Post-Exposure Therapeutic Efficacy of COX-2 Inhibition against Burkholderia pseudomallei

Burkholderia pseudomallei is a Gram-negative, facultative intracellular bacillus and the etiologic agent of melioidosis, a severe disease in Southeast Asia and Northern Australia. Like other multidrug-resistant pathogens, the inherent antibiotic resistance of B. pseudomallei impedes treatment and highlights the need for alternative therapeutic strategies that can circumvent antimicrobial resistance mechanisms. In this work, we demonstrate that host prostaglandin E2 (PGE2) production plays a regulatory role in the pathogenesis of B. pseudomallei. PGE2 promotes B. pseudomallei intracellular survival within macrophages and bacterial virulence in a mouse model of pneumonic melioidosis. PGE2-mediated immunosuppression of macrophage bactericidal effector functions is associated with increased arginase 2 (Arg2) expression and decreased nitric oxide (NO) production. Treatment with a commercially-available COX-2 inhibitor suppresses the growth of B. pseudomallei in macrophages and affords significant protection against rapidly lethal pneumonic melioidosis when administered post-exposure to B. pseudomallei-infected mice. COX-2 inhibition may represent a novel immunotherapeutic strategy to control infection with B. pseudomallei and other intracellular pathogens.


Introduction
Development of new therapeutics effective against intracellular bacterial pathogens remains a high priority. In addition to the global impact of intracellular bacterial infections on public health, the alarming increase in multidrug resistant strains and the potential threat of biological attack with select agents, such as Burkholderia pseudomallei, highlight the urgent need for safe and effective therapies against this collective group of pathogens. B. pseudomallei is a Gram-negative, facultative intracellular bacillus and the causative agent of melioidosis, a disease associated with high morbidity and mortality in Southeast Asia and Northern Australia. Although melioidosis is not endemic in the United States, B. pseudomallei is classified as a Tier 1 select agent due to its ease of respiratory transmission, high mortality rate, multidrug resistance, and the absence of a protective vaccine [1]. Furthermore, malicious use of B. pseudomallei and B. mallei during World Wars I and II provides historical precedence for use of these agents as bioweapons and validates the need for post-exposure therapeutics that can be quickly administered to military personnel and civilians [2].
The inherent antibiotic resistance of B. pseudomallei limits chemotherapeutic options for melioidosis and the particular choice of antibiotic regimen has not been shown to impact mortality within the first 48 hours of hospitalization [3]. Current treatment requires intravenous administration of ceftazidime or meropenem, with or without trimethoprim-sulphamethoxazole (TMP-SMX), for two weeks of intensive phase therapy. The intensive phase of treatment may be extended up to eight weeks for deep-seated infections. Upon completion of this intensive phase, an eradication phase utilizing oral TMP-SMX or doxycycline for outpatient use is recommended for 8-12 weeks in order to prevent relapse. Despite this aggressive therapy, case fatality rates for severe melioidosis approach 40% in Thailand and 15% in Australia [4]. Therefore, it is necessary to develop new modalities of treatment that can replace or complement existing antibiotics to improve patient survival.
An appealing alternative as a first line therapeutic strategy is to enhance the host innate immune response during the early course of bacterial infection. In human trials, complementary use of granulocyte colony-stimulating factor improved the duration of survival for melioidosis patients with severe sepsis but did not decrease mortality rates [5]. In pre-clinical studies, treatment of BALB/c mice with cationic liposomal DNA complexes (CLDC) 24 h prior to intranasal B. pseudomallei challenge enhanced natural killer (NK) cell recruitment and afforded complete protection from a lethal infectious dose [6]. Similarly, treatment of BALB/c mice with the TLR9 agonist, CPG ODN, 48 h prior to B. pseudomallei infection led to significantly lower tissue bacterial burdens and improved overall survival [7,8]. Combining vaccination with CpG treatment that was given up to 18 h post-infection provided significantly greater protection against B. pseudomallei than either treatment alone, indicating that immune modulation with CpG can also enhance the efficacy of other countermeasures [9]. In contrast, post-exposure prophylaxis with CpG alone was not effective against B. pseudomallei, as initial control of bacterial growth appears dependent upon prior recruitment of inflammatory cells to the lung [8]. Since bacterial infection cannot be predicted, it is imperative to identify immunotherapeutics that can mediate protection when administered post-exposure.
In the present study, we identify the prostaglandin E2 (PGE 2 ) pathway as a novel therapeutic target during pneumonic melioidosis. PGE 2 is a potent lipid mediator derived from cyclooxygenase (COX) metabolism of the cell membrane fatty acid, arachidonic acid [10]. PGE 2 is produced in response to inflammation via the COX-2 enzyme and is a key mediator of immunopathology in chronic disease, autoimmunity, and cancer [10]. While PGE 2 -mediated immunoregulation is essential for maintaining homeostasis, its suppressive effects on innate and adaptive immunity may be counter-productive during infection. In this work, we demonstrate that B. pseudomallei rapidly induces macrophage COX-2 expression and PGE 2 production which establishes a permissive environment for B. pseudomallei intracellular persistence. Pulmonary infection with B. pseudomallei leads to increased concentrations of lung PGE 2 , and lung PGE 2 levels significantly correlate with disease progression in mice. Postexposure administration of a COX-2 inhibitor provides significant protection against lethal pulmonary challenge with B. pseudomallei. This is the first demonstration of a non-antibiotic post-exposure therapeutic that provides significant protection on its own against lethal pulmonary infection with B. pseudomallei. Therapeutic strategies targeting the PGE 2 pathway may delay disease progression in pneumonic melioidosis and afford a window of opportunity for antibiotic intervention and/or development of adaptive immunity. Furthermore, COX-2 inhibition may represent a novel and universal immunotherapeutic strategy against other intracellular pathogens.

