on RAW cells to assure the sp. act. of Pla. Murine Pla but not HSA induced RAW migration (Fig. 2E). Collectively, these results suggest that MAPKs are activated and play a crucial role in Pla-induced macrophage and fibroblast in vitro migration.

Plg and Pla bind to cells, fibrin, and other extracellular matrix proteins via their LBS present in the kringle domains of these proteins. These interactions localize Plg activation and Pla activity to target surfaces (2, 4, 7). The protease activity of Pla is required for several processes, including cell migration (4). To investigate the importance of Pla protease activity and LBS to Pla-induced cell migration, we used the serine protease inhibitor Leup and the lysine analog TXA. The pretreatment with Leup or TXA impaired fibroblast (Fig. 2A) and macrophage (Fig. 2B, 2D) migration in vitro. U0126 was used as a control drug. Leup and TXA also inhibited Pla-induced ERK1/2 phosphorylation in both cell types, as presented in representative Fig. 2C (RAW cells). Therefore, both LBS and protease activity seem to be required for Pla-induced in vitro migration of fibroblasts and macrophages.

To investigate the ability of Pla to induce cell recruitment in vivo, we injected Pla into the pleural cavity of mice. There was no previous stimulus with any phlogistic agent other than Pla. Intrapleural injection of Pla induced a time-dependent influx of leukocytes into the pleural cavity of mice that was increased at 24 and 48 h (Fig. 3A). In agreement with previous reports on Plg-deficient mice (10, 11), the cells recruited to the pleural cavity after Pla injection were almost entirely mononuclear cells without any significant modification in neutrophil numbers (Fig. 3A). As performed in the in vitro setting, both murine Pla and HSA were injected in the pleural cavity of mice to evaluate for the specificity of the human Pla injected in our experiments. Similarly to human Pla, murine Pla induced leukocyte influx that was composed mostly of mononuclear cells (Supplemental Fig. 1A). HSA had no effect whatsoever and results were similar to those obtained in mice injected with PBS (Supplemental Fig. 1C). Interestingly, Plg also induced leukocyte recruitment to the pleural cavity that was similar to that induced by Pla (48 h PBS-injected mice, 13.2 ± 2.3 × 10⁵ mononuclear cells/cavity; 48 h Plg-injected mice, 26 ± 3.6 × 10⁵ mononuclear cells/cavity; n = 5 mice/group; p < 0.05). To identify cell populations that migrated into the cavity, we analyzed recruited cells for surface expression of CD45, F4/80, CD3, and Gr-1. Our results show that most cells that migrate in response to the Pla and Plg stimuli (Fig. 3D, 3E) are F4/80⁺, a marker for monocytes/macrophages. Confirming the pleural counts of the cytospin slides (showed in Fig. 3A), flow cytometry analysis showed that Pla or Plg injection was not able to increase the frequency of neutrophils or lymphocytes in pleural cavity when compared with PBS-injected mice (Fig. 3D). The gating strategy used in the present study is shown in Supplemental Fig. 2. Corroborating our in vitro data, we observed increased ERK1/2 phosphorylation (Fig. 3B) in recruited cells recovered from the pleural cavity.

NF-κB is a transcriptional factor associated with migration of many cell types including leukocytes (26, 32), and it is a target of ERK1/2. It has been shown that Pla activates NF-κB in human monocytes and dendritic cells (18, 19, 31, 33) and promotes nuclear translocation of NF-κB/Rel proteins subsequently to IκB-α phosphorylation/degradation. In our experimental settings the activation of NF-κB, inferred by the phosphorylation of its inhibitor (IκB-α), was increased in recruited cells recovered from the pleural cavity (Fig. 3B). We also evaluated NF-κB activation in a more direct way by analyzing p65/RelA phosphorylation levels in pleural cells recovered after Pla injection. Our in vivo results are in agreement with in vitro findings (18, 33) that have shown that Pla promotes NF-κB activation in multiple levels (Supplemental Fig. 1B).

It has been shown that Pla is able to activate monocytes and trigger the release of lipid mediators, cytokines, and chemokines in vitro (16, 18, 19). Owing to the importance of cytokines and chemokines to cell migration, we investigated pleural levels of the cytokines IL-1β, IL-6, and TNF-α and the chemokines MCP-1 (CCL2), KC/CXCL1, and MIP-2 (CXCL2) during Pla-induced mononuclear traffic into the pleural cavity. Interestingly, we observed increased levels of IL-6 and MCP-1 six hours after Pla injection, whereas the levels of TNF-α, IL-1β, KC/CXCL1, and MIP-2/CXCL2 did not change during the analyzed periods of time (Fig. 3C).

