Beneficial and Detrimental Effects of Plasmin(ogen) during Infection and Sepsis in Mice

Plasmin has been proposed to be an important mediator during inflammation/infection. In this study, by using mice lacking genes for plasminogen, tissue-type plasminogen activator (tPA), and urokinase-type PA (uPA), we have investigated the functional roles of active plasmin in infection and sepsis. Two models were used: an infection model by intravenous injection of 1×107 CFU of S. aureus, and a sepsis model by intravenous injection of 1.6×108 CFU of S. aureus. We found that in the infection model, wild-type (WT) mice showed significantly higher survival rates than plasminogen-deficient (plg-/-) mice. However, in the sepsis model, plg-/- or tPA-/-/uPA-/- mice showed the highest survival rate whereas WT and tPA+/-/uPA+/- mice showed the lowest survival rate, and plg+/-, tPA-/-, and uPA-/- mice had an intermediate survival rate. These results indicate that the levels of active plasmin are critical in determining the survival rate in the sepsis, partly through high levels of inflammatory cytokines and enhanced STAT3 activation. We conclude that plasmin is beneficial in infection but promotes the production of inflammatory cytokines in sepsis that may cause tissue destruction, diminished neutrophil function, and an impaired capacity to kill bacteria which eventually causes death of these mice.


Introduction
Sepsis is the culmination of complex interactions between the infecting microorganism and the host immune, inflammatory, coagulation and anti-coagulation responses [1]. Sepsis caused by infections with bacteria such as Staphylococcus aureus is a lifethreatening condition that may lead to septic shock, resulting in multiple organ failure and death [2]. It is well known that disorders of coagulation and fibrinolysis play a major role in the development of organ dysfunction during sepsis [3]. Plasmin, a potent serine protease in fibrinolysis and the key component of the plasminogen activator (PA) system, is generated by conversion from its precursor, plasminogen, by either of two physiological PAs, tissue-type PA (tPA) and urokinase-type PA (uPA) [4]. Besides fibrinolysis, plasmin also degrades a large range of extracellular matrix substrates and activates pre-matrix metalloproteinases [5]. Plasmin has therefore been suggested to be an important upstream regulator of extracellular matrix remodeling in many tissue degradation-related innate immune processes such as cell migration, tissue remodeling, inflammation, and complement activation [6][7][8].
In addition to its roles in extracellular proteolysis, the PA system is also involved in generation of pro-inflammatory responses in the extracellular environment. Studies using plasminogen-deficient (plg -/-) mice have provided evidence supporting a role of the PA system in mediating the migration of inflammatory cells towards inflammatory sites [9]. In vitro studies have also indicated that plasmin cleaves components of the complement system, thereby releasing chemotactic complement fragments [10,11]. Moreover, recent in vitro studies suggest that the PA system appears to be involved in the intracellular signaling events during inflammation. For instance, plasmin can activate the p38 mitogen-activated protein kinase (MAPK), Janus kinase (JAK), signal transducers and activators of transcription (STAT) signaling pathways in monocytes, which have been shown to be important for the inflammatory response [12]. Plasmin is also known to stimulate the release of cytokines and other inflammatory mediators by different cell types [13]. During severe infection, uncontrolled release of cytokines such as tumor necrosis factor-alpha (TNF-a) and interleukin-6 (IL-6) may cause a so-called cytokine storm [14]. An uncontrolled cytokine storm leads to sepsis, and is therefore fatal. However, although various mechanisms underlying the inflammatory response during infection have been proposed, the possible functional roles of the PA system during infection, and during sepsis especially, remain largely unknown.
In the current study, we have used single gene-deficient mice lacking plasminogen (plg -/-), uPA (uPA -/-), tPA (tPA -/-), and doubly deficient mice lacking both tPA and uPA (tPA -/-/uPA -/-), to study the functional roles of plasmin during S. aureus-induced infection and sepsis. Our data show that the plasmin shows contrasting roles in infection and sepsis.

