Platelet-activating Factor Contributes to Bacillus anthracis Lethal Toxin-associated Damage*

Background: PAF has been implicated as a potent lipid mediator of endotoxin-induced sepsis, but its role in anthrax-associated shock is unknown. Results: Increased serum levels of PAF were present in LeTx-challenged mice. Inhibition of PAF activity prolonged survival ameliorated increased vascular permeability and hepatic necrosis. Conclusion: PAF appears to be a mediator in lethal toxin-associated damage. Significance: PAF antagonists may be helpful as adjunctive therapy for anthrax-associated shock. The lethal toxin (LeTx) of Bacillus anthracis plays a central role in the pathogenesis of anthrax-associated shock. Platelet-activating factor (PAF) is a potent lipid mediator that has been implicated in endotoxin-associated shock. In this study, we examined the contribution of PAF to the manifestations of lethal toxin challenge in WT mice. LeTx challenge resulted in transient increase in serum PAF levels and a concurrent decrease in PAF acetylhydrolase activity. Inhibition of PAF activity using PAF antagonists or toxin challenge of PAF receptor negative mice reversed or ameliorated many of the pathologic features of LeTx-induced damage, including changes in vascular permeability, hepatic necrosis, and cellular apoptosis. In contrast, PAF inhibition had minimal effects on cytokine levels. Findings from these studies support the continued study of PAF antagonists as potential adjunctive agents in the treatment of anthrax-associated shock.

The lethal toxin (LeTx) of Bacillus anthracis plays a central role in the pathogenesis of anthrax-associated shock. Plateletactivating factor (PAF) is a potent lipid mediator that has been implicated in endotoxin-associated shock. In this study, we examined the contribution of PAF to the manifestations of lethal toxin challenge in WT mice. LeTx challenge resulted in transient increase in serum PAF levels and a concurrent decrease in PAF acetylhydrolase activity. Inhibition of PAF activity using PAF antagonists or toxin challenge of PAF receptor negative mice reversed or ameliorated many of the pathologic features of LeTx-induced damage, including changes in vascular permeability, hepatic necrosis, and cellular apoptosis. In contrast, PAF inhibition had minimal effects on cytokine levels. Findings from these studies support the continued study of PAF antagonists as potential adjunctive agents in the treatment of anthrax-associated shock.
Severe respiratory distress, shock, multiorgan dysfunction, and bleeding are characteristics of anthrax. Despite the availability of effective anti-microbial therapy, the morbidity and mortality of this disease remains exceedingly high (1), possibly reflecting the action of tissue-damaging toxins. Lethal toxin (LeTx), 3 edema toxin, and anthrolysin O have each been shown to contribute to Bacillus anthracis virulence, though LeTx is considered particularly important (reviewed in Ref. 2). LeTx is a Zn 2ϩ -dependent endoprotease that cleaves MAPK kinases and alters cell signaling. In vitro, LeTx challenge induces rapid lysis of macrophages from susceptible mouse strains (3,4). Furthermore, LeTx challenge reproduces many of the clinical features of anthrax in animal models, and the administration of toxin specific antibody significantly reduces the mortality of experimentally B. anthracis-infected animals (5-7).
Platelet-activating factor (PAF) is a potent lipid mediator that was originally described in the context of its ability to alter platelet function (8). PAF is produced in response to stimuli by a variety of cell types, including monocytes/macrophages, polymorphonuclear leukocytes, eosinophils, basophils, platelets, mast cells, vascular endothelial cells, and lymphocytes. This lipid mediator exerts diverse biologic effects and has been implicated in several pathologic conditions including systemic inflammatory response and shock. Elevated serum PAF levels have been reported in septic patients (9,10), and administration of PAF to animals reproduces many features of shock (11). PAF is rapidly inactivated by serum PAF acetylhydrolase (PAF-AH), and decreased levels of PAF-AH have been noted in patients with anaphylactic and septic shock (12)(13)(14). PAF antagonists have been studied as potential therapeutics in endotoxin-mediated shock (reviewed in Ref. 15), although initial results have been equivocal (16). Given the overlap in the clinical syndromes induced by LeTx and PAF, we sought to determine whether PAF contributes to the pathologic disturbances induced by LeTx. used in PCR with 2.5 mM deoxynucloside triphosphate and 20 M each primer under the following conditions with Taq polymerase Gold (Applied Biosystems, Foster City, CA): 95°C for 10 min, 95°C for 1 min, 55°C for 1 min, and 72°C for 2.5 min for 33 cycles. All animal studies were carried out with protocols approved by the Albert Einstein College of Medicine Animal Care and Use Committee.