Because the protease activity of Pla is required for cell migration in vitro (Fig. 2A) and in vivo (11) and PAR-1 is implicated in Pla- and Plg-inducing cell effects (21, 22, 34, 35), we decided to investigate whether PAR-1 was involved in Pla-induced leukocyte migration to the pleural cavity of mice. As in our in vitro settings, the Pla-induced mononuclear migration was inhibited by pretreating mice with Leup (Fig. 4A), denoting a requirement of protease activity in such process. Similarly to Leup, pretreatment of mice with SCH79797 or RWJ56110, highly potent and selective PAR-1 antagonists, prevented Pla-induced mononuclear migration (Fig. 4C, 4D) and Pla-induced ERK1/2 and IκB-α phosphorylation (Fig. 4E). Consistent with the role of PAR-1 for the actions of Pla, the i.pl. administration of RP19979, a selective PAR-1 agonist, mimicked Pla effects and promoted mononuclear influx into the pleural cavity of mice (Fig. 4B).

Because inhibition of PAR-1 signaling impaired Pla-induced monocyte recruitment by affecting ERK activation, we wondered whether the MEK/ERK pathway was involved in Pla-induced leukocyte migration in vivo. In agreement with our in vitro data, pretreatment of mice with the MEK/ERK inhibitor U0126 inhibited Pla-induced leukocyte accumulation into the pleural cavity (Fig. 5A), NF-κB activation (Fig. 5B), and CCL2 release (Fig. 5C). Taken together, these results suggest that PAR-1 and MEK/ERK/NF-κB activation are required for the Pla-induced CCL2 release and leukocyte migration to the pleural cavity.

Because MCP-1/CCL2 levels were modulated during Pla-induced mononuclear recruitment (Figs. 3C, 5C) and this chemokine is a known chemoattractant to monocytes and lymphocytes by acting via the CCR2 receptor (25), we next evaluated whether Pla-induced monocyte recruitment was dependent on CCR2 activation. Interestingly, we found that CCR2 knockout (KO) mice were refractory to Pla stimulation (wild-type [WT] plus PBS, 14.3 ± 2.8 × 10⁵ mononuclear cells/cavity; WT plus Pla, 25.4 ± 2.4 × 10⁵ mononuclear cells/cavity; CCR2 KO plus PBS, 12.2 ± 1.8 × 10⁵ mononuclear cells/cavity; CCR2 KO plus Pla, 14.4 ± 0.9 × 10⁵ mononuclear cells/cavity; n = 8 mice per group; p < 0.001 when comparing WT plus PBS versus WT plus Pla and WT plus Pla versus CCR2 KO plus Pla). Because CCR2 is also involved in monocyte egress from bone marrow and it is known that CCR2 KO mice have fewer circulating monocytes (36), our results showing a decreased migration of mononuclear cells in Pla-injected CCR2 KO mice could be the result of a decreased monocyte frequency in the circulation rather than a minor recruitment of monocytes from blood to the tissues. Thus, we used a pharmacological approach to further investigate the role of the CCR2 receptor in Plg/Pla-induced cell migration. The administration of the CCR2 antagonist RS504393 significantly decreased Plg-induced migration in vivo (Fig. 6A) and the chemotaxis of RAW cells in vitro (Fig. 6B), suggesting an important role of the CCL2/CCR2 axis in Pla-induced mononuclear cell migration. The proposed model for the Pla-induced mononuclear cell migration is shown in Fig. 7.

The involvement of the Plg/Pla system in physiological and pathological events related to cell migration has been shown, but the mechanism of Pla-dependent transduction pathways underlying these processes have not been fully investigated in vivo. In this study, we performed assays using in vitro and in vivo models to investigate the effects of Pla on mononuclear cell migration into the pleural cavity of mice and also detailed signaling mechanisms behind this process. Our major findings are as follows: 1) Pla induced migration of both fibroblasts and macrophages in vitro through a MEK-ERK1/2–dependent pathway; 2) inhibition of serine protease activity or LBS prevented Pla-induced migration, demonstrating a requirement of proteolysis and lysine-bearing receptors for Pla activity; 3) Pla injection into the pleural cavity of mice promoted influx of mononuclear cells that was associated with ERK1/2 and NF-κB activation and increased levels of CCL2 and IL-6 in the pleural exudates; 4) inhibition of either PAR-1 or the MEK/ERK pathway impaired Pla-induced mononuclear cell migration associated with inhibition of ERK1/2, NF-κB, and CCL2 release; and 5) and the use of a CCR2 antagonist prevents Plg-induced monocytes migration in vivo and in vitro, reinforcing the role of the CCL2/CCR2 axis in this process. Therefore, we provide evidence that Pla is a mediator of cell migration in vivo by a mechanism dependent on activation of PAR-1 → MEK/ERK → NF-κB, leading to CCL2 production/release and CCR2-dependent recruitment of mononuclear cells to the pleural cavity of mice (Fig. 7).

