Syndecan 1 Shedding Contributes to Pseudomonas aeruginosa Sepsis
http://www.100md.com
感染与免疫杂志 2005年第12期
Departments of Surgery Pathology Microbiology and Immunology, Texas Tech University Health Sciences Center, 3601 4th St., Lubbock, Texas 79430
Department of Medicine, Baylor College of Medicine, Houston, Texas 77030
ABSTRACT
The innate immune system is comprised of many components that function coordinately to prevent bacterial sepsis. However, thermal injury suppresses many of these factors, and the opportunistic pathogen Pseudomonas aeruginosa takes advantage of this condition, making it one of the leading causes of morbidity and mortality in the setting of thermal injury. P. aeruginosa is extremely efficient at colonizing burn wounds, spreading systemically, and causing sepsis, which often results in a systemic inflammatory response, multiple-organ failure, and death. The pathogenicity of P. aeruginosa is due to the arsenal of virulence factors produced by the pathogen and the immunocompromised state of the host. Syndecan 1 is a major heparan sulfate proteoglycan present on many host cells involved in thermal injury. Syndecan 1 anchored to the cell surface can be cleaved in a process termed ectodomain shedding. Syndecan 1 shedding results in the release of intact, soluble proteoglycan ectodomains that have diverse roles in innate immunity. Here we show for the first time that thermal injury results in shedding of syndecan 1 from host tissue. Our data show that syndecan 1 null mice are significantly less susceptible to P. aeruginosa infection than their wild-type counterparts, as demonstrated by (i) significantly lower mortality; (ii) absence of systemic spread of P. aeruginosa; and (iii) significant reductions in some proinflammatory cytokines. These results suggest that shed syndecan 1 plays an important role in the pathogenesis of P. aeruginosa infection of thermal injury and that syndecan 1-neutralizing agents may be effective supplements to current P. aeruginosa treatments.
INTRODUCTION
The gram-negative opportunistic pathogen Pseudomonas aeruginosa is particularly prevalent in burn wound infections. P. aeruginosa bacteremia is one of the leading causes of morbidity and mortality in the setting of thermal injury. It often causes multiorgan failure and results in mortality rates of up to 50% (2, 25, 26, 39). The extensive damage induced during P. aeruginosa infection is due to the ability of the microorganism to produce a large array of virulence factors (25) and the immunocompromised state of the individual. However, the increased ability of P. aeruginosa over many other pathogens to cause sepsis in thermally injured individuals is not fully understood.
Beyond the initial barrier of the skin, several components of the host's innate immune system are designed to act as sentinels that guard against systemic bacterial attacks. Toll-like receptors (TLRs) are transmembrane receptors found on host cells that recognize pathogen-associated molecular patterns (19). Upon recognition of a pathogen-associated molecular pattern, TLRs activate downstream signaling events, which induce an immediate inflammatory response (19). Acute inflammation, as seen in thermal injury, is characterized by vasodilatation and neutrophil infiltration. Consequences of inflammation include the production of cytokines and chemokines. Chemokines, such as interleukin-8 (IL-8), recruit neutrophils into the damaged area, which normally engulf the infecting bacteria before high numbers are reached. Proinflammatory cytokines, such as IL-1, IL-6, and tumor necrosis factor alpha (TNF-), are produced predominately by activated macrophages. Among their many roles, cytokines regulate neutrophil functions and perpetuate the inflammatory response. Although these first lines of defense are normally effective, the highly devascularized environment of a burn wound and overall dysregulation of immunity in burned individuals provide an ideal setting for the proliferation and spread of bacteria within the injured tissue and subsequent dissemination throughout the immunocompromised host. In this environment, inflammation can shift from a protective process to a destructive one. The massive neutrophil infiltration, characteristic of thermal injury, may actually have a negative biological effect. Activated neutrophils release oxidants and hydrolytic enzymes that, when combined with bacterial degradative enzymes, can result in increased proteolysis within the wound and may actually aid spread and eventual dissemination of bacteria (14, 33). This neutrophil paradox has been implicated in the pathology of many chronic inflammatory conditions (36).