B. anthracis and Toxin Components-B. anthracis Sterne 34F2 (pXO1ϩ, pXO2Ϫ) was obtained from Dr. Alex Hoffmaster at the Centers for Disease Control (Atlanta, GA). Bacterial cultures were grown from frozen stock in brain-heart infusion broth (Difco, Detroit, MI) at 37°C for 18 h with shaking. Recombinant protective antigen (PA) and lethal factor (LF) proteins and endotoxin-reduced PA and LF were obtained from the Northeast Biodefense Center Expression Core of the NYS Department of Health (Albany, NY). Briefly, histidine-tagged PA and LF were expressed in Escherichia coli and purified by affinity chromatography using a ready to use column prepacked with precharged high performance nickel-Sepharose (HisTrap HP) (GE Life Sciences). Proteins were further purified by ion exchange (Mono Q) chromatography (GE Life Sciences). LPS measurements on these preparations revealed levels of Ͻ12.3 endotoxin units/ml. To further reduce endotoxin, proteins were purified by affinity chromatography using Endotrap Blue resin (Hyglos, Chandler, NC), which significantly reduced LPS levels (0.023 enzyme unit/mg). Studies done with the endotoxin reduced and nonendotoxin reduced preparations gave comparable results. All proteins were quantitated using the colorimetric Bradford reagent (ThermoScientific Pierce). SDS-PAGE analysis revealed more than 95% of the protein in one band at molecular masses of 83 kDa (PA) and 89 kDa (LF).
PAF Antagonists-CV3988, WEB 2086, and quinacrine were solubilized in ethanol and diluted in either PBS or normal saline and administered at doses of 3 and 5 mg/kg. Ginkgolide B was solubilized in DMSO, diluted in PBS, and administered at a dose of 5 mg/kg. CV3988 and WEB 2086 are competitive PAF receptor antagonists. Ginkgolide B accelerates PAF degradation by promoting PAF-AH I ␣ 2 homodimer activity and quinacrine inhibits PAF synthesis. For hematocrit studies, PAF antagonists were administered at 5 mg/kg intravenously 1 h prior to toxin challenge. All antagonists except quinacrine (Sigma) were obtained from Enzo Life Science (Farmingdale, NY).
Macrophage Depletion-Dichloromethylene bisphosphonate (CL 2 MBP), also known as clodronate, was a gift from Roche and was encapsulated in liposomes as previously described (41). Liposome clodronate selectively depletes macrophages after intravenous administration (5,41). Clodronate liposomes and PBS liposomes were given to WT mice (n ϭ 6 per group) ϳ48 h prior to toxin challenge. To confirm macrophage depletion, mice (n ϭ 3 per group) were given 0.1 ml of clodronate liposomes or PBS liposomes intravenously. Two days later, the mice were sacrificed, the spleens and livers were removed, and cells were prepared for FACS analysis. Briefly, the cells (10 6 ) were stained for 30 min on ice with 100 l of the following antibodies diluted in staining buffer (1% FCS/PBS): 2 g/ml of R-phycoerythrin-labeled anti-CD45 and 5 g/ml of FITC-labeled anti-mouse MAC-3 (Pharmingen, San Diego, CA). The samples were washed twice in staining buffer and fixed in 1% paraformaldehyde. Stained samples were stored in the dark at 4°C overnight and analyzed on a Calibur FACscan flow cytometer (Becton Dickinson, Mountainview CA) using the CELLQuest (Becton Dickinson) software. Live cells were gated as judged from forward and side laser scatter and CD45 ϩ cells. Controls consisted of isoptype-matched irrelevant antibodies.
Survival Studies-WT and PAFr Ϫ/Ϫ mice (n ϭ 10 per group) were injected into the tail vein with 120 g of PA and 50 g of LF in 100 l of PBS as described (7). For some experiments, mice (n ϭ 10 per group) were infected intravenously with 10 6 B. anthracis Sterne bacterial cells. For some experiments, WT mice were treated with 3 mg/kg CV3988 or WEB 2086 (n ϭ 5 per group) 2 h prior to LeTx injection. Control mice received PBS (n ϭ 5 per group). The mice were monitored daily for mortality.