Results
Reduced survival rate upon induction of infection and increased survival rate upon induction of sepsis in plg -/mice in comparison to WT mice To reveal the functional roles of plasmin in infection and sepsis, we used WT, plg +/-, and plg -/mice, which exhibit 100%, 50%, and 0% of the normal serum plasminogen level, respectively. Infection and sepsis were achieved by intravenous injection of 1610 7 or 1.6610 8 CFU of S. aureus, respectively, and the survival of mice was thereafter followed for 25 days. As shown in Figure 1A, in the first 4 days after bacterial injection no mice died in the infection group. From day 5 to day 25, the survival rates of WT, plg +/-, and plg -/mice were 86.7%, 80%, and 50%, respectively. The survival rate in plg -/mice upon induction of infection was significantly lower (P = 0.04) than that in the WT mice. A contrasting result was obtained when sepsis was induced ( Figure. 1B and Table 1). In this model, the median survival time in plg -/mice was 156 h, which was significantly longer (P = 0.009) than that in WT mice (40 h). Interestingly, the median survival time and the survival curve of plg +/mice were intermediate as compared to WT and plg -/mice. These data show that plasminogen deficiency has contrasting effects on the survival , plg -/-(n = 16) and plg +/-(n = 15) mice was followed for 25 days. Significant differences were found between WT and plg -/mice (log-rank test, P,0.05). B. Mice were inoculated with 1.6610 8 CFU of S. aureus to induce sepsis. Survival of WT (n = 28), tPA +/-/uPA +/-(n = 10), tPA -/-(n = 13), uPA -/-(n = 15), plg +/-(n = 20), plg -/-(n = 10), and tPA -/-/uPA -/-(n = 7) mice was followed for 8 days. Significant differences were found between WT mice and plg -/mice (log-rank test, P,0.01).

PA-deficient mice have delayed onset of death and an improved survival
Plasminogen is converted to the active enzyme plasmin by either uPA or tPA. We therefore studied S. aureus-induced sepsis in tPA -/or uPA -/mice, and also in doubly deficient (tPA -/-/uPA -/-) mice. As shown in Figure 1B and Table 1, the tPA -/and uPA -/mice showed delayed onset of death and significantly elevated survival rates as compared to WT mice (logrank test P = 0.04). Moreover, when tPA -/-/uPA -/mice were used to induce sepsis, they showed the longest median survival time compared to all other genotypes. The delayed onset of death in these tPA -/-/uPA -/mice was even more evident than in plg -/mice ( Figure 1B and Table 1). These data suggest that lack of tPA and/or uPA in mice significantly enhances the survival rate, and prolongs the onset of death in S. aureus-induced sepsis, thus confirming the notion that active plasmin plays a key role in the mortality of sepsis.

