Mycobacterium tuberculosis PPE18 Protein Reduces Inflammation and Increases Survival in Animal Model of Sepsis

Sepsis is a syndrome caused by an abnormal host response to infection. The infection is mostly due to bacteria but can also be due to viruses, fungi, or parasites. It is a major cause of mortality especially in hospitals (1). Many survivors have profound disabilities, such as amputated limbs, blindness, and cognitive problems. Sepsis leads to an overwhelming inflammatory response in the host, which manifests as well-known clinical symptoms (fever, tachycardia, leukocytosis) and the accompanying systemic inflammatory response syndrome (2). In sepsis, infection overstimulates the host immune response by activating monocytes/macrophages, neutrophils, and endothelial cells. The activation of these cells results in an elaborate and extensive array of proinflammatory mediators, which includes cytokines such as TNF-α, IL-1β, IL-6, and IL-8, as well as lipids, oxygen and nitrogen radical intermediates, components of the complement cascade, catecholamines, histamines, and others. The chemical mediators can cause local damage to cells and systemic toxic effects (3–5). Although sepsis is known to be a complex condition which involves immune suppression contributing to morbidity and mortality (6), the excessive production of proinflammatory cytokines, the primary being TNF-α and IL-1, are responsible for the pathophysiology of sepsis, which includes vasodilation and increased capillary permeability further leading to hypotension, hemoconcentration, macromolecular extravasation, cardiac dysfunction, and multiple organ failure. Also, inflammation-induced dysregulation of the coagulation system can result in disseminated intravascular coagulation (DIC) (7). Injection of TNF-α and IL-1 in experimental animals leads to development of a shock-like state characterized by pulmonary edema and hemorrhage (8). Similarly in humans, administration of TNF-α leads to systemic inflammatory response syndrome (9, 10). Clinical trials in sepsis patients have shown benefits of anti–TNF-α Ab (11). In patients with severe sepsis, anti–TNF-α Ab reduced mortality and in patients suffering from shock, there was a trend for better survival with use of anti–TNF-α Ab (12). Also, reduction in levels of TNF-α and overall inflammation correlates with increased survival in animal models of sepsis (13–15). Indeed, an exaggerated and dysregulated inflammatory response poses a challenge in sepsis and efforts need to be geared toward identifying effective anti-inflammatory agents which can reduce TNF-α and inflammation for successful resolution of sepsis (16–20).

We have earlier demonstrated that a Mycobacterium tuberculosis protein belonging to the PPE family, PPE18, also known as Mtb39a (Rv1196), binds to TLR2 and causes IL-10 induction in macrophages via activation of p38 MAPK (21). Also, its interaction with TLR2 leads to phosphorylation of suppressor of cytokine signaling 3 (SOCS3), which then physically interacts with the IκBα–NF-κB/rel complex, thus preventing phosphorylation and degradation of IκBα and nuclear translocation of p50 and p65 NF-κB and c-rel transcription factors. As a consequence of this, there is downregulation of transcription of NF-κB–regulated genes such as IL-12 and TNF-α (22). PPE18 selectively downregulates proinflammatory immune responses. At the same time, it increases secretion of IL-10 (21), which is an anti-inflammatory cytokine. Also, our earlier work demonstrated that macrophages infected with ppe18^−/−M. tuberculosis produce less IL-10 and more IL-12 p40 compared with those infected with wild-type M. tuberculosis (21). In M. tuberculosis infection, anti-inflammatory proteins such as PPE18 likely reduce levels of host protective proinflammatory cytokines such as TNF-α, thus aiding survival of the bacilli (23). These same properties of PPE18 can be exploited to dampen the effects of extreme inflammation observed in situations such as sepsis. With this rationale, we decided to test recombinant PPE18 (rPPE18) as a therapeutic in a mouse model of sepsis. Our studies revealed that treatment of mice with rPPE18 reduces TNF-α levels, generates M2 macrophages, inhibits DIC, and improves clinical symptoms and survival of mice suffering from septicemia induced by i.p. injection of a high dose of Escherichia coli. Additionally, rPPE18 reduced TNF-α, alanine transaminase (ALT), creatinine, and organ damage and improved survival in a mouse model of cecal ligation and puncture (CLP)-induced polymicrobial sepsis. In light of these observations, we believe that rPPE18 or peptides derived from it may have the potential for being developed as antisepsis therapeutics.

The rPPE18 was purified as described earlier (21). In brief, 6X-histidine–tagged PPE18 cloned in pRSETa was overexpressed in E. coli BL21 (DE3) cells. Protein expression was induced by addition of 1 mM isopropyl β-d-1-thiogalactopyranoside (Sigma-Aldrich, St. Louis, MO) to log phase cultures, which were grown further for 3 h. Bacterial cells were pelleted down, resuspended in PBS (pH 7.4) containing 0.3 mg/ml lysozyme (Sigma-Aldrich) and 1 mM PMSF (Sigma-Aldrich), and incubated at 37°C for 20 min. Next, 1% sodium lauryl sarcosine (Sigma-Aldrich) was added and the mixture was sonicated. TALON (Clontech Laboratories, Mountain View, CA) beads were added to the supernatant obtained postsonication. rPPE18-bound TALON beads were washed with 20 mM imidazole (Sigma-Aldrich) to remove bound impurities. rPPE18 was then eluted with 200 mM imidazole. Eluted fractions were resolved on a gel and stained with Coomassie brilliant blue to assess purity. Eluted rPPE18 was extensively dialyzed against PBS to remove imidazole. Dialyzed protein was incubated with 10% v/v polymyxin B-agarose (Sigma-Aldrich) beads, as described earlier, to remove contaminating LPS (21). Before use, the polymyxin B-agarose slurry was washed extensively with sterile PBS to remove any unconjugated polymyxin B present in the slurry. After incubation with protein, beads were pelleted down by centrifugation and LPS-free protein was carefully removed. Purified protein was filter-sterilized using a 0.4-μm syringe filter and then used for experiments. The protein concentration was determined using a kit that employs the bicinchoninic acid method of protein estimation (Thermo Fisher Scientific, Waltham, MA). In some cases, purified rPPE18 was subjected to repeated heating (100°C) and freezing (liquid N₂) to obtain denatured protein which was used as control. For some experiments BSA was treated with polymyxin B-agarose as described above and used as control.