PAF Measurements-WT mice were challenged with LeTx (120 g of PA and 50 g of LF) intravenously and euthanized at 30 min, 2 h, and 16 h. Mice were bled from the retroorbital sinus, and serum was collected and stored at Ϫ20°C until tested. The mice were then sacrificed, and the liver was removed and homogenized in 2 ml of PBS in the presence of protease inhibitors (Complete Mini; Roche Applied Science). Homogenates were centrifuged at 2000 ϫ g for 10 min to remove cell debris, and the supernatant was frozen at Ϫ80°C until tested. PAF measurements were using ELISA kit for PAF (Cedarlane Laboratories (USCN Life Science), Burlington, NC) as per the manufacturer's instructions. Briefly, samples were added into 96-well plate, 50 l of detection reagent A was then added, and the plate was incubated for 1 h at 37°C. ELISA plate was then washed 3ϫ with wash solution, and 100 l of detection reagent B was added. The plate was then incubated for 30 min at 37°C, washed 3ϫ with wash solution, and 90 l of substrate solution was then added. The plates were again incubated for 25 min at 37°C. Stop solution (50 l) was added to each well and immediately read at 450 nm (Labsystems Multiskan, Franklin, MA). Average values were obtained, and calculations of results were done based on the standard curve.
PAF Acetylhydrolase Measurements-PAF-AH activity was measured as per the manufacturer's instructions (Cayman, Ann Arbor, MI). Briefly, 10 l of 5,5Ј-dithiobis(nitrobenzoic acid), 10 l of sample, and 5 l of PAF-AH assay buffer were added to a 96-well plate. The reactions were initiated by adding 200 l of 2-thio PAF substrate solution. Absorbance was read every minute for 10 min at 405 nm (Labsystems Multiskan, Franklin, MA) to obtain 10 time points.
Histology-WT and PAFr Ϫ/Ϫ mice were evaluated histologically after intravenous PBS or LeTx treatment at 2 h (n ϭ 3 per group) and 24 or 48 h (n ϭ 5 per group). The mice were euthanized at 2, 24, or 48 h post-LeTx injection. Lung, liver, and spleen were isolated and fixed in 10% neutral buffered formalin (Fisher Scientific). One animal from each the 24-h and 48-h dose groups had a full tissue evaluation (liver, kidney, spleen, heart, lungs, adrenal glands, bone marrow, thymus, brain, skeletal muscle, nerve, small and large intestines, bladder, pancreas, submandibular salivary glands, and lymph nodes). Tissues were processed for paraffin embedding, and histological sections of 5 m were stained with hematoxylin and eosin. Sections were evaluated by a board-certified veterinary pathologist in a blinded manner.
Apoptosis Studies-Apoptosis was studied by examination of nuclear and cellular morphology in tissue sections stained with hematoxylin and eosin. Additional tissue samples were stained for cleaved caspase 3. Briefly, 5-m sections were deparaffinized in xylene followed by graded alcohols. Antigen retrieval was performed by incubating sections in 10 mM sodium citrate buffer (pH 6.0) and heated to 96°C for 20 min. Endogenous peroxidase activity was quenched using 3% hydrogen peroxide in PBS for 10 min. Sections were blocked with 5% normal donkey serum and 2% BSA in PBS for 1 h. The primary antibody to cleaved caspase 3 (Cell Signaling, Danvers, MA) was used at 1:50 for 1 h at room temperature. The primary species (rabbit IgG1) was substituted for the primary antibody to serve as a negative control. Sections were stained by routine immunohisthochemistry methods using HRP polymer conjugate (Invitrogen) to localize the antibody bound to antigen with diaminobenzidine as the final chromogen. All immunostained sections were lightly counterstained with hematoxylin.
Pulmonary Function Analysis-Whole body plethysmography (Buxco Electronics, Inc., Wilmington, NC) was used to measure respiratory parameters including tidal volume, respiratory rate, inspiratory, and expiratory time by monitoring the box flow pattern created by the animal's respiration (18 -20). Unrestrained mice (n ϭ 4 -5 per group) were placed in an enclosed chamber, and baseline readings were taken over a period of 3-5 min before intravenous injections of LeTx. Additional lung function measurements were taken at various time intervals after LeTx injection.
Hematocrit Measurements-WT and PAFr Ϫ/Ϫ mice (n ϭ 5-6 per group) were challenged with LeTx or PBS intravenously and then bled from the retroorbital sinus using heparinized microhematocrit capillary tubes (Fisher Scientific). Tubes were sealed with Hemato-Seal (Fisher Scientific) and spun in microcapillary centrifuge (Fisher Scientific) for 5 min. Hematocrits were measured by determining the red blood cell volume relative to the volume of whole blood.