Pla has been shown to induce movement of several cell types (17, 22, 37). In the present study we showed that Pla induced fibroblast and macrophage migration in vitro in a time-dependent manner. Fibroblast migration is a key event in dermal wound healing. After the skin is injured, several interacting events are initiated, including inflammation, tissue formation, angiogenesis, tissue contraction, and tissue remodeling (38). During the inflammatory response, there is formation of blood clots, and circulating leukocytes invade the wound region by migrating through the extracellular matrix. Fibroblasts subsequently migrate into the region, through biochemical alteration of the extracellular matrix by degrading fibrin, and replace the blood clot with collagen (38). Thus, several events regarding tissue growth and repair require fibroblast migration, and skin wound is just an example of such a fundamental physiological role of Pla. Although we did not investigate the in vivo migration of fibroblasts in the present study, our in vitro results provide evidence that Pla could participate in such an in vivo process. Accordingly, a recent work from Shen et al. (39) has shown that Plg is a key proinflammatory regulator that accelerates the healing of acute diabetic wounds. Indeed, the importance of Plg to wound healing was previously shown in Plg^−/− mice (6).

Cellular migration requires specific intracellular signaling cascade events. Although Pla-induced signal transduction pathways have been analyzed in vitro in a few previous reports (19, 21, 23, 24, 31, 33), the signaling pathways that govern Pla-induced cell migration in vivo have not been studied in detail. It has been shown that Plg/Pla activates ERK1/2 in fibroblasts (20, 21, 23, 24), monocytes (31), and dendritic cells (33). We have previously shown that Plg/Pla activates the MEK/ERK pathway, leading to the expression of the transcription factors c-fos and Egr-1 (20), which in turn promotes α-enolase expression (23, 24), an important Plg receptor that mediates monocyte recruitment in vivo (14). In the present study, we show that activation of the MEK/ERK pathway is essential to the Pla-inducing effects, because a specific MEK/ERK inhibitor abolished Pla-induced migration in in vitro (macrophages and fibroblast) and in vivo (mononuclear cells) settings. Indeed, the levels of ERK1/2 phosphorylation were in line with the increased influx of mononuclear cells into the pleural cavity.

Pla has been shown to activate NF-κB in human monocytes, macrophages, and dendritic cells in vitro (18, 19, 31, 33) and promotes nuclear translocation of NF-κB/Rel proteins subsequently to IκB-α phosphorylation/degradation. Such activation is involved in Pla-induced expression/release of proinflammatory cytokines (IL-1α/β and TNF-α) and tissue factor by human monocytes (18, 31). In our in vivo model, similarly to the ERK1/2 phosphorylation profile, Pla-induced NF-κB activation was paralleled with the influx of mononuclear cells. Importantly, the treatment with U0126 also prevented Pla-induced IκB-α phosphorylation, suggesting that NF-κB activation is downstream of ERK1/2 in our system.

We observed increased levels of IL-6 and CCL2 at 6 h after Pla injection, an interval that precedes the recruitment of mononuclear cells to the pleural cavity. In contrast to in vitro studies using human monocytes (18), levels of TNF-α, IL-1β, KC/CXCL1, and MIP-2/CXCL2 were not altered in pleural exudates following Pla injection at the times analyzed (6, 24, and 48 h). It is possible that these cytokines and chemokines are produced earlier after Pla injection, are not detectable in the conditions of our experimental situation, or may not be produced in vivo. Of note, Pla-activated monocyte-derived dendritic cells did not release these proinflammatory cytokines (33). Cailhier et al. (40) showed that the ablation of resident pleural macrophages in a model of carrageenan-induced pleurisy had marked effects on the production of several cytokines and chemokines, such as CXCL2, TNF-α, IL-6, and IL-10. However, the drop in levels of KC and MCP-1 was not so intense after macrophage ablation, suggesting that there was a source other than residential macrophages, such as mesothelial cells, in MCP-1 production in vivo and subsequent mononuclear cell recruitment (40). Although we did not perform experiments to identify the cell population responsible for the detected cytokine production in our model, we think that the findings on the carrageenan-induced pleurisy can be applied in the Pla-induced pleurisy as well, leading to the conclusion that residential macrophages are responsible for the production of IL-6 and mesothelial cells for the MCP-1 release (40). It was previously stated that Pla activates macrophages via the annexin A2 heterotetramer (composed of annexin A2 and S100A10) and the following signaling pathways lead to IL-6 production (31). IL-6 is able to induce neutrophilia and neutrophil activation (41), but we probably did not observe these effects owing to a lack of Pla-induced neutrophil chemoattractants, such as MIP-2/CXCL2 and KC/CXCL1 in our model.