BALB/c mice were maintained at the animal house facility of Vimta Labs, Hyderabad and the experimental protocols were approved by and performed as per the guidelines of the Institutional Animal Ethics Committee of Vimta Labs, Hyderabad. Log phase E. coli BL21 cells were washed, resuspended in 0.1 ml of sterile PBS, and injected i.p. in age- and weight-matched male BALB/c mice (8–12 wk old) at a dose of either 2.5 × 10⁸ or 5.0 × 10⁸ CFU to induce septic peritonitis as described by others (24, 25). Mice were administered native or denatured rPPE18 (100 μg per mice) or an equivalent volume of PBS (as buffer control) 1 h before or after injection with E. coli. Mice were bled retro-orbitally at specific times. Blood was left to coagulate at room temperature for ∼2 h. Sera were then separated by centrifugation and stored at −20°C until further use in ELISA. For some experiments, the peritoneal cavity of euthanized mice was washed with 5 ml sterile PBS to collect lavage. Clear lavage was obtained by sequential centrifugation first at 2500 rpm to remove mouse peritoneal cells and next at 5000 rpm to remove E. coli bacteria. Peritoneal lavage cells were used for Western blot analysis and cell free clear lavage was used for ELISA.

CLP was performed as described earlier (26). Briefly, 8–12-wk-old male, weight-matched BALB/c mice were anesthetized by administration of ketamine HCl (Troika Pharmaceuticals, Dehradun, India) and xylazine HCl (Indian Immunologicals, Hyderabad, India). A midline abdominal incision was made under sterile conditions and the cecum below the ileocecal valve was ligated with suture, thereby preserving intestinal continuity. Once ligated, the cecum was punctured through with a 16 G needle twice. The cecum was repositioned, the peritoneum was closed with absorbable suture, and the skin was closed using silk suture. Sham-operated mice underwent the same procedure including opening of the peritoneum and exposing the bowel but without ligation and needle perforation of the cecum. One milliliter sterile saline was injected s.c. immediately after surgery. rPPE18 (100 μg per mice) or an equivalent volume of PBS was injected i.p. at specific times after CLP procedure. Survival rates were determined over a period of 16 d with assessment every 12 h. From an independent set of mice, blood was collected 16 h post-CLP procedure for measurement of serum ALT, creatinine, and complete blood count analysis. Peritoneal lavage was collected in 1.5 ml sterile saline 16 h post-CLP procedure for measuring TNF-α.

Cytokines were quantified in sera, peritoneal lavage, and cell culture supernatants by sandwich ELISA according to the manufacturer’s instructions as described earlier (21, 22). Absolute concentrations of TNF-α (eBioscience, San Diego, CA), IL-1β (BD Biosciences, Sparks, MD), IL-6 (PeproTech, Rocky Hill, NJ), IL-12 p70 (eBioscience), and IL-10 (BD Biosciences) cytokines were measured using a standard curve provided by the manufacturers.

Liver and spleen were aseptically removed from euthanized mice, fixed in 10% neutral buffered formalin and then embedded in paraffin wax. Sections were then prepared and stained with H&E for visualizing cell death and organ damage that accompanies septicemia (27). Sections were visualized under a Nikon DS-Fi1 microscope (Shinagawa, Tokyo, Japan). Microphotographs were taken using a camera attached to the microscope.

BALB/c mice (6–8 wk old) of either sex were injected i.p. with 1 ml of 4% sterile thioglycollate broth (HiMedia Laboratories, Mumbai, Maharashtra, India). After 4 d, mice were sacrificed and peritoneal macrophages were harvested as described earlier (28). Macrophages were seeded either in 96-well plates at a density of 0.2 × 10⁶ or in 24-well plates at a density of 0.8 × 10⁶ and were treated with different concentrations of rPPE18 for 1 h and then activated with 1 μg/ml of bacterial LPS (Sigma-Aldrich). Cells were cultured in RPMI 1640 medium (HyClone, GE Healthcare, Chicago, IL) supplemented with 10% FBS (HyClone), HEPES (Sigma-Aldrich), glutamine (Sigma-Aldrich), and antibiotic-antimycotic mixture (Sigma-Aldrich) and maintained under sterile conditions of 37°C, 5% CO₂, and 85% humidity. After 24 h, culture supernatants were collected for estimation of NO and cytokines, whereas cells were harvested for Western blotting.

Blood samples were collected retro-orbitally from mice in either 4% EDTA or 3.2% sodium citrate to collect whole blood and plasma, respectively. Blood in sodium citrate was immediately centrifuged for plasma collection to measure prothrombin time (PT), fibrinogen (FBG), and activated partial thromboplastin time (aPTT) by Sysmex CA-50 Automated Blood Coagulation Analyzer (Sysmex, Kobe, Japan). From whole blood, WBC and platelet (PLT) counts were determined by Medonic CA-620 hematology analyzer (Medicon Ireland, Newry, Northern Ireland, U.K.).