Evans Blue Extravasation-A 100 l of intraperitoneal injection of 0.22-m filtered 10 mg/ml Evans blue PBS was administered 45 min prior to LeTx injection. For some experiments, WEB 2086 (3 mg/kg) was administered 2 h prior to LeTx injection. The mice were sacrificed and perfused with 30 ml PBS. Lungs were removed and homogenized in 1.5 ml of PBS. To extract the dye, TCA (60%) was added to each sample, vortexed, and centrifuged (1000 ϫ g) for 30 min at 4°C. Optical densities of the supernatants were measured at 620 nm (Labsystem Multiskan, Franklin, MA).
Biochemical Profile-After challenge with LeTx (24 h postchallenge), WT and PAFr Ϫ/Ϫ mice (n ϭ 5 per group) were bled from the retroorbital sinus by use of heparinized microhematocrit capillary tubes (Fisher Scientific). Serum was then separated and sent to a commercial veterinary laboratory (Antech Diagnostics, Lake Success, NY) for standard mammalian chemistry and liver function tests.
Cytokine and Chemokine Studies-WT and PAFr Ϫ/Ϫ mice (n ϭ 6 per group) were sacrificed 2 and 24 h post-intravenous injection of LeTx or PBS. The mice were sacrificed, and the lungs were homogenized in 2 ml of PBS in the presence of protease inhibitors (Complete Mini; Roche Applied Science). Homogenates were centrifuged at 2000 ϫ g for 10 min to remove cell debris, and the supernatant was frozen at Ϫ80°C until tested. Supernatants were assayed using mouse cytokine protein array I (Ray Biotech, Norcross, GA) as per the manufacturer's instructions. Briefly, membranes were blocked with 1ϫ blocking buffer, washed three times, and then incubated with samples. Membranes were washed again and incubated for 1 h with biotin-conjugated cytokines, which were detected by incubation with HRP-conjugated streptavidin. All incubations were done at 37°C for 1 h. Unbound reagents were removed by washing and the membranes developed. This kit assays for the following cytokines: GCSF, GM-CSF, IL-2, IL-3, IL-4, IL-5, IL-6, IL-9, IL-10, IL-12, IL-13, IL-17, IFN-␥, MCP-1, MCP-5, RANTES, SCF, sTNFRI, TNF-␣, and thrombopoietin. Supernatants were also assayed for IL-2, IL-4, IL-6, IL-10, MCP-1, and MIP-1␣ concentrations using ELISA kits (Pharmingen, San Diego, CA and R&D Systems Inc., Minneapolis, MN). The detection limits of cytokine assays are 3.1 pg/ml for IL-2, 7.8 pg/ml for IL-4, 15.6 pg/ml for IL-6 and TNF-␣, and 31.3 pg/ml for IL-10 and IFN-␣ as stated by the manufacturer. The detection limits of the chemokine assays are 4.7 pg/ml for MIP-1␣ and 15.6 pg/ml for MCP-1 as determined by the manufacturer.
Macrophage and Hepatocyte Isolation-Peritoneal macrophages were isolated from mice (PAFr Ϫ/Ϫ and WT) by peritoneal lavage without prior peritoneal irritation. Briefly, the abdominal cavity was washed five times with sterile Hanks balanced salt solution, 1% penicillin-streptomycin, 0.1 mM EGTA (Fisher Scientific) using a sterile Pasteur pipette. Following centrifugation, erythrocytes were lysed by resuspension in ice-cold 0.17 M NH 4 Cl for 10 min. A 10-fold excess of DMEM solution was then added to make the solution isotonic, the cells were collected by centrifugation, and live cells (trypan blue exclusion and morphological examination) were counted in a hemocytometer chamber. The cells were suspended in DMEM (Invitrogen), 10% NCTC-109 medium, 1% penicillin-streptomycin, and 1% nonessential amino acids (Mediatech, Inc, Manassas, VA). Cells were plated at a density of 1 ϫ 10 6 cells/well in a 96-well tissue culture plate and incubated overnight at 37°C. Primary mouse hepatocytes were isolated from WT and PAFr Ϫ/Ϫ mice as described with minor modifications (21).
MTT Assay-MTT (Sigma) assay was used to determine LeTx toxicity to peritoneal macrophages and primary hepatocytes as described (22). Macrophages (10 6 ) and hepatocytes (5 ϫ 10 4 ) per well were incubated in a 96-well plate with LeTx for 4 h at 37°C. Dose-response experiments were done. All wells were plated in duplicates. Macrophage supernatants were removed for PAF measurements as described above. A 25-l volume of 5 mg/ml stock-solution of MTT (in sterile PBS) was then added to each well and after 2 h of incubation of 37°C, 100 l of the extraction buffer (12.5% SDS, 45% DMF) was added, and cells were incubated overnight at 37°C. Optical densities were measured at 570 nm (Labsystem Multiskan, Franklin, MA).