Dozens of inflammatory mediators play important roles in balancing the protective and destructive effects of inflammation. One important group of inflammatory mediators are heparan sulfate proteoglycans (HSPGs) (7). HSPGs are ubiquitous molecules expressed by all adherent cells and are composed of a core protein and one or more covalently attached heparan sulfate (HS) chains (7). The HS chains are made up of repeating units of hexuronic acid and N-substituted glucosamine disaccharide units. One family of HSPGs implicated in inflammation is the syndecans. Four syndecan family members (syndecan 1 to 4) have been cloned in mammals and are implicated in mediating a wide range of effects, including leukocyte rolling on endothelial cells, wound repair, angiogenesis, vascular permeability, and modulation of chemokine activity (7). Syndecans attached to the cell surface can be cleaved in a process termed shedding. Syndecan shedding results in the release of intact, soluble proteoglycan ectodomains into the host cell environment that can mediate cellular physiology at distant sites (7). Although syndecan ectodomains are shed at low levels as part of normal cell turnover, tissue injury can markedly activate syndecan shedding (12). Thus, shed syndecan 1 (Sdc-1) is not found at detectable levels in the biological fluids of healthy individuals, but high levels of shed Sdc-1 and Sdc-4 have been documented in human dermal wound fluids (37) and in tracheal aspirates of ventilated preterm infants (13).
Syndecan shedding during tissue injury is presumably due to mechanical shearing of the ectodomains and/or cleavage by host cell-derived proteases (12). However, it has also been shown that the LasA protease produced by P. aeruginosa activates syndecan 1 shedding (23). Shed Sdc-1 contributed to P. aeruginosa pathogenesis in a mouse model of lung infection (22). Sdc-1 knockout (Sdc-1–/–) mice were significantly more resistant to P. aeruginosa lung infection than their wild-type (WT) counterparts, as demonstrated by decreased mortality and bacteremia (22). Furthermore, treating P. aeruginosa-infected WT mice with specific inhibitors of Sdc-1 shedding and HS was therapeutic (22). Considering the demonstrated importance of shed Sdc-1 to P. aeruginosa pathogenesis in the lung, and assuming that the significant mechanical damage that is inflicted on host tissue during thermal injury results in high levels of shed Sdc-1, we hypothesized that shed Sdc-1 is a decisive host factor contributing to P. aeruginosa pathogenesis in burn wound infections. Here we report that Sdc-1–/– mice are significantly less susceptible to P. aeruginosa burn wound infections than WT mice and that shed Sdc-1 may play a role in the blood vessel invasion and subsequent systemic spread of P. aeruginosa. These data provide support for the prospect of using Sdc-1 neutralizing agents in the treatment of burn wounds.
MATERIALS AND METHODS
Mice. Sdc-1–/– mice were generated and bred as previously described (22). Sdc-1–/– mice used in this study were backcrossed eight times onto the C57BL/6 background and were maintained by interbreeding at Baylor College of Medicine (Houston, Tex.). WT Sdc-1+/+ C57BL/6 mice were obtained from Charles River Laboratories (Wilmington, MA) or were derived from crossing mice heterozygous for the syndecan 1 gene. Mice used in experiments were 6 to 8 weeks old and weighed 17 to 20 g. Mice were housed and studied under protocols approved by the Institutional Animal Care and Use Committee in the animal facility of Texas Tech University Health Sciences Center (Lubbock, TX).
Bacterial growth and inoculum. P. aeruginosa strain PAO1 was grown in Luria-Bertani (LB) medium (1). Aliquots (50 μl) of overnight cultures of PAO1 were subcultured in fresh LB broth and grown at 37°C for 4 h to an optical density at 540 nm of approximately 0.9. A 100-μl aliquot of each culture was then pelleted, washed in phosphate-buffered saline (PBS), and serially diluted (10-fold serial dilutions) in PBS. A 100-μl aliquot of the 10–4 dilution was injected into each animal. As we have previously determined (9), this dilution contains approximately 2 x 102 to 2 x 103 CFU of P. aeruginosa. We have also shown previously that this dose of PAO1 produces 90 to 100% lethality in WT C57BL/6 mice by 48 h postburn infection (9). The exact inoculum of each strain was determined by plating serial dilutions of the inoculum on LB agar plates.