For differential leukocyte count, blood was collected in EDTA tubes and analyzed on automatic cell analyzers, ADVIA 2120 (Hematology System; Siemens Healthcare Diagnostics, Forchheim, Germany).

For Western blotting, cells were lysed using cold lysis buffer comprising PBS, 1% NP-40, and protease inhibitor mixture (Roche, Penzberg, Germany). Cells were kept in lysis buffer on ice for 10 min and then subjected to one round of freeze-thaw after which the lysates were centrifuged at 12,000 rpm for 10 min to remove debris. Approximately 30–40 μg of proteins were separated on a 12% SDS-PAGE and then transferred on to a nitrocellulose membrane. After blocking overnight with 5% fat-free milk, membranes were probed with anti–NO synthase 2 (NOS2) Ab or anti–arginase-1 Ab or anti–β-actin Ab (all from Santa Cruz Biotechnology, Santa Cruz, CA) followed by an appropriate HRP-conjugated secondary Ab (Sigma-Aldrich) and then developed using ECL reagent (GE Healthcare). For detecting total and phosphorylated p38 MAPK levels, lysates were prepared in a lysis buffer containing 20 mM HEPES (Sigma-Aldrich), 250 mM NaCl (Sigma-Aldrich), 1% NP-40 (Sigma-Aldrich), 2 mM EDTA (Sigma-Aldrich), aprotinin (Sigma-Aldrich), leupeptin (Sigma-Aldrich), DTT (Sigma-Aldrich), and PMSF and sodium orthovanadate (both Sigma-Aldrich). After resolving on a 12% SDS-PAGE, proteins were transferred onto a nitrocellulose membrane and blocked with 5% BSA (Sigma-Aldrich). The membranes were then incubated with Abs to either phospho-p38 MAPK (BD Pharmingen, San Jose, CA) or total p38 MAPK (BD Pharmingen). Membranes were washed and incubated with HRP-conjugated anti-mouse Ab (Sigma-Aldrich) and developed using ECL reagent (GE Healthcare).

Arginase activity in cell lysates was measured using a commercial kit (QuantiChrom Arginase Assay Kit; Bioassay Systems, CA). Briefly, 0.8 × 10⁶ thioglycollate-elicited peritoneal macrophages were treated with 3 μg/ml rPPE18, 1 h prior to being activated with 1 μg/ml LPS. After 24 h, cells were detached from the wells, washed, and lysed in a buffer containing Tris-HCl (pH 7.4) (Sigma-Aldrich), 0.4% Triton X-100 (Sigma-Aldrich), and protease inhibitor mixture (Roche). Lysates were centrifuged for 10 min at 12,000 rpm at 4°C. The resultant supernatant was used for the arginase assay after protein estimation using bicinchoninic acid reagent (Thermo Fisher Scientific). The assay was performed according to the manufacturer’s instructions using 10 μg of protein. The assay measures urea generated as a result of arginase activity as a chromogenic product whose absorbance is measured at 430 nm. Cell lysates were incubated with substrate in the presence of manganese and urea generation was allowed for 2 h at 37°C. The urea formed was converted to a chromogenic product by addition of urea stop reagent and incubated for a further 1 h at room temperature. Arginase activity was calculated with respect to a urea standard and expressed as U/L.

Levels of ALT were measured in sera of septic mice subjected to E. coli BL21–induced septic shock using a commercial ALT assay kit according to the manufacturer’s instructions (Sigma-Aldrich). ALT activity was measured as a coupled enzymatic reaction, which results in product formation measured colorimetrically. ALT enzyme activity was proportional to pyruvate generated. Concentration of pyruvate generated was calculated with the help of a pyruvate standard curve. In brief, BALB/c mice were injected i.p. with 100 μg of rPPE18 or an equivalent volume of PBS before being administered either 2.5 × 10⁸ or 5.0 × 10⁸ E. coli BL21 CFU. After 3 or 24 h, mice were bled retro-orbitally and sera were separated and used for ALT assay. The assay was performed for 45 min at 37°C. ALT enzyme activity was expressed as U/L. For experiments involving CLP, analysis for creatinine and ALT in sera was done on Johnson & Johnson Vitros 250 Chemistry Analyzer.

BALB/c mice were injected with rPPE18 (100 μg) or PBS and 1 h later infected with 2.5 × 10⁸ or 5.0 × 10⁸ E. coli BL21 CFU. After 24 h, clinical symptoms, such as conjunctivitis, ruffling of fur coats, and activity on stimulation, were observed in a group of 10 mice and graded as described earlier (29). A collective grade was assigned to each group depending on the severity of the symptom observed in the majority of the animals. To evaluate the effect of rPPE18 on survival of mice, E. coli BL21–infected mice were administered rPPE18 1 h pre- or postinfection. Survival of mice was monitored every 6–8 h. Time of death was noted as accurately as possible.

The nonparametric Mann–Whitney two-tailed test was performed to determine significance. The log rank (Mantel–Cox) test was performed to determine significant differences in survival curves. The p values <0.05 were considered significant.