Statistics-For parametric data, individual comparisons were done using a Student's t test. Multiple comparisons were done with an ordinary one-way analysis of variance, and post hoc analyses were done with a Dunnett multiple comparison test. For nonparametric data, individual comparisons were done with a Mann-Whitney test. Multiple comparisons were done using the Kruskal-Wallis test, and post hoc analyses were done with the Dunn test. Survival curves were compared by log rank analysis (Graphpad Prism; GraphPad Software, Inc., La Jolla, CA). All variance listed is standard deviation.

RESULTS
PAF and PAF-AH Activity-Serum and liver levels of PAF were measured several times after injection of LeTx. At 30 min, serum from WT mice challenged with LeTx contained higher amounts of PAF (48 ng/ml Ϯ 30.2 ng/ml) compared with PBStreated mice (15 ng/ml Ϯ 2.7 ng/ml) (p Ͻ 0.05). In addition, serum PAF levels at 2 and 16 h were not significantly different between mice treated with LeTx and controls (data not shown). Administration of LeTx did not affect PAF levels in the liver (Fig. 1B). Serum PAF-AH activity was significantly lower in LeTx-treated mice (129 Ϯ 24 mol/min/ml) compared with PBS-treated (155 Ϯ 22 mol/min/ml) and untreated mice (178 Ϯ 18 mol/min/ml) (p Ͻ 0.05) (Fig. 1C). There were no differences in serum PAF-AH activity between PBS-treated and PA-treated mice (data not shown).
Because macrophages play a central role in LeTx pathogenesis and produce PAF (23,24), we investigated the possibility that macrophage depletion would provide protection against lethal toxicity. Macrophage depletion by administration of clodronate liposomes provided a survival benefit to WT mice (median survival, 6 days) from LeTx-induced death compared with WT mice treated with PBS liposomes (median survival, 1 day) (p Ͻ 0.05) (Fig. 2D).
Hematocrit-Administration of LeTx to WT mice resulted in an average 11% increase in HCT compared with PBS-treated mice, 2 h following LeTx injection (Fig. 3A). This increase in HCT was still present at 6 h following LeTx injection. However, at 24 h WT mice manifested a 42% decrease in HCT compared with mice injected with PBS (Fig. 3B). Administration of some PAF antagonists (CV3988 and WEB 2086), but not others (quinacrine and ginkgolide B) ameliorated, but did not entirely prevent, the increase in HCT produced by LeTx injection at 2 h (Fig. 3C). No effects on HCT by antagonists in the absence of LeTx were observed (data not shown). In addition, PAFr Ϫ/Ϫ mice experienced smaller changes in HCT at 2 and 24 h in response to LeTx when compared with WT mice (Fig. 3, A and B).
Respiratory Function-Within 1.5 h of LeTx injection, WT mice developed labored breathing. Visible changes in breathing correlated with changes in respiratory function as measured by whole body plethysmography, including a greater than 50% decrease in respiratory rate and a 20% decline in tidal volume (Fig. 4A). Inspiratory and expiratory times were also markedly prolonged (Fig. 4B). Mice that showed improved respiratory parameters by 24 h generally recovered, whereas animals with persistent respiratory compromise generally died within 24 h (data not shown). In contrast to hematocrit studies, we were unable to demonstrate an effect of the PAF antagonist WEB 2086 on respiratory parameters (data not shown). Furthermore, PAFr Ϫ/Ϫ mice still developed alterations in respiratory function in response to LeTx, although tidal volume and FIGURE 1. In vivo PAF measurements following LeTx injection. A, WT mice (n ϭ 4 -6) were injected intravenously with LeTx (120 g of PA and 50 g of LF) and sacrificed 30 min post-injection. Serum from LeTx-treated WT mice contained significantly higher amounts of PAF (p Ͻ 0.05) compared with PBS-treated WT mice. No differences were detected in serum PAF levels of untreated and PA-treated WT mice compared with PBS-treated mice (p Ͼ 0.05). B, no differences were detected in PAF levels in the livers of untreated and PBS-or LeTx-treated WT mice. C, serum from LeTx-treated WT mice contained lower amounts of PAF-AH (*, p ϭ 0.031) compared with untouched and PBS-treated WT mice. The lines represent medians for experimental groups. Statistical analyses were done using the Kruskal-Wallis test, and post hoc analyses were done using the Dunn multiple comparison test.
peak expiratory flow were not as affected compared with WT mice (Fig. 4C).