Migration of circulating leukocytes to sites of injury or inflammation is a crucial step of both innate and adaptive immunity. To achieve this, finely coordinated mechanisms exist by which intravascular leukocytes become able to penetrate the vascular wall and migrate to the sites of injury or infection (8). The administration of Pla to the cremaster muscle of mice, at the same dose used in the present work (2 μg), induced firm adhesion and subsequent transmigration of leukocytes (42). In an attempt to understand better how Pla injection induced leukocyte migration, we used an in vivo model by injecting human Pla into the pleural cavity of mice and detected significant leukocyte recruitment. The total number of leukocytes (mirrored by number of mononuclear cells) was increased in a time-dependent manner, but the number of neutrophils (and neutrophil chemoattractant proteins) was not altered significantly after Pla injection, suggesting specific recruitment of mononuclear cells in this model. These data are in agreement with others that applied the model of thioglycollate-induced cell migration to the peritoneal cavity using Plg^−/− mice (10, 11). Those mice presented decreased recruitment of lymphocytes, monocytes, and macrophages, whereas neutrophils remained similar to the WT (10). Additionally, in in vitro settings Pla induced potent monocyte chemotaxis and actin polymerization that were similar to those induced by the chemoattractant fMLF, and such effect was not seen in neutrophils (17), reinforcing the idea of cell-specific recruitment. In contrast to these findings, a study accessing the inflammatory response to implanted biomaterials in the peritoneal cavity demonstrated that both neutrophil and macrophage recruitment to the peritoneal lavage were decreased in Plg^−/− mice (43). These different results might be due to the nature of the stimulus used to promote cell recruitment (Ag X implanted), suggesting that the Plg influence on leukocyte migration may be stimulus specific (44). We showed in the present study that Plg induced cell migration to a similar extent as did Pla. It has been stated that Plg is quickly converted to Pla by urokinase-type Plg activator or tissue-type Plg activator, and protease activity of Pla is essential on Plg-induced migration (4, 11). Thus, it is likely that urokinase-type Plg activator generated in the pleural milieu after the injection may be activating the exogenously administered Plg and promoting cell migration. One must bear in mind that Plg was injected into the cavity using a fine sterile syringe. This small injury caused in the injection site is insufficient to generate leukocyte influx but may initiate events leading to Pla activation. This suggestion is currently under investigation in our laboratory.

Several Plg receptors have been shown to participate in Pla-induced migration in vivo and in vitro (12–15, 45). These receptors have two common properties: they possess a carboxyl-terminal lysine residue that can bind Plg and they are not transmembrane proteins but become expressed on the cell surface by yet undefined pathways (46). The binding of Plg to receptors in the surfaces of cells via its LBS enhances its activation to Pla by protecting it from inactivation by α₂-anti-Pla, and by directing its proteolytic activity to specific substrates (2, 7). In the present study, although we have not investigated in much detail the relevance of any specific Plg receptor to Pla-induced cell migration, we showed that migration of macrophages was impaired by TXA treatment, a lysine analog that impairs the ligation of LBS present on the Pla kringle domain to the carboxyl-terminal lysine residues present on the cell surfaces. Indeed, lysine analogs such as TXA and ε-aminocaproic acid have been reported to inhibit efficiently Pla activity and migration (11), showing a key role for Plg receptors in Pla-induced migration.