Earlier work had demonstrated that rPPE18 reduces LPS-induced TNF-α levels in human macrophages (22). Similar to human macrophages, rPPE18 inhibited TNF-α induction in mouse macrophages activated with LPS in vitro in a dose-dependent manner (Supplemental Fig. 1A). Also, the suppression in TNF-α was observed at both early (3 h) and late (24 h) time points postactivaton with LPS (Supplemental Fig. 1B). The effects were found to be specific to rPPE18 as polymyxin B–treated BSA failed to do so (Supplemental Fig. 1C). Next, we studied if rPPE18 by virtue of its ability to reduce TNF-α could be efficacious in reducing severity of sepsis. To test this hypothesis, we employed an animal model of peritonitis, where mice were injected i.p. with high doses of Gram-negative E. coli BL21 bacteria (30). Therefore, we next studied the effects of rPPE18 on E. coli BL21–induced sepsis. Peritoneal administration of high numbers of E. coli BL21 is known to result in systemic inflammation (24, 25) and we also found that i.p. injection of 2.5 × 10⁸ E. coli BL21 CFU resulted in a rapid rise in the serum TNF-α levels by 3 h (Fig. 1A) and dipped by 24 h (Fig. 1A). However, at both the early and late time points tested, mice treated with rPPE18 before induction of septicemia showed significant reduction in TNF-α (Fig. 1A). Similar reduction in TNF-α levels upon rPPE18 treatment was also observed in peritoneal lavage (Fig. 1B). Use of denatured rPPE18 could not significantly reduce TNF-α levels in septic mice confirming that the observed effect was indeed due to signals triggered by rPPE18 and not by some contaminant in the protein preparation (Fig. 1A, 1B). IL-1β levels were also measured in infected mice either left untreated or treated with rPPE18. IL-1β is a proinflammatory cytokine whose levels rise during peritonitis and LPS-induced septicemia (24, 31, 32). Predictably, peritoneal infection with E. coli BL21 resulted in elevated serum levels of IL-1β which were reduced upon pretreatment with rPPE18 in both serum and peritoneal lavage (Supplemental Fig. 2A, 2B). Another cytokine that is routinely studied as a marker for inflammatory responses is IL-6, levels of which are elevated during sepsis (33). Treatment with rPPE18 before induction of sepsis resulted in decreased IL-6 induction in sera and peritoneal lavage (Supplemental Fig. 2C, 2D). Pretreatment with rPPE18 was able to inhibit the proinflammatory cytokine storm not only at 2.5 × 10⁸ CFU but also at a higher dose of 5.0 × 10⁸ E. coli BL21 CFU (Supplemental Fig. 2E–G) as observed by reduction in serum levels of IL-1β (Supplemental Fig. 2E), IL-6 (Supplemental Fig. 2F), and TNF-α (Supplemental Fig. 2G) at 3 h postinfection.

To assess the effect of rPPE18 on sepsis-induced organ damage, levels of serum ALT, a marker for liver damage, were determined (18, 19, 32). As expected, injection of 2.5 × 10⁸ E. coli BL21 resulted in exacerbated liver injury as indicated by elevated serum ALT activity at 24 h postinfection. However, mice which had received rPPE18 prior to i.p. injection of E. coli BL21 had significantly reduced levels of serum ALT indicating protection against sepsis-induced liver damage (Fig. 1C). The severity of liver damage as reflected in elevated serum ALT levels was also observed in histological sections of tissue obtained from septic mice. The H&E-stained sections of the liver (Fig. 1D) and spleen (Fig. 1E) from mice infected i.p. with 2.5 × 10⁸ E. coli BL21 CFU showed severe lesions and damage resulting from severe sepsis. In contrast, mice which had received rPPE18 prior to injection with 2.5 × 10⁸ E. coli BL21 CFU had almost normal tissue architecture (Fig. 1D, 1E). Even at a high dose of 5.0 × 10⁸ E. coli BL21 CFU–induced sepsis, rPPE18 protein was able to inhibit ALT levels in sera (Supplemental Fig. 3A), as well as spleen and liver organ damage as assessed by histopathological analysis of tissue sections (Supplemental Fig. 3B, 3C).

Severity of liver and spleen damage ascertained by measurement of ALT activity and observation of H&E-stained tissue sections was found to correlate with external clinical features of septic mice. Mice suffering from sepsis exhibit clinical features which are reflective of the shock that they are experiencing. The clinical features such as conjunctivitis, ruffling of hair coat, huddling, and activity upon stimulation can be graded to reflect the severity of sepsis (29). Groups of 10 mice were examined at 24 h postinjection of either 2.5 × 10⁸ or 5.0 × 10⁸ E. coli BL21 CFU. Comparisons were made between groups which either received PBS or rPPE18 prior to administration of E. coli BL21. A grade of normal, mild, marked, and severe was given depending on the severity of the clinical symptoms. The observations (Supplemental Fig. 2H) and grades are summarized in Table I, which clearly indicate that mice pretreated with rPPE18 are clinically healthier with reduced sepsis symptoms. However, rPPE18 by itself had no effect on the kinetics of bacterial growth in vitro (Supplemental Fig. 2I) indicating that the observed ameliorative effect of rPPE18 is due to its ability to modulate the host immune responses.