Biochemical Profile Studies-WT mice exhibited a dose-dependent increase in serum glutamic oxaloacetic transaminase (SGOT) and serum glutamate pyruvate transaminase (SGPT) levels ( Fig. 5A) that did not occur in PAFr Ϫ/Ϫ mice (Fig. 5B) at 24 h post-LeTx challenge. WT mice, but not PAFr Ϫ/Ϫ mice, injected with LeTx also exhibited a dose-dependent increase in blood urea nitrogen and a decrease in serum albumin concentration, also consistent with volume loss (Fig. 5, C and D).
LeTx-related Histological Findings-Two hours after LeTx challenge, there were minimal inflammatory changes. Several histologic findings were common to WT and PAFr Ϫ/Ϫ mice, including the presence of small cells with condensed nuclei suggestive of apoptotic cells within blood vessels in bone marrow and lung (data not shown). Loss of red blood cells within the red pulp was noted, consistent with splenic contracture (Fig. 7, A  and B), which is consistent with a physiologic response to volume loss. The lungs of WT and PAFr Ϫ/Ϫ mice also had increased alveolar capillary cellularity and minimal acute fibrinous pneumonitis (data not shown).
Twenty-four hours after LeTx challenge, WT mice exhibited hepatic necrosis that was widespread, with a centrilobular to midzonal distribution (Fig. 7C, panel i). In contrast, PAFr Ϫ/Ϫ mice did not have evidence of liver necrosis (Fig. 7C, panel ii).
Other findings were generally similar to those identified at 2 h (circulating micronucleated/apoptotic cells) and splenic contracture without significant differences between WT and PAFr Ϫ/Ϫ mice. At 24 h, significant lymphocytic apoptosis was evident in the spleen of WT (Fig. 7C, panel iii) and PAFr Ϫ/Ϫ mice.
Apoptosis Studies-Examination of the bone marrow and lung of LeTx-challenged PAFr Ϫ/Ϫ mice revealed negative staining for cleaved caspase 3 (Fig. 8, A and C). In contrast, numerous cells were positive for cleaved caspase 3 in the bone marrow and lungs of LeTx-challenged WT mice (Fig. 8, B and D). Negative staining for cleaved caspase 3 was noted in PBS-treated WT and PAFr Ϫ/Ϫ mice (data not shown).
Cytokine Expression-Given the role of PAF in systemic inflammatory response, we investigated cytokine/chemokine expression in response to LeTx. Twenty-four hours after LeTx injection, alterations in cytokine/chemokine expression were observed in the lungs of mice (WT and PAFr Ϫ/Ϫ ) by cytokine array (Fig. 9A). At 24 h, WT and PAFr Ϫ/Ϫ mice challenged with LeTx exhibited increased expression of MCP-1 and RANTES compared with mice injected with PBS (Fig. 9, A and B). This induction was confirmed by ELISA studies (Fig. 9B and data not shown). In addition, WT and PAFr Ϫ/Ϫ mice injected with LeTx exhibited a small, but statistically significant decrease in soluble TNFR-1 expression relative to PBS-injected counterparts. IL-12 expression was only detected in PAFr Ϫ/Ϫ WT mice injected with LeTx. MIP-1␣ induction was not detected following LeTx injection (data not shown).
Cellular Toxicity-To determine whether survival differences among WT and KO mice were related to differences in macrophage susceptibility to LeTx, in vitro macrophage experiments were performed. Peritoneal macrophages isolated from WT and PAFr Ϫ/Ϫ mice exhibited a dose-dependent decrease in MTT signal in response to LeTx treatment. However, there were no differences between WT and PAFr Ϫ/Ϫ mouse macrophage susceptibility to LeTx at all concentrations tested (data not shown). At a LeTx concentration of 1 g/ml, there were 29 and 31% decreases in MTT signal for WT and PAFr Ϫ/Ϫ mouse macrophages compared with cells treated with media alone (data not shown). In contrast, LeTx did not affect viability of hepatocytes isolated from WT mice (data not shown).