Thermally injured mouse model. The virulence of PAO1 was examined by using the modified burned-mouse model of Stieritz and Holder (35). In this modified model, a scald burn is induced. Mice were anesthetized by intraperitoneal injection of 0.4 ml of Nembutal/20 g of body weight at 5 mg/ml (5% sodium pentobarbital; Abbott Laboratories, North Chicago, Ill.), and their backs were shaved. The mice were securely placed into a template with an opening (4.5 by 1.8 cm) exposing their shaved backs. About 15% of the total surface area of the mouse was exposed through the opening on the template. The thermal injury was induced by placing the exposed area of the shaved skin in 90°C water for 10 s. Such an injury is nonlethal but causes a third-degree (full-thickness) burn. Fluid replacement therapy consisting of a subcutaneous injection of 0.8 ml of a 0.9% NaCl solution was administered immediately following the burn. Mice were challenged by the subcutaneous inoculation of 100 μl of the bacterial inoculum (see above) directly under the burn. Control mice were subcutaneously injected with 100 μl of sterile phosphate-buffered saline (PBS) directly under the burn. In some experiments, porcine mucosal heparan sulfate (Sigma, St. Louis, Mo.) dissolved in PBS was simultaneously injected under the burn with the bacterial inoculum. During recovery, the mice were continuously observed under warming lights.
Mortality among infected mice was recorded at 24 h intervals, up to 5 days postburn and postinfection (postburn/infection). Animals were treated humanely and in accordance with the protocol approved by the Animal Care and Use Committee at Texas Tech University Health Sciences Center (Lubbock, TX).
Quantitation of bacteria within the skin and livers. At 24 h postburn/infection, mice were euthanized by intracardial injection of 0.2 ml of Sleepaway (sodium pentobarbital-7.8% isopropyl alcohol euthanasia solution; Fort Dodge Laboratories, Inc., Fort Dodge, Iowa). Skin sections of approximately 5 by 5 mm were obtained from the burned skin of both the control and challenged mice. Simultaneously the entire liver of each animal was obtained. Tissues were weighed, suspended in 2 ml of PBS, and homogenized (overhead stirrer; Wheaton Instruments, Millville, N.J.). Homogenates were serially diluted, and a 100-μl aliquot of each diluent was plated on LB agar plates to determine the number of CFU, which was calculated per gram of tissue.
Immunoblotting experiments. Skin sections from control (nonburned) mice, mice that had been administered a third-degree scald burn, and mice that were burned and infected with P. aeruginosa (PAO1) were harvested at 24 h postburn. The tissue was placed in phosphate-buffered saline (PBS) and gently vortexed for 10 s. Five-microliter samples of the wound fluid extracts were then dot blotted in triplicate onto a cationic nylon membrane (Immobilon Ny+; Millipore) and quantified by dot immunoblotting with the 281-2 rat anti-mouse Sdc-1 ectodomain monoclonal antibody as described previously (21, 23). The same blot was stripped and then reprobed with an antibody directed against the cytoplasmic domain of Sdc-1 (21) to confirm that Sdc-1 detected by the 281-2 antibody was shed ectodomain and not intact Sdc-1 released from cells damaged by vortexing.
Hematoxylin and eosin (H&E) staining. Mouse skin was harvested and immediately placed into 4% paraformaldehyde in phosphate buffer, pH 7.3, and fixed overnight at 4°C. The tissue was then processed through a graded series of ethanol, cleared in xylene, and infiltrated with paraffin. The tissue was embedded on edge in paraffin, and blocks were sectioned with a Cut 4055 Olympus microtome. Five-micrometer sections of the blocks were collected on Superfrost/Plus slides (Fisher Scientific) and H&E stained using standard laboratory techniques. Slides were visualized by light microscopy at 400x with an Olympus BH-2 microscope and imaged with an Olympus camera model C-35AD-4 (Tokyo, Japan).
Scanning electron microscopy. Mouse skin that had been burned and infected with P. aeruginosa was harvested, cut into 1-mm-thick cross-sections, and placed into fixative containing 1.5% glutaraldehyde and 3% paraformaldehyde in 0.1 M Millonig's buffer with dextrose. Mouse skin that had been burned but not infected was treated in a similar manner to use as a control. The tissues were fixed overnight at 4°C. Tissue was then frozen in liquid nitrogen and fractured with a frozen blade to expose more surface. Tissues were then washed in buffer, postfixed in 1% osmium tetroxide, and taken through a graded series of alcohols. The tissue was critically point dried with a Tousimis critical point dryer (Rockville, MD). The sections were mounted on edge on stubs coated with carbon tape. Some of the infected skin tissue samples were carved using a razor blade under a dissecting scope to reveal more of the inner structure. All the tissues were gold coated using a Polaron sputter coater. The sections were then scanned for bacteria, and pictures were taken using a Hitachi S-500 scanning electron microscope (Hitachi America, Ltd., Brisbane, CA).