Mice and Bacterial Challenges
Ethics Statement: Animal experiments were performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the Tulane University Institutional Animal Care and Use Committee (protocol number 4042). Six to eight week old, female BALB/c mice (Charles River) were maintained under pathogen-free conditions and fed sterile food and water ad libitum. Infections utilizing Bps were performed under Animal Biosafety Level 3 containment.
Burkholderia pseudomallei strain 1026b (BEI Resources) was used in this study. For infectious challenge, mice were anesthetized with Ketamine/xylazine (88 mg/kg) (Fort Dodge Animal Health). The bacterial inoculum contained 3610 3 cfu ( , 4 LD 50 ) suspended in 40 ml sterile saline and 20 mL was delivered to each nostril via pipet. Bacterial cfu were confirmed by plating the inoculum on LB agar. Euthanasia endpoints used in this study included loss of .20% body weight, hunched posture and decreased movement or response to stimuli, or paralysis. In a subset of experiments, mice were treated with the selective COX-2 inhibitor, (N-[2-(cyclohexyloxy)-4-nitrophenyl]-methanesulfonamide) (NS398) (Cayman Chemicals), 3 h post-exposure. Mice received 50 ml of NS398 (15 mg/kg) dissolved in DMSO or vehicle control (DMSO) by intraperitoneal injection. Treatments were repeated for two consecutive days. After euthanasia, tissues were removed, weighed and homogenized in 1 ml 0.9% sterile saline. Serial dilutions of tissue homogenates were plated on LB agar and bacterial cfu were counted after 2-4 days of incubation at 37uC.
In some experiments, cells were treated with 100 mM of NS398, 100 mM nor-N v -hydroxy-L-arginine (nor-NOHA; Cayman Chemical), or 1 mM PGE 2 (Sigma). PGE 2 was measured in cell culture supernatants and lung homogenates by competitive ELISA (Pierce). Nitric oxide was measured as its stable end product nitrite by Griess assay (Invitrogen).

TLR Pathway PCR Array
Fold-change in mRNA expression of 84 genes central to TLRmediated signal transduction and innate immunity were measured by PCR array following the manufacturer's protocol and data analysis software (SABiosciences).