DISCUSSION
PAF is induced during inflammation by both exogenous (e.g., LPS, HIV infection) and endogenous stimuli and is an active mediator of inflammation (25)(26)(27). In animal models, many of the features of endotoxin-induced shock can be reproduced by PAF injections (28). Our findings demonstrate that serum PAF levels are transiently increased in response to LeTx challenge in WT mice. Our findings are consistent with recent studies that reveal a rapid induction of inflammatory lipid mediators by LeTx-mediated activation of the inflammasome (29). In addition, our studies suggest that PAF contributes to the mortality of LeTx in WT mice because PAFr Ϫ/Ϫ mice and WT mice treated with PAF receptor antagonists exhibited prolonged survival. We also observed that increased serum PAF levels corre- Brackets denote standard deviation. Experiments were done in separate sets of mice and repeated with comparable results. *, p Ͻ 0.05 for comparison with control; **, p Ͻ 0.05 for comparisons between WT and PAFr Ϫ/Ϫ mice. C, HCT in WT mice treated with PAF antagonists prior to LeTx injection. HCT were measured 2 h following LeTx injection. *, p Ͻ 0.05 for comparison between animals pretreated with PAF antagonist versus PBS prior to LeTx challenge. For individual comparisons (A and B), a Mann-Whitney test was used. For multiple comparisons (C), the Kruskall-Wallis test was used, and post hoc analyses were done using the Dunn multiple comparison test. . WT mice challenged with LeTx exhibited shallow and slower breathing compared with baseline. A, bars show the average percent decreases from baseline for different respiratory functions following LeTx injection. B, bars show the average percent increase for inspiratory and expiratory times. Brackets denote standard deviation. All changes shown are statistically significant from baseline. C, WT and PAFr Ϫ/Ϫ mice developed changes in respiratory function compared with baseline. However, decreases in tidal volume and peak expiratory flow were less prominent for PAFr Ϫ/Ϫ compared with WT mice. Bars represent the average percent decrease in respiratory rate, tidal volume, peak inspiratory flow, and peak expiratory flow. Bars denote averages of n ϭ 4 -5 mice, and brackets denote standard deviation WT (*, p Ͻ 0.05). FREQ, respiratory rate; TV, tidal volume; PIF, peak inspiratory flow; PEF, peak expiratory flow; Ti, inspiratory time; Te, expiratory time. Statistical analyses were done using a Mann-Whitney test.
lated with a decrease in PAF-AH activity. PAF-AH is the enzyme primarily responsible for the degradation of PAF, normally limiting the half-life of PAF to minutes. Thus, decreased activity could result in increased PAF levels in our experiments. The magnitude of the observed decrease in PAF-AH in response to LeTx in WT mice is similar to that described in endotoxin challenged gerbils and comparable to the decreased activity observed in critically ill patients on presentation (30,31).
Early studies linked LeTx susceptibility in mice to macrophage toxicity, although more recent studies dispute this association (32). The effects of LeTx on macrophage viability are mediated by the NOD-like receptor sensor, NLRP1, and related to inflammasome assembly and caspase-1 activation (33)(34)(35). Our results suggest that PAF produced by macrophages contributes to LeTx-related pathology. In this regard, macrophages from WT and PAFr Ϫ/Ϫ mice were both susceptible to LeTxinduced death. Furthermore, macrophage depletion provided survival benefits to WT mice from LeTx-induced death. Therefore, we hypothesize that LeTx-induced macrophage damage contributes to increased PAF levels, either directly from macrophages or from other cells in response to macrophage damage.
To understand the contribution of PAF in disease following LeTx challenge in a toxin-susceptible mouse strain, we performed a variety of physiologic and biochemical studies. Several of the clinical features of anthrax (e.g., pleural effusions, hemoconcentration, and bleeding) indicate significant alterations in vascular permeability. Many of these features can be reproduced by LeTx injections in animal models (36,37). Following LeTx challenge, we observed several features consistent with increased vascular permeability in WT mice including hemoconcentration, respiratory distress, increased serum blood urine nitrogen, and decreased protein levels. Furthermore, FIGURE 5. Effects of LeTx on Biochemical profile. The effects of LeTx on liver function and serum chemistry for WT (A and C) and PAFr Ϫ/Ϫ mice (B and D) are shown. Liver toxicity was assayed by changes in serum aspartate (SGOT), alanine transaminase (SGPT), and alkaline phosphate (AP) levels. Serum chemistry tests included blood urine nitrogen (BUN) and total protein (TPR). Mice were injected intravenously with 120 g of PA and 50 g of LF or with 60 g of PA and 25 g of LF. *, p Ͻ 0.05. The legend applies to all panels. Statistical analyses were done using the Kruskal-Wallis test, and post hoc analyses were done using the Dunn multiple comparison test. FIGURE 6. Evans blue extravasation. Vascular permeability was measured in WT mice by Evans blue extravasation. WT mice treated with LeTx had higher absorbance readings compared with PBS-treated WT mice (p Ͻ 0.05). WT mice treated with WEB 2086 prior to LeTx injection had lower absorbance readings. n ϭ 4 -6 mice. Statistical analyses were done using the Kruskal-Wallis test, and post hoc analyses were done using the Dunn multiple comparison test.
LeTx-challenged mice also exhibited splenic contracture, which is likely a physiologic response to volume loss in this context.