Cytokine expression. Total RNA from mouse skin sections was extracted using TRI Reagent-RNA/DNA/Protein isolation reagent per the manufacturer's guidelines (Molecular Research Center, Inc., Cincinnati, OH). RNA was treated with RQ1 RNase-Free DNase (Promega, Madison, WI) to remove any contaminating genomic DNA. A total of 20 μg of DNA-free RNA was used to detect the expression of murine cytokine genes using the mck4 and mck3b Multi-Probe Template Sets and the Riboquant RNase Protection Assay System (PharMingen, San Diego, CA). The probes were hybridized overnight at 56°C with 20 μg of total RNA from either the liver or the skin. The mixtures were then RNase treated, purified, and resolved on a 6% acrylamide-8 M urea gel per the manufacturer's guidelines. Gels were dried, and protected bands representing different cytokines were detected after exposure of the gels to a phosphorimager screen. Radioactive signals were imaged using a Typhoon variable mode imager (Amersham Biosciences, Piscataway, NJ). Signals were quantitated using Image-Quant TL (Amersham Biosciences, Piscataway, NJ). The signal of the band representing a cytokine of interest was normalized to the signal of the L32 (ribosomal gene) band from the same sample.
Statistical analysis. The Student's t test and the Fisher's exact test (Statview; Abacus Concepts, Inc., Berkeley, CA) were used to determine significant differences between the numbers of CFU/gram of tissue in bacterial colonization experiments and between groups for the mortality experiments, respectively. Significant differences between cytokine mRNA levels in the RNase protection assays were determined by the Student's t test (Microsoft Excel 2002, Austin, TX).
RESULTS
Thermally injured syndecan 1 knockout mice are less susceptible to P. aeruginosa infection. Tissue injury activates syndecan shedding (12), and shed Sdc-1 is found at high levels in human dermal wound fluids (37) and in tracheal aspirates of ventilated preterm infants (13). However, Sdc-1 has not been demonstrated in burn wounds. To determine if thermal injury results in Sdc-1 shedding, Sdc-1 ectodomain levels in burn fluid were measured by dot immunoblotting. Skin sections from control (nonburned) mice, mice that had been administered a third-degree scald burn, and mice that were burned and infected with P. aeruginosa (PAO1) were harvested at 24 h postburn. The tissue was placed in phosphate-buffered saline (PBS) and gently vortexed for 10 s. Samples were then dot blotted onto a cationic nylon membrane and quantified by immunoblotting with an Sdc-1 ectodomain antibody. Sdc-1 was detected in the fluid extract of burned, but not nonburned, skin; however, increased levels of Sdc-1 were detected in the wound fluid extracts of mice that were burned and infected with PAO1 (Fig. 1). Sdc-1 ectodomains were not detected when the membrane was reprobed with an antibody against the cytoplasmic domain of Sdc-1, confirming that Sdc-1 was shed and not released intact from cells (data not shown).
To determine if the increased levels of shed Sdc-1 that are associated with thermal injury correlated with the severity of P. aeruginosa infection, we utilized Sdc-1–/– mice, which carry a deletion in the gene encoding Sdc-1, and the thermally injured mouse model. Sdc-1–/– mice are fertile and develop normally (8). The thermally injured mouse model of P. aeruginosa infection was originally described by Stieritz and Holder and closely resembles the pathology of human thermal injury (35). We have previously determined that the P. aeruginosa strain PAO1 causes approximately 90% mortality to thermally injured mice by 48 h postburn/infection (30, 31).
Groups of WT or Sdc-1–/– mice on the C57BL/6 background were given full-thickness scald burns and infected with PAO1. The mice were observed for 120 h (5 days) postburn/infection, and percent mortality was recorded for each group (Fig. 2). Eighty-two percent (18 out of 22) of the burned and infected C57Bl/6 WT mice died in the first 48 h, and no further mortality was observed. However, the percent mortality of Sdc-1–/– mice was significantly reduced at all time points (Fig. 2; 13% [2 out of 15] at 48 h [P = 0.00005] and 40% [6 out of 15] at 120 [P = 0.014]).