Comparison of viable bacterial counts in mouse organs after induction of infection and sepsis
In order to study whether there was a difference in the presence of bacteria in different organs in WT and plg -/mice after induction of infection, both genotypes of mice were inoculated with 1610 7 CFU of S. aureus, and 24 h later numbers of viable bacteria were counted from various organs. As shown in Figure 2A, the bacterial counts in liver, kidney, spleen, and blood were similar in WT and plg -/mice. The bacterial counts in brain, however, were significantly higher in WT mice than in plg -/mice. These results suggest that in the blood and in most organs, the ability to kill bacteria was similar in WT and plg -/mice at this early stage of infection. The difference in bacterial counts in the brain suggests that plasminogen may be required for bacterial invasion through the blood-brain barrier (BBB) [15,16].
We also investigated the presence of viable bacteria in various organs after induction of sepsis in WT, tPA -/-, uPA -/-, tPA -/-/uPA -/-, and plg -/mice. As shown in Figure 2B, the numbers of viable bacteria in the blood were not significantly different in all genotypes studied at 24 h after bacterial injection. However, as shown in Figure 2C-2F, the viable bacterial counts in the spleen, brain, liver, and kidney of WT, tPA -/-, and uPA -/mice were significantly higher than those in tPA -/-/uPA -/mice and plg -/mice. The viable bacterial counts in all organs studied were not significantly different in WT, tPA -/-, and uPA -/mice. The bacterial counts in the brains of tPA -/-/uPA -/mice were significantly higher than those in plg -/mice, but this was not the case for other organs ( Figure 2D). These data suggest that in sepsis, the bacterial killing ability is improved in mice with reduced levels of functional plasmin. Taken together, our data show that plasmin exhibits contrasting functions in bacterial killing in infection and sepsis.
Higher serum levels of TNF-a, IL-6, and IL-10 in WT mice than in plg -/mice during sepsis In vitro studies have indicated that plasmin stimulates the expressions of several cytokines [13]. A dramatic increase in inflammatory cytokine levels is one of the major clinical features of sepsis [17,18]. To study how plasminogen affects the cytokine production, we studied the profiles of inflammatory cytokines in the sera of WT and plg -/mice during infection and sepsis. As shown in Figure 3A (sepsis) and 3B (infection), serum levels of TNF-a reached a maximum level at 2 h after the injection of S. aureus, and gradually fell thereafter in both genotypes of mice and in both infection and sepsis. After induction of sepsis ( Figure 3A), significantly higher TNF-a levels were seen in WT mice than in plg -/mice at all 4 time points studied. In contrast, in the infection model ( Figure 3B), the TNF-a levels were either comparable (at 6 h and 24 h) or significantly lower (at 2 h and 36 h) in WT mice than in plg -/mice.
As shown in Figure 3C and 3D, IL-6 levels were comparable in WT mice and plg -/mice at 6 h, for both the infection model and the sepsis model. However, 24 h after induction of sepsis, WT mice had 8-fold higher levels of IL-6 than plg -/mice ( Figure 3C), while no difference was observed at the same time point in the infection model ( Figure 3D). At 24 h after induction of sepsis, the IL-6 levels in the sera of plg +/mice were also significantly lower than those in WT mice, but significantly higher than those in plg -/and tPA -/-/uPA -/mice ( Figure 3E).
As shown in Figure 3F, at 24 h after induction of sepsis, the IL-10 levels were significantly higher in WT mice than in plg -/mice, while the levels of IL-10 were similar between the two genotypes in the infection model. Furthermore, in both genotypes of mice, IL-10 levels were significantly higher in sepsis than in infection. Taken together, these data demonstrate that during sepsis, plg -/mice have significantly lower cytokine levels than WT mice. In addition, during sepsis, since the IL-6 levels are highest in WT mice, intermediate in plg +/mice, and lowest in both tPA -/-/uPA -/and plg -/mice, the increase in cytokine levels appears to be related not only to levels of plasminogen but also to levels of active plasmin. Improved survival in septic mice treated with IL-6 antibodies IL-6 is an important pro-inflammatory cytokine that is upregulated during sepsis [19]. As we found that the serum levels of IL-6 were significantly higher in WT mice than in plg -/mice 24 h after induction of sepsis, we tested the hypothesis that neutralization of IL-6 may delay the onset of death, or improve the survival rate in our sepsis model. As shown in Figure 4 and Table 2, when WT mice were treated with an anti-IL-6 neutralizing antibody (1.33 mg per kg body weight) at 1 h before and 24 h after the induction of sepsis, the onset of death was delayed for a period of 45 h, and the survival rate was significantly elevated as compared to mice treated with control IgG (P,0.05). WT mice treated with anti-IL-6 antibody showed a similar survival curve to that of plg -/mice, which suggests that lack of plasminogen improves the chances of survival from sepsis by reducing the production of IL-6. Activation of STAT3 during infection and sepsis STAT3 is a key molecule in mediating signaling of several inflammatory cytokines during an acute inflammatory response [20]. To study if plasminogen deficiency affect STAT3 activation, we measured the levels of phosphorylated and total STAT3 in spleen neutrophils obtained 24 h after infection and sepsis induction ( Figure 5A and 5B). In the infection model, levels of STAT3 and phosphorylated STAT3 are either moderately increased (in WT mice) or remained unchanged (in plg -/mice) as compared to uninfected controls. In the sepsis model, levels of total and phosphorylated STAT3 in WT mice increased significantly upon induction of sepsis. However, the levels remained largely unchanged in plg -/mice under sepsis challenge, indicating that plasmin is involved in the activation of the STAT3 signaling pathway in sepsis, which may contribute to the differences in cytokine levels, bacterial killing ability, and mortality between WT and plg -/mice during infection and sepsis.
Comparison of intensity of C5a on blood neutrophils in WT and plg -/mice during sepsis Studies using a murine cecal ligation and puncture sepsis model have shown that exposure of neutrophils to C5a at concentrations occurring in human plasma during sepsis leads to neutrophil dysfunction and paralysis of signaling pathways [21]. We therefore investigated whether plasminogen deficiency may influence the surface deposition of C5a on blood neutrophils after induction of sepsis. Twenty-four and 36 h after induction of sepsis, the intensity of C5a on neutrophils was determined. As shown in Figure 6, the intensity of C5a on neutrophils was lower in plg -/mice than in WT mice at 24 h after sepsis induction and was further reduced at 36 h. The diminished deposition of C5a on neutrophils in plg -/mice during sepsis suggests that C5a may contribute to the underlying molecular mechanisms in our sepsis model.