Real Time-PCR
RT-PCR was conducted using an iCycler (BioRad) with iScript cDNA Synthesis Kit (BioRad). 1 mg of RNA was converted to cDNA following the manufacturer's protocol. 1 ml of cDNA was added to 12.5 mL of iQ SYBR Green Super Mix containing 350 nM of each forward and reverse primer. Primer sequences

Author Summary
Burkholderia pseudomallei is the etiologic agent of melioidosis, a severe disease endemic in Southeast Asia and Northern Australia. B. pseudomallei is also classified as a Tier 1 select agent due to the threat of malicious use of the organism. Treatment of melioidosis is complicated by the inherent multidrug resistance of B. pseudomallei, leading to high case fatality rates or disease relapse. New therapeutic strategies are urgently needed to improve patient survival and to protect against a deliberate release of B. pseudomallei. Immunotherapeutics that can enhance the host immune response and delay disease progression represent a significant area of research interest. A number of immunomodulatory agents delivered locally to the lung prior to B. pseudomallei infection have afforded significant protection against pulmonary disease in animal models of melioidosis; however, their protective capacity significantly wanes upon post-exposure administration. In this work, we identify the PGE 2 pathway as an immunotherapeutic target in pulmonary melioidosis and show that postexposure COX-2 inhibition provides significant protection against lethal B. pseudomallei lung infection in mice. Further research examining FDA-approved COX-2 inhibitors as post-exposure prophylaxis for B. pseudomallei is warranted, as this may represent a safe, affordable, and efficacious immunotherapeutic strategy.

Detection of COX-2 and Arginase
For Western blot, equal amounts of protein (50 mg) from cell lysates or lung homogenates were resolved by SDS-PAGE and transferred to nitrocellulose using an iBLOT (Invitrogen). Detection of COX-2 enzyme was performed using a 1:1000 dilution of rabbit polyclonal anti-COX-2 (Cell Signaling Technology), followed by peroxidase-conjugated, donkey anti-rabbit IgG (1:5000). Arg2 was detected using a 1:200 dilution of rabbit polyclonal anti-mouse Arg2 (sc-20151, Santa Cruz) and Arg1 was detected using a 1:500 dilution of Arg1 antibody 18351 (Santa Cruz). Mouse Arg2-transfected 293T cell lysate (Santa Cruz) was used a positive control for Arg2 and 5 mg of mouse liver extract was used a positive control for Arg1. b-actin, a protein loading control, was detected using 1:1000 dilution of polyclonal rabbit anti-mouse b-actin (Cell Signaling). Detection of bound antibodies was visualized by a chromogenic reaction using Opti-4CN Substrate (BioRad).

Statistical Analysis
Statistical analyses were performed using Prism 5.0 software (Graph Pad). Kaplan-Meier survival curves were compared by log-rank analysis. All other data were analyzed using a one-way or two-way ANOVA followed by the Bonferroni post-test to determine statistical differences between groups. p,0.05 was considered statistically significant. All data are representative of at least two independent experiments.