Macrophages are innate cells that are exposed to various microbes and microbial products (34). The varying stimuli that act on macrophages can trigger different activation programs leading to generation of subsets of macrophages with specific phenotypes and functions (35). LPS and IFN-γ stimulation results in generation of M1 macrophages that are inflammatory and produce cytokines such as TNF-α, IL-1β, IL-6, and IL-12. On the other hand, cytokines such as IL-4, IL-10, and IL-13 activate macrophages through the alternate pathway giving rise to M2 macrophages (36, 37). Elevated numbers of M1 macrophages and high levels of M1-derived cytokines are correlates of sepsis severity (38, 39). On the other hand, M2 macrophages which produce anti-inflammatory mediators and promote tissue regeneration (40) are crucial for resolution of sepsis (38, 39). M1 macrophages have high NOS2 levels, whereas M2 macrophages have high arginase-1 levels (41). Because rPPE18 inhibits production of proinflammatory cytokines such as TNF-α, IL-1β, and IL-6 (Fig. 1, Supplemental Fig. 2), we next explored the possibility of rPPE18 promoting the polarization of macrophages toward the M2 phenotype. To determine the effect of rPPE18 on polarization of macrophages, initially levels of arginase-1 and NOS2 in mouse macrophages activated (in vitro) with LPS in the absence or presence of rPPE18 were compared. Indeed, NOS2 and arginase-1 were found to be reciprocally regulated in macrophages treated with rPPE18 (Fig. 2A). In LPS-activated macrophages, rPPE18 increased arginase-1 and decreased NOS2 in a dose-dependent manner (Fig. 2A). Also, decreased NOS2 levels correlated with reduced NO production in LPS-activated macrophages treated with rPPE18 when compared with macrophages activated with LPS alone (Fig. 2B). As compared with LPS-treated macrophages, arginase activity was higher in rPPE18 plus LPS–treated macrophages (Fig. 2C) which correlated with increased arginase-1 protein levels in these cells (Fig. 2A).

Previous work in our laboratory demonstrated that rPPE18 after binding to TLR2 increased phosphorylation of p38 MAPK, which in turn was responsible for downstream activation of IL-10 (21). The phosphorylation of p38 MAPK was reconfirmed in our system of study (Fig. 2D). Addition of rPPE18 to thioglycollate-elicited mouse peritoneal macrophages resulted in increased p38 MAPK phosphorylation (Fig. 2D, lane 3). Addition of LPS also led to increased p38 MAPK phosphorylation (Fig. 2D, lane 2). However, in the presence of rPPE18, phospho-p38 MAPK levels were greater in LPS-activated cells (Fig. 2D, lane 4). Next, we examined whether activation of p38 MAPK by rPPE18 was crucial for the elevated arginase-1 levels in these cells. For this purpose, SB203580, a specific p38 MAPK inhibitor, was used. The decrease in arginase-1 with SB203580 in cells treated with just LPS was not very evident (Fig. 2E, lanes 5 and 6). However, pretreatment with SB203580 resulted in reduction of arginase-1 levels in cells administered both rPPE18 and LPS (Fig. 2E, lanes 8 and 10). These results suggest that the rPPE18-mediated increase in arginase-1 level is due to downstream signaling events triggered by phosphorylation of p38 MAPK.

So far, it was observed that rPPE18 decreased NOS2 expression and increased arginase-1 expression and its activity in LPS-stimulated mouse peritoneal macrophages (Fig. 2A–C) indicating that rPPE18 could cause generation of M2 macrophages in vitro. Apart from increased arginase-1 levels, M2 macrophages secrete higher amounts of IL-10 (42), whereas M1 macrophages secrete higher amounts of IL-12 (35, 43, 44). In the next experiment, therefore, effect of rPPE18 on macrophage polarization in the in vivo system of sepsis was tested. IL-10 and IL-12 p70 levels were measured in sera of mice which either received PBS or rPPE18 before being injected with 2.5 × 10⁸ or 5.0 × 10⁸ E. coli BL21 CFU. IL-10 levels were higher (Fig. 3A), whereas IL-12 p70 levels were lower (Fig. 3B) in sera of septic mice which had received rPPE18. Next, levels of NOS2 and arginase-1 in the peritoneal exudate cells harvested from these mice were measured. Mice that received a combination of PBS and E. coli BL21 showed higher NOS2 levels compared with mice which received rPPE18 and E. coli BL21 (Fig. 3C, upper panel). Conversely, arginase-1 levels were higher in cells from mice which received a combination of rPPE18 and E. coli BL21 CFU compared with cells from mice which received PBS and E. coli BL21 CFU (Fig. 3C, middle panel). Even though arginase-1 expression was evident in peritoneal macrophages harvested 4 d post-thioglycollate administration (Fig. 2A, 2E), it was not detected in peritoneal cells harvested from E. coli–treated but rPPE18-untreated mice (Fig. 3C). It has been reported that LPS stimulation and sepsis induces arginase-1 expression (45). However, there are also reports which show endotoxin-mediated arginase-1 suppression in macrophages (46). It is evident in our system of study that in peritoneal exudate cells from E. coli–treated and rPPE18-untreated mice, there is elevated NOS2 which could be the possible reason of repressed arginase-1 expression. Overall, these results suggest that rPPE18 leads to generation of IL-10^hiNOS2^loarginase-1^hi M2 macrophages in mice suffering from septicemia.