The requirement of protease activity for Pla-induced cell migration seems to be a common mechanism to different cell types (4). Similar to the requirement of LBS, the protease activity of Pla is also necessary to induce cell migration in vitro and in vivo as shown in this work by using a serine protease inhibitor (Leup) and the PAR-1 inhibitor SCH79797. Indeed, when a lysine analog is present, Pla is unable to cleave and activate PAR-1, suggesting that Pla interaction with the surface is necessary for PAR-1 activation (34, 47). In this context, we could speculate that PAR-1 would be a Plg/Pla coreceptor, and it may be necessary for the action of Pla. In this study, we have not explored in which receptor containing LBS would Pla bind to promote activation of PAR-1. Receptors such as α-enolase, histone H2B, or Plg-R_KT are potential ligands and have been described to be involved with monocyte migration in vivo (13–15). Pla is known to activate PAR-1 by proteolysis, a process that is involved in several Pla-induced effects, such as antiapoptotic signaling in monocytes (34), induction of Cyr61 gene expression in fibroblasts (21), migration of Chinese hamster ovary cells (22), expression of IL-8 and PGE₂ release (48), and promotion of epithelial-to-mesenchymal transition in renal interstitial fibrosis (35). In the present study, we showed that PAR-1 inhibition impaired Pla-induced migration of mononuclear cells and also led to the inhibition of ERK1/2 and NF-κB activation. Indeed, macrophages and monocytes express PAR-1 (49). Pendurthi et al. (21) showed that PAR-1^−/− fibroblasts are not responsive to Pla induction of ERK1/2, suggesting that the mechanism of the Pla-induced MEK/ERK pathway is dependent on PAR-1. It has also been shown that Pla induces MER/ERK through PAR-1 in renal interstitial fibrosis (35), and PAR-1 activation is necessary for CCL2-dependent leukocyte recruitment in vivo (50). Therefore, the aforementioned studies are supportive of our results showing Pla-induced PAR-1–dependent mononuclear migration.

The expression of cytokines and chemokines is regulated primarily at the transcriptional level. Thus, the expression of CCL2 in response to stimulus with thrombin may occur by the MEK/ERK pathway (51), and CCL2 is a known NF-κB target gene (52). Alternatively, Burysek et al. (19) showed that Pla-induced expression of CCL2 in vitro did not require the MEK/ERK signaling pathway by using U0126 at dose of 1 μM. In our in vivo study, i.pl. injection of Pla induced CCL2 levels that were sensitive to MEK/ERK inhibition with U0126 at dose of 60 μg/cavity. CCR2 is mainly expressed in monocytes and binds to a group of MCP chemokines, including MCP-1/CCL2, MCP-2/CCL8, and MCP-3/CCL7 playing a key role in monocyte migration in different experimental models and diseases (25). In our settings, CCR2^−/− mice were unable to respond to i.pl. injection of Pla, and the use of a CCR2 antagonist inhibited Plg- and Pla-induced macrophage migration in vivo and in vitro, showing the importance of the CCL2/CCR2 axis on the Pla-mediated cell recruitment. In agreement with our findings, it has been shown that after inflammatory stimulus Plg^−/− mice present a 50% decrease in the levels of CCL2 expression compared with Plg^+/+ mice, suggesting that differences in CCL2 levels might contribute to the reduced recruitment of macrophages and lymphocytes in the Plg^−/− mouse (43). These finding reinforce the idea that release of CCL2 is relevant for Pla-induced mononuclear recruitment. Of note, it has been shown that Pla also cleaves CCL2 at lysine 104, increasing its chemotactic potency (53), and that Pla-mediated truncation of CCL2 may be an indispensable step to full activation of this chemokine in vivo (54). Thus, the Pla-induced CCL2 release (shown in the present study) and putative increase of potency (54) (but not shown in the present study) seems to be governing Pla-induced leukocyte migration into the pleural cavity.

Although we focused our study on CCL2 and CCR2 and show the clear importance of this axis in our model, we cannot exclude the possibility of having other cytokines and chemokines participating in the process. Indeed, other chemokines such as CCL3 (MIP-1α), CCL4 (MIP-1β), CCL5 (RANTES), and their main receptors (CCR1 and CCR5) are also involved in macrophage recruitment (55), but we did not investigate the role of these chemokines in our experimental model.

Taken together, our data show that Pla induces cell migration by cleaving PAR-1 and triggering the activation of MEK/ERK/NF-κB pathway, which culminates with the induction of CCL2 and IL-6 release in the pleural cavity of mice. Furthermore, we provide evidence that the CCL2/CCR2 axis contributes to Pla-induced migration of mononuclear cells to the sites of inflammation (Fig. 7). To our knowledge, this is the first study that details the underlying mechanism of Pla-induced cell migration in vivo, from receptor binding activation of signaling pathways to chemokine release and action. authors declare no conflict of interest.

We thank Dr. Massimo Locatti for providing the PAR-1 inhibitor, Dr. André Klein for providing the PAR-1 agonist, Dr. Mauro Perretti by providing Pla, and Dr. Fernando Cunha for providing the CCR2 knockout mice. We also thank Frankcinéia Assis and Ilma Marçal for technical assistance.

The authors have no financial conflicts of interest.