B. pseudomallei Rapidly Induces PGE 2 Production by Macrophages
B. pseudomallei is remarkable in its ability to establish chronic infection that can reactivate decades after the initial infection and yet virtually nothing is known regarding the mechanisms by which B. pseudomallei evades immune clearance [1]. In order to identify host cell signaling pathways that might contribute to B. pseudomallei intracellular persistence, we performed a Toll-like receptor (TLR) PCR array on J744A.1 macrophages infected with B. thailandensis. B. thailandensis is a commonly used biosafety level 2 surrogate organism for the study of B. pseudomallei and, with the exception of capsular polysaccharide, possesses all of the known B. pseudomallei virulence determinants such as Type 3 and Type 6 secretion systems [13][14][15][16][17]. Although B. thailandensis is 1,000-to 100,000-fold less virulent than B. pseudomallei in animal models, the organisms behave very similarly in vitro. B. thailandensis and B. pseudomallei induce pyroptosis in macrophages as early as 8 h post infection at a multiplicity of infection (MOI) 10 or greater [18]. In pilot experiments, we determined that J774A.1 macrophages infected with B. thailandensis at MOI 10 or 1 displayed 80% and 28% cytotoxicity, respectively at 8 h post-infection (not shown). Therefore, experiments utilizing J774A.1 macrophages or primary bone marrow-derived macrophages (BMDM) were limited to an eight hour experimental time course using B. thailandensis or B. pseudomallei at MOI 1 or lower (0.1).
Consistent with previous reports [19][20][21][22], B. thailandensis upregulated expression of TLR1 and TLR2 by two h postinfection, and increases in TLR1, TLR2, TLR3, TLR4, and TLR5 mRNA expression were observed by eight h post-infection (Supporting information, Table S1). No change in mRNA expression was observed for TLR6, 7, 8, or 9. One of the most striking changes in expression occurred in COX-2, the enzyme responsible for the production of PGE 2 . A rapid increase (430-fold) in COX-2 mRNA expression occurred by two h post-infection and further increased by .16,000-fold at eight h (Supplementary Table S1).
To confirm the TLR array results obtained for B. thailandensisinfected J774A.1 macrophages, BMDM were infected with B. thailandensis and B. pseudomallei (MOI 1) and COX-2 mRNA expression was measured by RT-PCR. B. thailandensis and B. pseudomallei both up-regulated COX-2 mRNA expression in BMDM to a similar extent (Fig. 1A), although the levels of mRNA expression were lower than that observed for B. thailandensis-infected J774A.1 cells and may reflect differences between immortalized and primary cell lines (Fig. 1A, Supplementary Table S1). COX-2 enzyme and its end product, PGE 2 , were also produced by macrophages in response to B. pseudomallei in a time-and dose-dependent manner (Fig. 1B-C). Since lipopolysaccharide (LPS) of Gram-negative bacteria is known to induce COX-2 and PGE 2 production, we evaluated whether the PGE 2 response of infected macrophages was simply a passive signaling event mediated by TLR4 recognition of LPS. Notably, heat inactivation of B. pseudomallei significantly abolished COX-2 and PGE 2 expression (Fig. 1A-C) indicating that viable bacteria and/or bacterial proteins are required for early PGE 2 production by macrophages.

PGE 2 Enhances B. pseudomallei Intracellular Survival
Because PGE 2 has been shown to suppress macrophage bactericidal mechanisms [23], we assessed the impact of COX-2 activation and PGE 2 production on B. pseudomallei intracellular survival using the selective COX-2 inhibitor, NS398. Preliminary dose-response experiments were conducted using 10 to 200 mM NS398 (not shown). BMDM treated with $100 mM NS398 demonstrated enhanced intracellular killing of B. pseudomallei compared to non-treated cells by six h post-infection ( Fig. 2A). To verify the specificity of NS398 and that endogenous PGE 2 is responsible for the suppression of bacterial killing, exogenous PGE 2 was added to NS398-treated cells. Addition of PGE 2 to the cell cultures restored B. pseudomallei intracellular survival ( Fig. 2A) confirming that PGE 2 promotes a favorable environment for B. pseudomallei.
Previous work has shown that macrophage bactericidal activity against B. pseudomallei is mediated to a large extent by reactive nitrogen species and to a lesser extent by reactive oxygen species (ROS) [24,25]. PGE 2 has been shown to suppress nitric oxide (NO) synthesis in Kupffer cells, hepatocytes, murine peritoneal macrophages, and RAW 264.7 murine macrophages [26]. Therefore, we evaluated the downstream effect of PGE 2 on the macrophage NO response to B. pseudomallei infection. Treatment of BMDM with the COX-2 inhibitor NS398 led to a significant increase in nitrite, the stable end product of NO (Fig. 2B). This effect was not drug-specific because similar results were obtained using the COX inhibitor, indomethacin (not shown). Conversely, the addition of exogenous PGE 2 to NS398-treated macrophages significantly reduced nitrite levels in B. pseudomallei-infected cells (Fig. 2B). This suggests that PGE 2 -mediated suppression of NO production may partially contribute to B. pseudomallei intracellular survival.