In sepsis, the proinflammatory cytokine surge is known to cause activation of coagulation leading to DIC (47). Development of DIC and its severity correlates with mortality in severe sepsis (48, 49). DIC is characterized by systemic activation of blood coagulation, which results in generation and deposition of fibrin, leading to microvascular thrombi in various organs, depleting the organ’s supply of essential nutrients and oxygen, thus contributing to multiple organ dysfunction syndrome (50). Due to extensive coagulation, finally a stage arrives where consumption and subsequent exhaustion of coagulation proteins and PLTs (from ongoing activation of coagulation) may lead to severe bleeding (48). Studies have shown that IL-6 can cause activation of coagulation (51, 52), whereas TNF-α and IL-1β inhibit the anticoagulation cascades (53, 54). Because rPPE18 inhibits IL-6, TNF-α, and IL-1β, responsible for disturbance of coagulation and anticoagulation signaling events, the effect of rPPE18 on DIC during sepsis was examined. For this, BALB/c mice were administered rPPE18 1 h before infection with 2.5 × 10⁸ E. coli BL21 CFU and after 12 h, blood was collected to evaluate coagulation parameters for DIC such as PT, aPTT, FBG, PLT count, and WBC count. Prophylactic administration of rPPE18 helped reduce clotting time as evident by decreased PT (Fig. 4A) and aPTT (Fig. 4B) values in the rPPE18-treated group as compared with PBS-treated group. Moreover, in mice infected with E. coli, rPPE18 increased FBG levels (Fig. 4C), PLT count (Fig. 4D), and WBC count (Fig. 4E). Therefore, rPPE18 inhibits DIC in septic mice.

Experiments carried out so far showed that pretreatment with rPPE18 could reduce severity of E. coli BL21–induced sepsis in terms of reducing levels of proinflammatory cytokines (Fig. 1A, 1B, Supplemental Fig. 2), reducing organ damage (Fig. 1C–E, Supplemental Fig. 3), and preventing DIC (Fig. 4A–E). Mice administered rPPE18 also appeared clinically healthier (Supplemental Fig. 2H, Table I). Whether all this translated to an ultimate survival advantage in mice which received rPPE18 remained to be investigated. Therefore, it was further examined whether rPPE18-pretreated mice survived better upon induction of sepsis. The mean survival time of mice injected with 2.5 × 10⁸ E. coli BL21 CFU was 35 h (Fig. 4F), whereas for mice injected with 5.0 × 10⁸ E. coli BL21 CFU, it was 24 h (Fig. 4G), indicating that both these doses of bacteria are lethal. However, mice given rPPE18 i.p. before administration of 2.5 × 10⁸ E. coli BL21 CFU were almost completely protected from sepsis-induced lethality. Their survival was 75% when the last death in the PBS-injected group was recorded at 37 h (Fig. 4F). Also, in mice which were given rPPE18 and a higher dose of 5.0 × 10⁸ E. coli BL21 CFU, the mean survival time was 46 h compared with 24 h for mice which did not receive rPPE18 (Fig. 4G). This shows that administration of rPPE18 eventually confers survival advantage during sepsis.

So far for our experiments we had used a system where rPPE18 was given to mice prophylactically, that is, before induction of septicemia by E. coli BL21. This helped in unraveling its anti-inflammatory properties, its ability to inhibit DIC in infected mice, and its ability to polarize macrophages toward the M2 phenotype. However, to have potential as a tool for treatment of sepsis, it is important to study if rPPE18 can reduce inflammation and DIC when administered therapeutically. For this, rPPE18 was given 1 h postinduction of sepsis and it was found that rPPE18 could significantly reduce levels of TNF-α (Fig. 5A), IL-1β (Fig. 5B), IL-12 p70 (Supplemental Fig. 4A), and IL-6 (Supplemental Fig. 4B) in mice infected with 2.5 × 10⁸ E. coli BL21, indicating that rPPE18 is a potent inhibitor of proinflammatory cytokines when administered therapeutically. Next, it was examined whether rPPE18 could also inhibit DIC when injected therapeutically. BALB/c mice were administered rPPE18 1 h postinfection with 2.5 × 10⁸ E. coli BL21 CFU and 12 h later, blood was collected and coagulation parameters for DIC were examined. Administration of rPPE18 therapeutically helped to reduce clotting time as evidenced by decreased PT (Fig. 5C) and aPTT (Fig. 5D) values in rPPE18-treated group as compared with PBS-treated septic mice. Interestingly, rPPE18 also increased FBG levels (Fig. 5E), PLT count (Fig. 5F), and WBC count (Fig. 5G) in mice infected with E. coli in comparison to PBS-treated septic mice. Thus, rPPE18 is able to prevent inflammation and DIC in septic mice, even when administered therapeutically.

Because rPPE18 reduced induction of both proinflammatory cytokines and DIC in mice infected with a higher dose of E. coli when administered therapeutically, we next investigated its effects on survival of these mice. Mice which were not administered rPPE18 died by 23 h postinduction of sepsis (Fig. 5H). However, therapeutic administration of rPPE18 at 1, 24, and 72 h postinjection of E. coli BL21 dramatically increased survival to 71% as observed at the last time point of 384 h (Fig. 5H). Thus, rPPE18 improves survival when administered therapeutically to septic mice.

In all our studies so far, we had used a system where i.p. injection of high doses of E. coli BL21 was used to induce sepsis. Administration of high doses of E. coli BL21 led to a rapid rise in TNF-α and resulted in 100% mortality by 40 h. This model allowed us to study the effects of rPPE18 and dissect its mechanism of action in sepsis. However, we next wished to study the effect of rPPE18 in polymicrobial sepsis induced by CLP, which is a more physiological model of sepsis (55, 56). In this model, the effects of rPPE18 were studied on TNF-α levels; WBC count; numbers of lymphocytes, monocytes, and neutrophils; liver and kidney function; and ultimately survival.