Shed Sdc-1, via its HS chains, can function as a soluble effector to regulate various cellular processes, including proliferation, adhesion, and differentiation (3). In a P. aeruginosa lung infection model, HS was as effective as purified shed Sdc-1 at increasing the susceptibility of Sdc-1–/– mice to lung infection from P. aeruginosa (22). To determine if the absence of shed Sdc-1 was responsible for the lower susceptibility of burned Sdc-1–/– mice to P. aeruginosa infection, we tested whether HS could increase mortality. Sdc-1–/– mice were burned and infected with PAO1 as before; however, some groups of mice were also administered a single 20-μg, 50-μg, or 100-μg dose of HS locally at the injury site. These doses were chosen because we have previously determined that 100 μg of protamine sulfate is required to neutralize syndecan 1 in vivo (9). As before, the percent mortality at 48 h in Sdc-1–/– mice infected with PAO1 was relatively low 25% (2 out of 8 mice). However, the percent mortality was significantly higher (83%, 5 out of 6 mice; P = 0.05; for 100 μg) in mice that received a simultaneous HS injection and increased to 100% by the 72-h time point (Fig. 3). The percent mortality among HS-injected mice was similar to that seen for WT mice that possess Sdc-1 (compare Fig. 2 to 3). No mortality was observed in nonburned mice that were given 100 μg HS or infected with PAO1 plus 100 μg HS (data not shown). Taken together, these results indicate that the high levels of Sdc-1 shed due to thermal injury exacerbate the virulence of P. aeruginosa.
Shed syndecan 1 facilitates the systemic spread of P. aeruginosa. P. aeruginosa infection in the murine thermal injury model can be divided into two phases (10, 11). The first phase involves colonization of the burned tissue by P. aeruginosa. The devascularized state of the burn eschar (burn wound area) provides an environment ideally suited for bacterial growth, and P. aeruginosa proliferates extremely quickly. Once a threshold concentration of bacteria is reached in the eschar (approximately 109 CFU/g tissue), P. aeruginosa spreads systemically, causing sepsis followed by multiple-organ failure and eventually death (10, 11). We have previously seen that by 24 h postburn/infection, P. aeruginosa CFU in the eschar increase from 102 CFU/g tissue (infecting dose) to approximately 109 CFU/g tissue and can also be isolated from the liver and spleen (29, 32). By 48 h postburn/infection, the majority of mice succumb to the infection (29, 32).
To determine if Sdc-1 influenced either stage of P. aeruginosa infection, we examined both the local colonization within the eschar and the systemic spread of P. aeruginosa in burned and infected WT and Sdc-1–/– mice. We examined the local colonization of bacteria within the eschar by determining the number of PAO1 CFU in epithelial sections taken at two specific sites within the area of thermal injury, the inoculation site and a site 15 mm distal to the inoculation site. Using this approach, we have previously shown that PAO1 spreads efficiently from the inoculation site to the distal site by 24 h postburn (32). As in the mortality experiments, mice were thermally injured and inoculated with PAO1. At 24 h postburn/infection, the mice were euthanized. Epithelial sections were isolated and homogenized, and the number of PAO1 CFU was determined as described in Materials and Methods. WT and Sdc-1–/– mice had similar P. aeruginosa CFU counts within both the "inoculation" and "distal" epithelial sections (Fig. 4), and administration of HS did not affect local spread. These results indicate that within 24 h, P. aeruginosa is able to efficiently multiply and colonize the burn eschar despite the lack of Sdc-1.
To examine the systemic spread of P. aeruginosa, mice were thermally injured and inoculated with PAO1 as described above. At 24 h postinfection, the mice were euthanized, the livers were harvested, and the CFU/gram of tissue was determined. On average, 3.4 x 105 ± 9 x 104 CFU were recovered from livers of WT mice (5 mice/group), while no bacteria were recovered from livers of Sdc-1–/– mice (5 mice/group). Administration of heparan sulfate, as described above, resulted in similar CFU counts as seen in WT mice (9 x 103 ± 5 x 103). Taken together, these results indicate that P. aeruginosa can efficiently colonize the burn eschar and multiply to threshold levels despite the absence of Sdc-1. However, Sdc-1 appears to be needed for dissemination of PAO1 into the bloodstream and secondary colonization of internal organs.
Syndecan 1 is involved in P. aeruginosa perivascular cuffing. Two distinctive clinical features of P. aeruginosa bacteremia are invasion and necrosis of blood vessels (34). In humans, blood vessel invasion by P. aeruginosa correlates with the appearance of skin lesions termed ecthyma gangrenosum (17). Ecthyma gangrenosum is classically considered a pathognomonic sign of sepsis by P. aeruginosa (17, 18). However, a few other bacterial species have now been associated with ecthyma gangrenosum (27). Upon microscopic examination, blood vessel invasion by P. aeruginosa can be recognized by the presence of perivascular cuffing (PVC) (34). Bacilli form a circumferential pattern surrounding the vessel and are often aligned single file or in stacks between muscle cells of the venous walls (34). The mechanisms of PVC by P. aeruginosa and its role in pathogenesis are not fully understood. However, in the thermally injured mouse model, we have consistently seen that PVC accompanies sepsis and subsequent death from P. aeruginosa infection (K. P. Rumbaugh, unpublished observations).