Arginase 2 Enhances Bps Intracellular Survival
We next examined the effect of endogenous PGE 2 production on the expression of iNOS, which is required for the synthesis of NO. We did not observe any significant change in iNOS mRNA expression in NS398-or PGE 2 -treated cells compared to controls infected with B. pseudomallei (Fig. 3A). This suggested that PGE 2 did not directly regulate iNOS in B. pseudomallei-infected cells and that other mechanisms were responsible for the reduced levels of NO.
Since the enzymes arginase 1 (Arg1) and 2 (Arg2) compete with iNOS for the substrate, L-arginine, we postulated that PGE 2 induction of arginase could alter the level of NO production during B. pseudomallei infection. PGE 2 induction of macrophage arginase promotes tumor cell growth by suppressing NO-mediated tumor cytotoxicity [27,28]. Arg1 expression was not detected after four h of B. pseudomallei-infection, but the expression of Arg2 was significantly increased (155-fold) in B. pseudomallei-infected BMDM (Fig. 3A). NS398-treated macrophages demonstrated a significant reduction in Arg2 expression while treatment with exogenous PGE 2 increased Arg2 expression by 376-fold (Fig. 3A). These data suggest that endogenous PGE 2 may interfere with NO production by enhancing Arg2 expression.
Modulation of the arginase pathway contributes to the intracellular survival of multiple pathogens, including Salmonella and Mycobacterium spp. [29]. To determine whether Arg2 directly interferes with NO production and enhances B. pseudomallei intracellular survival, we treated macrophages with the arginase inhibitor, nor-NOHA. A significant decrease in B. pseudomallei intracellular survival was observed in nor-NOHA-treated BMDM (Fig. 3B) and this corresponded to a significant increase in nitrite levels (Fig. 3C). Collectively, these results indicate that Arg2 expression promotes B. pseudomallei intracellular survival, in part, through suppression of macrophage NO synthesis.

PGE 2 is Produced during Bps Pulmonary Infection
Inhalational infection with B. pseudomallei is a natural route of exposure and represents the most likely route of infection in a deliberate biological attack [2]. In order to evaluate the role of PGE 2 during pneumonic melioidosis, genetically-susceptible BALB/c mice were challenged by the intranasal route with a lethal dose of B. pseudomallei (3610 3 cfu) [30]. Pulmonary infection with B. pseudomallei progressed rapidly in mice leading to greater than 20% weight loss by 72 h post-infection (Fig. 4). A significant increase in lung PGE 2 was observed by 72 h post-infection and significantly correlated with disease progression (p = 0.029 by Pearson statistical analysis) (Fig. 4). These results indicate that PGE 2 may play an important role in pneumonic melioidosis during the early stages of infection.

Protective Efficacy of COX-2 Inhibition against Pneumonic Melioidosis
Because PGE 2 inhibition enhanced bacterial clearance in vitro and because PGE 2 is elevated in B. pseudomallei-infected lungs, we evaluated the efficacy of COX-2 inhibition as a postexposure therapeutic strategy against lethal B. pseudomallei pulmonary challenge. Mice were given NS398 or mock control by i.p. administration three h after B. pseudomallei intranasal infection, and treatments were repeated for two consecutive days. Initiation of therapy within three h is clinically relevant in the case of a known biological exposure to B. pseudomallei, such as a laboratory accident. A daily maximum dose of 15 mg/kg of NS398 was selected based upon previously documented pharmacological efficacy in mice (particularly in reducing lung PGE 2 ) without any associated toxicity [31]. Mock-treated mice infected with B. pseudomallei rapidly displayed signs of pulmonary disease and all had to be euthanized within 72 h (Fig. 5). Lungs of mock-treated mice all contained greater than 10 6 cfu of B. pseudomallei at the time of euthanasia. In contrast, none of the NS398-treated mice showed signs of illness until day 5 postinfection. On day 5, one mouse in the NS398-treated group displayed hind leg paralysis and was humanely euthanized. This was observed again in another animal on day 7. No bacteria were recovered from the lungs of either animal. Intranasal infection of mice with B. pseudomallei often manifests in colonization of the brain with subsequent neurologic complications [30], and we believe that this, and not pulmonary disease, likely accounted for the animals' morbidity. By day 10, all of the remaining NS398-treated mice appeared to have recovered from the infection. NS398-treated mice showed no evidence of weight loss throughout the study (not shown). No bacteria were recovered from the lungs of NS398-treated mice at the study endpoint with the exception of one animal that contained 10 4 cfu. All of the mice were colonized with 20-100 cfu in the spleen and liver, indicating that bacterial dissemination from the lung had occurred. These results indicate that host PGE 2 production promotes the pathogenesis of B. pseudomallei during pneumonic melioidosis and that inhibition of COX-2 enhances bacterial clearance from the lung and improves host survival. Consistent with these findings, COX-2 inhibition also significantly reduced tissue bacterial burdens and pulmonary inflammation in mice infected with B. thailandensis (Supporting information, Figures S1, S2).