Polymicrobial sepsis induced by CLP increased levels of TNF-α in the peritoneal lavage, which were significantly reduced in mice that received rPPE18 therapy (Fig. 6A). We, however, could not measure TNF-α in blood serum. This is consistent with earlier reports (57, 58). CLP also resulted in dramatic reduction in numbers of lymphocytes (Fig. 6B), monocytes (Fig. 6C), and neutrophils (Fig. 6D). This reduction in numbers was reflected in the decrease in WBCs post–CLP-induced sepsis (Fig. 6E). Administration of rPPE18 restored lymphocyte (Fig. 6B), monocyte (Fig. 6C), and WBC (Fig. 6E) numbers comparable to sham-operated mice. CLP also resulted in reduction in blood neutrophil levels. However, neutrophil numbers were not significantly but only marginally higher in rPPE18-treated mice (Fig. 6D).

As CLP is known to cause liver and kidney damage (59, 60), we next assessed liver and kidney function by measuring serum ALT and creatinine levels. We observed an increase in serum ALT (Fig. 6F) and creatinine (Fig. 6G) levels 16 h post-CLP. Therapeutic administration of rPPE18 significantly reduced ALT and creatinine levels. This effect on ALT levels was also reflected in differences in liver pathology between the two groups of animals. Mice subjected to CLP but not treated with rPPE18 showed marked necrosis and liver damage compared with sham-operated mice, whereas rPPE18-treated CLP mice had more hypertrophy (Fig. 6H), which can be a nonadverse effect (61), and hypervacuolation, which may be a host protective response (62). These data suggest that rPPE18 can provide protection from polymicrobial sepsis–induced organ damage.

Finally, we studied the effect of rPPE18 on survival of mice subjected to CLP-induced polymicrobial sepsis. Mice that received PBS post-CLP had a median survival time of 22 h, whereas those that received rPPE18 had significantly improved survival. In the rPPE18 treatment group, 75% of mice were still surviving when the last death in the PBS group was registered at ∼60 h post-CLP (Fig. 6I). Mice were monitored until 16 d and the survival percentage in the rPPE18-treated group at this time was 70% (Fig. 6I). These data indicate that rPPE18 by virtue of its ability to reduce TNF-α levels and prevent organ damage can provide protection in polymicrobial sepsis. Importantly, results from these experiments validate previous observations made using the model of E. coli BL21–induced septic shock.

Sepsis is a condition described by systemic hyperinflammation induced because of excessive production of proinflammatory cytokines such as TNF-α, IL-1β, and IL-6. Its causes are uncontrolled bacteremia resulting from situations of pneumonia, peritonitis, and surgical procedures (63, 64). Specifically, in the majority of cases, infection is caused by Gram-negative bacteria and LPS from the outer membrane of the bacteria overstimulates the host immune response. Treatment for sepsis consists of eradication of infection through early and aggressive treatment with appropriate antibacterials. However, despite advances in the development of powerful antibiotics, sepsis is still life-threatening. Therefore, there remains a need for development of novel antisepsis therapies combining antibiotic therapy with the application of a nonsteroidal anti-inflammatory regimen.

In this study, septicemia was induced in mice by injection of very high doses (2.5 × 10⁸ and 5 × 10⁸ CFU) of Gram-negative E. coli BL21 i.p. Such high doses of bacteria were potent inducers of proinflammatory cytokines (TNF-α, IL-1β, and IL-6) and organ damage, which are hallmarks of septicemia. Also, apart from inducing a cytokine storm, high doses of E. coli BL21 resulted in 100% fatality by 40 h. Therefore, this is a robust model to study excessive inflammation and septicemia as has been reported previously (24, 25). In this system we tested the effect of rPPE18 belonging to the mycobacterial PE/PPE family which comprises several immunomodulatory proteins (65). We found that rPPE18 significantly reduces levels of TNF-α, IL-1β, and IL-6. Both IL-1β and IL-6 have been shown to be elevated during septicemia (31–33). However, the role of IL-6 in experimental sepsis models is controversial as IL-6 has both anti- and proinflammatory properties (66). Blockade of IL-6 has been shown to be beneficial in sepsis as well as other inflammatory diseases (66, 67) indicating a positive correlation between elevated IL-6 levels and sepsis severity (68, 69). Interestingly, in rPPE18-treated mice, we observed reduced coagulation time, indicated by decreased PT and aPTT, with a simultaneous increase in FBG, PLT, and WBC levels during infection with E. coli. Thus, rPPE18 may reduce the inflammation and thereby decrease toxicity of sepsis by inhibiting TNF-α, IL-1β, and IL-6 production and DIC. Pretreatment with rPPE18 indeed reduced sepsis-induced organ damage as studied both by release of ALT and histopathology. Also, these mice were clinically healthier and most importantly, rPPE18-treated septic mice had better survival rates.

In a separate mouse model of polymicrobial sepsis induced by CLP, rPPE18 reduced TNF-α, attenuated liver and kidney damage, prevented sepsis-induced depletion of monocytes and lymphocytes, and ultimately increased survival. The protective effects of rPPE18 were, therefore, demonstrated in two models of sepsis. Ability to reduce TNF-α and attenuate organ damage in CLP-induced sepsis correlates with increased survival as has been demonstrated previously (13, 14, 70–72). rPPE18 was able to consistently reduce elevated levels of TNF-α in E. coli BL21–induced and CLP-induced sepsis when administered therapeutically. The former is a model where LPS is likely to play a major role (73, 74) and the latter is a model of polymicrobial sepsis. rPPE18 works efficiently to reduce mortality in both model systems. Timely administration of antibiotics controls bacterial replication, but cannot undo the damaging effects of the systemic cytokine storm. A strategy of controlling bacterial multiplication along with inhibition of excessive proinflammatory cytokines and DIC by use of agents such as rPPE18 might be more effective in controlling human sepsis.