Discussion
In the present study, we report that plasmin plays a beneficial role in S. aureus-induced infection, but a detrimental role in S. aureus-induced sepsis. Several pro-and anti-inflammatory cytokines showed different profiles in WT and plg -/mice during infection and sepsis. In sepsis, blockade of IL-6 improved the survival of WT mice to a level that is close to plg -/mice. Furthermore, deposition of the active complement mediator C5a on the surface of neutrophils as well as activation of STAT3 is impaired in plg -/mice after induction of sepsis. These data demonstrate that plasmin has important functions in sepsis, which may go beyond its classical roles of degrading fibrin and extracellular matrix. In recent years, in vitro studies have suggested that plasmin plays a role in regulating signaling pathways, and in stimulating the release of cytokines and other inflammatory mediators [13]. To our knowledge, this is the first study that targets the roles of plasmin during host responses to infection and sepsis with in vivo models. Our results indicate that plasmin plays important roles in regulating cytokine expression, inflammatory signal transduction, bacterial killing ability, and mice survival rate during infection and sepsis.
We have previously shown that plasminogen levels are important for the incidence and severity of collagen II-induced arthritis [22]. Similarly, in the current study, our data with the sepsis model showed three types of survival phenotypes: (1) tPA -/-/uPA -/and plg -/mice, which cannot form active plasmin, have the highest survival rate, the longest median survival time, and the latest time of onset of death; (2) WT and tPA +/-/uPA +/mice have the lowest survival rate, the shortest median survival time, and the earliest time of onset of death; and (3) plg +/-, uPA -/-, and tPA -/mice have intermediate survival rate and time of onset of death. Because both tPA and uPA can activate plasminogen to functional plasmin, tPA or uPA singly deficient mice still have the capacity to convert plasminogen to active plasmin. Furthermore, tPA -/-/uPA -/mice have normal levels of plasminogen, but they cannot activate plasminogen to active plasmin. Thus, the survival data from the sepsis study suggest that it is the relative plasmin levels that determine the survival outcomes during sepsis.
During infection, however, WT and plg +/mice showed similar survival rates. These data indicate that unlike in sepsis, the average time of onset of death in the infection model was less dependent on the level of plasminogen. Furthermore, in sepsis, mice with normal or apparently normal ability to form plasmin (WT, tPA -/-, and uPA -/mice) all had significantly impaired bacterial killing ability as compared to mice that are unable to form plasmin (tPA -/-/uPA -/or plg -/mice). Moreover, the IL-6 levels in plg +/mice were approximately half of those in WT mice at 24 h after induction of sepsis, indicating that the relative levels of plasminogen also affect cytokine production. Based on these results, we propose that the available level of plasmin is critical for the onset of death, bacterial killing ability, and cytokine production in the development of sepsis.
STAT3 is a key molecule in mediating the signaling of many inflammatory cytokines such as TNF-a, IL-6, and IL-10 during inflammatory responses. TNF-a and IL-6 have been considered to be the primary mediators of sepsis [23], and a positive correlation has been found between serum IL-6 and TNF-a levels and multiple organ failure [24,25]. IL-10 has a pronounced antiinflammatory effect by reducing the level of superoxide production in neutrophils, which interferes with neutrophil-mediated cellular cytotoxicity [26]. Previous in vitro studies have shown that plasmin stimulates the expression of cytokines in human monocytes [13].  In our studies, we found that levels of total and phosphorylated STAT3 remained largely unchanged in plg -/mice undergoing sepsis, which is in contrast to the dramatic increased STAT3 activation in WT mice undergoing sepsis. In addition, we also found that TNF-a, IL-6, and IL-10 levels were significantly lower in plg -/mice than in WT mice during sepsis. In tPA -/-/uPA -/mice, the IL-6 was at a low level similar to that in plg -/mice. It is well known that the cytokine storm during sepsis induces an overwhelming inflammatory response, which leads to multiple organ failure and subsequent death. Blocking cytokines such as TNF-a, IL-6 and IL-1 did show benefit in patients with bacteremia [27]. Diminished cytokines levels in plg -/and tPA -/-/uPA -/mice indicate that there is a strong correlation between levels of active plasmin, cytokine expression and STAT3 activation. The data presented here suggest that reducing or removing functional plasmin in mice, leads to a higher survival rate during sepsis due to an impaired cytokine production. Studies of how plasmin regulate cytokine expression and downstream signal pathways during sepsis are being carried out in our laboratory.
Several studies have shown that mice with low complement activity have impaired host defense during infection [28][29][30], whereas during sepsis excessive complement activation leads to compromised innate immune functions [31]. Blockade of C5a or the C5a receptor with antibodies has been shown to greatly improve survival of rodents during sepsis [32,33]. In the current study, the contrasting phenotypes in terms of survival rate and cytokine expression between plg -/and WT mice during infection and sepsis are in accordance with the results of previous studies in mice with dysfunction of the complement system [34]. We have also shown that both the phosphorylated and total protein expression levels of STAT3 in spleen neutrophils of plg -/mice were markedly lower than those in WT mice after induction of sepsis. Thus, one possible explanation is that plasmin is important in mediating complement activation, which subsequently activates the STAT3 signaling pathway and contributes to the different phenotypes observed in plg -/and WT mice during infection and sepsis.
In summary, the current study reveals a contrasting role of plasminogen deficiency in infection and sepsis, and suggests that pro-inflammatory plasmin plays deleterious roles during systemic inflammation (sepsis). This pro-inflammatory function of plasmin is distinctly different from its classical roles as a protease that degrades fibrin and extracellular matrix. Our findings may be potentially useful for the development of novel therapeutic strategies against infection and sepsis in humans.