Effect of COX-2 Inhibition on Lung Arginase
B. pseudomallei infection led to increased PGE 2 and Arg2 expression in macrophages and both PGE 2 and Arg2 enhanced B. pseudomallei intracellular survival. Furthermore, PGE 2 positively regulated Arg2 expression in response to B. pseudomallei in vitro. We therefore evaluated Arg2 expression in the lungs of mice in response to bacterial infection and COX-2 inhibition. Similar to our in vitro observations, an increase in lung Arg2, but not Arg1, was observed in B. pseudomallei-infected animals compared to uninfected animals (Fig. 6). Upon COX-2 inhibition, a reduction in lung Arg2 was observed in B.
pseudomallei-infected mice as evident by Western blot and densitometry analysis (Fig. 6). These results corroborate our observations in murine macrophages and advocate a supporting role for Arg2 in PGE 2 -mediated immunosuppression during B. pseudomallei infection.

Discussion
Enhancement of non-specific innate immunity represents an attractive therapeutic strategy to combat infection with multidrug resistant bacterial pathogens. In this study, we demonstrate a critical role for PGE 2 in the early pathogenesis of B. pseudomallei pulmonary infection and identify the PGE 2 pathway as an immunotherapeutic target in melioidosis. PGE 2 can negatively regulate innate immunity by suppression of leukocyte activation [32], macrophage microbicidal activity [23], and NK cell function [33]. Over-production of PGE 2 has also been observed in a number of clinical conditions associated with an increased  susceptibility to bacterial infection, including AIDS [34]. Therefore, the results presented here may be applicable to other intracellular bacteria, particularly those that infect the lung. PGE 2 is a potent pro-inflammatory mediator in most tissues but plays an opposite role in the lung and gastric mucosa in order to limit inflammation and tissue injury upon mucosal insult [35,36]. PGE 2 concentrations in the lung are much higher than in plasma [37] so the bacterial survival advantage afforded by PGE 2 in the present study may be exclusive to pulmonary infection and requires further study.
Macrophages play an important role in early host defense against B. pseudomallei as macrophage-depleted mice display an accelerated mortality during experimentally-induced melioidosis [38]. Our in vitro studies demonstrated that endogenous and exogenous PGE 2 promoted B. pseudomallei intracellular survival, while inhibition of COX-2 eliminated endogenous PGE 2 production and restricted bacterial growth in macrophages. The therapeutic efficacy of COX-2 inhibition at the cellular level may account for its significant protective efficacy against pulmonary melioidosis when administered three h after B. pseudomallei challenge. A number of studies have shown the clinical effectiveness of various immunostimulants delivered intranasally to mice prior to B. pseudomallei pulmonary challenge but none have demonstrated significant protective efficacy as a stand-alone postexposure therapeutic [6,8,9,39]. For example, CpG ODN was ineffective when given as early as one h after intranasal B. pseudomallei challenge due to a delay in recruitment of inflammatory monocytes and neutrophils to the lung [8]. Unlike these studies, the COX-2 inhibitor was administered intraperitoneally to mice after B. pseudomallei pulmonary delivery, indicating that its efficacy does not rely upon local administration and subsequent inflammatory cell recruitment. We have not yet assessed the postexposure window of efficacy for COX-2 inhibition but we postulate that its direct action on infected macrophages, with less dependence on additional phagocyte recruitment, may allow a greater time frame for therapeutic intervention than we have already shown. PGE 2 -mediated suppression of macrophage bactericidal ability has been observed in other bacterial pulmonary infections and is not restricted to the mouse model. For example, pre-treatment of rat alveolar macrophages with the COX-1/2 inhibitor, indomethacin, or antagonists of the PGE 2 receptors, EP-2 and EP-4, augmented NADPH oxidase and ROS production and improved killing of Klebsiella pneumoniae [23]. It has been proposed that the timing and concentration of PGE 2 determines the macrophage bactericidal response to stimuli such as IFN-c and LPS [40]. Therefore, rapid production of high concentrations of PGE 2 in response to B. pseudomallei may suppress macrophage control of bacterial growth early in infection. While our results demonstrated that PGE 2 suppressed NO and enhanced Arg2, the mechanism(s) by which PGE 2 exerts its suppressive effects on B. pseudomalleiinfected macrophages may involve inhibition of additional reactive oxygen [23,41] and nitrogen species [42] or pro-inflammatory cytokines [43] and warrants further study. Nonetheless, our studies add B. pseudomallei to the growing list of intracellular pathogens that utilize host arginase to facilitate their survival [29]. PGE 2mediated immunosuppression through arginase induction is wellestablished in cancer [44]. Our study highlights a similar mechanism operating during an intracellular bacterial infection. The clinical benefit of COX-2 inhibitors against lung carcinoma [44] and other cancers [10,45] lends support to the potential use of this class of inhibitors against respiratory bacterial infection.
Interestingly, heat-inactivated B. pseudomallei did not induce appreciable levels of COX-2 expression or PGE 2 production by macrophages indicating that the response is not stimulated by bacterial LPS. Although bacterial flagellin has been shown to induce COX-2 expression through TLR5 recognition and p38 MAPK signaling [46], P. aeruginosa strains lacking flagellin induce COX-2 expression comparable to wild type strains [41]. In addition, the majority of our in vitro studies were performed using murine BMDM which have been shown to be unresponsive to flagellin due to an absence of TLR5 [47] [48]. Collectively, these findings suggest that the macrophage PGE 2 response is predominantly regulated by active bacterial processes as opposed to a passive signaling event mediated by bacterial PAMPs. In support of this view, COX-2 expression in the lungs of mice infected with Pseudomonas aeruginosa was dependent upon viable bacteria and the presence of the type three secretion system effector, ExoU, a member of the phospholipase A family [41]. Although the B. pseudomallei effector(s) responsible for the induction of PGE 2 remain to be identified, it is plausible that secreted bacterial phospholipases accelerate phospholipid release and turnover in the host cell leading to increased COX-2 and PGE 2 expression [49].
To our knowledge, this study is the first to characterize a postexposure immunotherapeutic that provides significant protection against lethal B. pseudomallei pulmonary infection in mice. In an experimental mouse model of tuberculosis, inhibition of PGE 2 reduced bacillary loads and increased granuloma formation, concomitant with increased IFN-c, TNF-c, and iNOS expression, suggesting that PGE 2 may contribute to M. tuberculosis persistence by down-regulation of cell-mediated immunity (CMI) [50]. Similar findings were reported for F. tularensis live vaccine strain (LVS) pulmonary challenge in mice [51]. Inhibition of PGE 2 reduced bacterial loads in the tissues and enhanced CMI responses. COX-2 2/2 and EP2 2/2 mice demonstrated accelerated clearance of Pseudomonas aeruginosa from the lungs compared to wild type mice, and PGE 2 signaling via EP2 suppressed macrophage ROS production in vitro [41]. COX-2 inhibition also improved bacterial clearance [32,41] and enhanced host survival [41] during intratracheal infection with P. aeruginosa. Collectively, these studies and ours suggest that PGE 2 production promotes bacterial pathogenesis in the lung and that inhibition of COX-2 may represent a broad-spectrum immunotherapeutic against multiple bacterial pathogens. These results compel further investigation of the role of PGE 2 in human melioidosis, particularly in patients with pneumonia. Use of commercially-available selective COX-2 inhibitors as an adjunct therapy to antibiotic treatment should also be explored in animal models of melioidosis as combination therapy may further eradicate persistent bacteria. Table S1 Fold-change in mRNA expression of 84 different genes from the Toll-like receptor pathway. J774A.1 macrophages were infected with B. thailandensis E264 (MOI 1) and gene expression was analyzed at 2 and 8 hours postinfection. Change in mRNA expression is represented as fold change over uninfected controls. n.c. indicates no change in expression.