The basis by which LeTx enhances vascular permeability is poorly understood. Incubation of endothelial cells with LeTx resulted in apoptosis and altered function in some studies (38,39). Furthermore, LeTx-induced macrophage death has been reported to augment apoptotic death of endothelial cells (40). We found that some (e.g., hemoconcentration and increased serum blood urine nitrogen levels) but not all of the changes indicative of vascular permeability in WT mice were partially ameliorated by the administration of PAF antagonists and in PAFr Ϫ/Ϫ mice. These effects on vascular permeability were confirmed with Evans blue studies. Overall, we interpret these findings as suggestive that PAF contributes to, but is not entirely responsible for, the changes in vascular permeability induced by LeTx. PAF is well known to alter endothelial function, and these effects may be related to direct effects (41) or macrophage and endothelial cell activation (42,43).
Extensive apoptosis following LeTx was present in a variety of organs and cell types, including endothelial cells. In contrast, apoptosis was dramatically reduced in PAFr Ϫ/Ϫ mice. The pattern of LeTx-induced macrophage death has been related to polymorphisms in Nalp1b and the macrophage activation state (33). Typically, macrophages from sensitive BALB/c mice undergo lysis and not apoptosis in response to LeTx (35,44), although apoptosis may occur in the context of sublytic levels of  toxin (45). Endothelial cell apoptosis has also been described in HUVEC cells by some but not all investigators (38,39,46). Interestingly, PAF was shown to promote apoptosis in a variety of cell types including neurons, corneal cells, and tumors (47)(48)(49)(50). Additional study will be needed to determine the role of PAF in LeTx-induced apoptosis.
PAF can induce many of the cellular pathways involved in inflammation and shock, including superoxide production, lymphocyte and neutrophil migration, mast and basophil degranulation, and NF-B activation (reviewed in Refs. 51 and 52). However, the role of cytokine storm in LeTx-induced damage is not clear (53). We observed little LeTx-mediated induc-tion of inflammatory cytokines in our mouse studies at 2 or 24 h, with minimal differences between WT and PAFr Ϫ/Ϫ mice.
We found that LeTx challenge produced substantial centrilobular hepatic necrosis (as evidence by an increase in serum transaminase levels and histologic studies), which was PAF receptor-dependent. Hepatic necrosis in response to LeTx was previously described in mice (53) and was also reported in the autopsy series of human anthrax cases associated with the Sverdlovsk epidemic (54). In murine studies, necrosis did not correlate with inflammatory cytokine changes, nor was it dependent on the presence of FASL, but it did correlate with evidence of tissue hypoxia (53). The presence of widespread centrilobular hepatocellular necrosis in our studies is consistent with vascular insufficiency, although centrilobular necrosis may also occur in the context of a variety of intoxications. PAF has been reported to play a role in hepatic necrosis induced by both toxins (acetaminophen and ethanol) and ischemia (55)(56)(57). In animal studies, PAF infusions have been shown to increase hepatic glycogenolysis (58,59) and induce vasoconstriction resulting in liver hypoxia (60). Both these effects have been linked to PAFinduced prostaglandin synthesis (58,60).
In summary, our results implicate PAF in the toxic effects of B. anthracis LeTx, especially as it relates to alterations in vascular permeability and hepatotoxicity. PAF appears to play a role downstream of macrophage death specifically in BALB/c mice. Extrapolations of findings from mice to humans must be done cautiously and judiciously, especially because human macrophages/monoctyes are more displayed decreased susceptibility to experimental infection. Our findings suggest the possibility that PAF antagonists may be helpful as an adjunctive therapy for anthrax. Given the high mortality associated with anthrax-associated shock despite antimicrobial therapy and supportive therapy, additional study of PAF antagonists is warranted. FIGURE 9. Differences in cytokine and chemokine expression induced by LeTx challenge in WT and PAFr ؊/؊ mice. Cytokine and chemokine levels in lung were measured using mouse cytokine protein array 24 h post-LeTx (120 g of PA and 50 g of LF) challenge. A, graph shows intensity of expression for individual cytokines/chemokines relative to an internal standard. B, array blots show reactivity for corresponding cytokines/chemokines. C, ELISA experiments revealed induction of MCP-1 in the lungs of WT and PAFr Ϫ/Ϫ mice following LeTx injection. n ϭ 6; PBS: n ϭ 3. Bars denote mean cytokine/ chemokine levels, and error bars denote standard deviations. *, p Ͻ 0.05 for comparison to PBS-injected animals. Statistical analysis was done using a Mann-Whitney test.