Bacterial strains
The Phillips strain of S. aureus isolated from a patient with osteomyelitis was kindly supplied by Dr. Höök, Gothenburg, Sweden. S. aureus was grown either in LB broth or on LB agar plates.

Models for induction of infection and sepsis
Infection model. Mice were challenged by intravenous injection with 1610 7 CFU of S. aureus Phillips. The survival rate was examined over the whole experiment (up to 25 days).
Sepsis model. Mice were challenged by intravenous injection with 1.6610 8 CFU of S. aureus Phillips. The survival rate was examined over the whole experiment (up to 8 days).

Bacterial counts in blood and different organs
Twenty-four hours after injection of the bacteria in amounts compatible with either infection or sepsis, the mice were killed and the blood, brain, spleen, liver, and kidney were collected and homogenized in 2 ml sterile PBS. After serial dilution, the homogenates and the blood were spread on LB agar plates and incubated at 37uC for 24 h. Colonies were counted and expressed as CFU/organ.

ELISA
At the indicated time points after induction of infection or sepsis, approximately 600 ml blood were collected, placed on ice and allowed to clot before centrifugation at 30006g for 10 min. Serum samples were then used for ELISA. The serum levels of TNF-a, IL-6, and IL-10 were measured using ELISA kits according to the instructions of the manufacturer (Nordic Biosite, Täby, Sweden).

Western blot analysis
Twenty-four hours after induction of infection or sepsis, spleens were collected and homogenized on ice, and the neutrophils were isolated by Percoll gradients (P4937; Sigma-Aldrich, Stockholm, Sweden). Neutrophils were lysed with 1% NP-40 and protease inhibitor cocktail. Thereafter, protein concentrations were measured by the BCA protein assay method (Pierce Biotechnology, Rockford, IL). Thirty micrograms of protein was electrophoresed by SDS-PAGE (7.5%), and transferred to polyvinylidene difluoride membranes (Amersham, UK) for western blot. Anti-STAT3 and anti-phosphorylated-STAT3 (Tyr 705 ) antibodies were obtained from Cell Signaling Technology (Boston, MA). The mouse monoclonal antibodies to b-actin were purchased from Sigma-Aldrich, Sweden. Western blot analyses were carried out as described by Persau et al [38].

Detection of C5a on blood neutrophils during sepsis by flow cytometry
Flow cytometric staining was performed on whole blood from 5 WT mice and 5 plg -/mice at 24 and 36 h, respectively, after sepsis induction. Leukoctes isolation was essentially followed by Zhang et al [39]. Cells were incubated with rat antibody to mouse neutrophils (MCA771G; Serotec, UK) and goat antibody to mouse C5a (AF2150; Minneapolis, MN) for 30 min on ice. After washing the secondary antibodies with FACS buffer, FITClabeled sheep anti-rat IgG diluted 1:100 (AAR10F; Serotec, UK) or PE-labeled donkey anti-goat IgG diluted 1:200 (Rockland Immunochemicals, PA) was added. The cells were washed once with FACS buffer and resuspended in 500 ml PBS. The samples were analyzed by double-colour fluorescence flow cytometry with cytomics FC500 (Beckman Coulter) for the detection of C5a intensity on blood neutrophils.

IL-6 blockade in WT mice 1 h prior to and 24 h after bacterial challenge
A rat monoclonal antibody (anti-IL-6; BD PharMingen, San Diego, CA) that specifically neutralize recombinant mouse IL-6 were used. The antibody were given to 9 WT mice at a dose of 1.33 mg per kg body weight at 1 h prior to and 24 h after the induction of sepsis. As controls, 14 WT mice were given control IgG (R3-34; BD PharMingen, San Diego, CA) at the same dosage. The mice were monitored every 8 h, and were killed 192 h after induction of sepsis.

Statistical analysis
Results of bacterial counts are expressed as the mean 6 SD. Data sets were analyzed using One-way ANOVA, with Bonferroni's Multiple Comparison test. Results of cytokine levels are expressed as the mean 6 SD. Data sets were analyzed using 2tailed t-test. To compare the survival curves among different genotypes of mice, log-rank test was used. P values less than 0.05 were considered significant.