Macrophages are versatile cells. Their microbicidal function and their participation in the inflammatory response can have immense bearing on the outcome of septicemia (34). In our model of E. coli BL21–induced sepsis, we observed accumulation of cells with reduced NOS2 and elevated arginase-1 expression in the peritoneal cavity in rPPE18-treated mice. These are characteristics of M2 macrophages (35, 41, 75). M2 macrophages are immunomodulatory and have low microbicidal activity compared with M1 macrophages and participate in processes of wound healing (35, 44). In cases of septicemia, studies have shown that a balance of M1 and M2 macrophages may determine the severity of sepsis and survival of infected animals (38). Previous studies using experimental animal models have shown that one of the underlying mechanisms of protection from sepsis is the increased proportion of M2 macrophages (38, 45, 76–78). PPE18 is capable of acting as an immunomodulator in the recombinant form as well as when present as part of whole M. tuberculosis (21–23). Our earlier studies (21) and results presented in this study clearly show rPPE18 to be a highly potent inducer of IL-10, a cytokine responsible for polarization of macrophages toward the M2 phenotype and also a cytokine that is secreted by M2 macrophages. In addition, rPPE18 reduces levels of endotoxin- and infection-induced IL-12 and TNF-α, also a typical characteristic of M2 macrophages. Earlier, we demonstrated that rPPE18 reduces NF-κB signaling by upregulating SOCS3 protein resulting in decreased production of IL-12 and TNF-α (22). Incidentally, a deficiency in SOCS3 has been found to increase polarization of macrophages toward the M1 phenotype and inflammation (79). Apart from conditions of sepsis, arginase-1 has been found to be considerably increased during M. tuberculosis infection (80–82). By upregulating arginase-1, M. tuberculosis can counteract the bactericidal effect of NOS2/NO, and survive better inside macrophages (83). Therefore, activation of arginase-1 is an important virulence mechanism of M. tuberculosis. Upregulation of arginase-1 by M. tuberculosis is dependent on TLR2 and STAT-6 (80). rPPE18 is known to bind to TLR2 and induce IL-10 (21). We now know that rPPE18-mediated signaling indeed leads to activation of arginase-1 and probably increases polarization of macrophages toward the M2 phenotype. Although this mechanism might prove useful for M. tuberculosis survival as observed in a mouse model of infection (23), it can also be used to provide protection during E. coli–induced and CLP-induced septicemia. Use of rPPE18 in a model of sepsis not only helped us to identify a potential candidate for sepsis therapy but also provided novel insights into the mechanism of action of a mycobacterial virulence factor.

Clinical trials have shown benefits of use of anti–TNF-α Ab (11, 12); however, no major benefits of activated protein C and TLR4 antagonist eritoran were observed in sepsis treatment (84–86). Immune adjuvants such as IL-7, IL-15, and GM-CSF have emerged as potential therapies for sepsis (87). Immune modulators such as thymosin α 1 (88) and curcumin (89–91), anticoagulants such as heparin (92), the PI3K-Akt pathway inhibitor wortmannin, LY294002 that suppresses coagulation and inflammation (93), and i.v. Ig (94) have shown promise in animal studies or clinical trials. Also, novel targets for sepsis therapy such as regulatory receptors PD-1 and BTLA have emerged (87). rPPE18 inhibits both inflammation and DIC and is hence a promising therapeutic agent for treatment of sepsis. Interestingly, the N-terminal of rPPE18 spanning amino acids 1–179 is sufficient to bind to TLR2 and trigger SOCS3-mediated activation of IL-10 (21, 22). Future study of the therapeutic use of rPPE18 in sepsis could focus on the 179 aa long N-terminal or its shorter segments, which might be more biologically effective and easier to deliver. Individuals latently infected with M. tuberculosis or having a history of active tuberculosis disease or vaccinated with M. bovis bacillus Calmette–Guérin may harbor Abs to PPE18 that might hinder its future therapeutic use. In such situations peptide fragments which are as effective as the full length protein may be used. Another advantage of using rPPE18 as opposed to anti–TNF-α Ab is that the efficacy of an Ab is reliant on its ability to completely neutralize TNF-α, unlike agents such as rPPE18 that inhibit the generation of TNF-α.

Proteins from pathogens have been used as vaccines for many years. More recently, focus has shifted to exploiting virulence protein factors from bacteria and viruses for therapy of immune-related disorders, coronary syndromes, and so on (95, 96). A good case in point is the potential therapeutic use of viral Serp-1 and Serp-2 proteins, which have been shown to reduce vascular inflammation and T cell apoptosis in a murine model (97, 98). Results from our study show that rPPE18 is a potent inhibitor of proinflammatory cytokines and DIC when given both prophylactically and therapeutically in mouse models of E. coli BL21–induced and CLP-induced septicemia. Also, rPPE18 was found to be a potent inducer of arginase-1 and IL-10, markers for M2 macrophages. A therapy which inhibits the production of inflammatory cytokines by polarizing macrophages to the M2 phenotype perhaps has a better chance of success along with antibiotics. Thus, its ability to dampen harmful responses and elevate protective responses in sepsis makes rPPE18 a promising candidate for antisepsis therapy.

We thank Dr. Jayant P. Hole, Suman Komjeti, Sridhar Kavela, and Sravani Edula for providing help in animal experiment work.

A patent application based on the results described in this paper is being filed by the Centre for DNA Fingerprinting and Diagnostics, in which S.M. and A.A. are listed as inventors. The other authors have no financial conflicts of interest.