As Sdc-1 is a component of normal vascular endothelium and Sdc-1–/– mice displayed significantly less sepsis from P. aeruginosa infection, we wanted to determine if Sdc-1 played a role in PVC. Groups of mice were thermally injured and infected with PAO1 as described in Materials and Methods. At 24 h postburn/infection, mice were euthanized and skin sections were harvested for histological evaluation. Sections measuring approximately 1 by 1.5 cm were extracted from the burn margin (the area where burned and healthy tissue merge). PVC was visualized in hematoxylin and eosin (H&E)-stained sections by the appearance of circumferentially arranged gram-negative bacilli around one or more blood vessels (Fig. 5A). PVC was usually seen in association with vessel thrombosis or blood clotting within the vessel, as seen in Fig. 5A. PVC was observed in skin sections from 4 of 5 thermally injured, P. aeruginosa-infected WT mice. However, none of the five sections obtained from Sdc-1–/– mice displayed PVC. Although bacilli could be seen in tissue sections from Sdc-1–/– mice, they were dispersed throughout the section and the blood vessels appeared normal without thrombosis (Fig. 5B). Sdc-1–/– mice that were administered HS in addition to PAO1 also displayed extensive PVC (data not shown). Skin tissue from thermally injured WT mice infected with PAO1 was also visualized by electron microscopy. Scanning electron microscopy (SEM) revealed organized arrangements of P. aeruginosa bacilli in areas surrounding blood vessels (Fig. 5C).
Syndecan 1 modulates the host cytokine response to P. aeruginosa infection in thermally injured mice. Extensive inflammation is an important initiating event in thermal injury that predisposes the host to sepsis, multiorgan failure, and death (6, 40). We have previously shown that the combination of thermal injury and P. aeruginosa infection results in an acute and exaggerated inflammatory response that is characterized by elevated mRNA levels of several cytokine and chemokine genes (29). This "cytokine storm" can be detected locally in the skin and systemically within the livers of burned and PAO1-infected mice by 24 h postburn/infection (29). Shed Sdc-1 is known to regulate the expression and function of many cytokines (7). Sdc-1 stimulates the release of TNF-, IL-1, and IL-6 from dendritic cells and IL-1, IL-6, TNF-, IL-12, transforming growth factor- (TGF-), and prostaglandin E2 (PGE2) from macrophages (7), and it is known to play a key role in IL-8-mediated chemotaxis of neutrophils to sites of tissue injury (15). As Sdc-1 shedding on epithelial cells accompanies thermal tissue injury, we wanted to determine if Sdc-1 was a mediator in the extreme upregulation of cytokines seen at the onset of thermal injury. Therefore, we examined the mRNA levels of 19 murine cytokine and chemokine genes in both the skin and livers of thermally injured, PAO1-infected WT or Sdc-1–/– mice. The murine genes examined were TNF-, TNF-, IL-3, IL-6, IL-7, IL-11, TGF-1, TGF-2, TGF-3, gamma interferon (IFN-) and IFN-, lymphotoxin (LT), macrophage migration inhibitory factor (MIF), granulocyte-macrophage colony-stimulating factor (GM-CSF), macrophage colony-stimulating factor (M-CSF), granulocyte colony-stimulating factor (G-CSF), leukocyte inhibitory factor (LIF), and stem cell factor (SCF).
WT and Sdc-1–/– mice were thermally injured and infected with PAO1. Mice were euthanized at 24 h postburn/infection, and RNA was harvested from skin and liver sections as described in Materials and Methods. RNase protection assays (RPA) were run using the Multi-Probe RPA System (BD RiboQuant). Transcripts for many of the cytokines and chemokines examined were detected in the skin of both WT and Sdc-1–/– mice. As we have previously reported (29), MIF, M-CSF, TGF-, TNF-, IL-6, IFN-, IFN-, LIF, and G-CSF were all expressed in thermally injured, PAO1-infected skin (data not shown). However, the mRNA levels of G-CSF, LIF, and IL-6 were significantly decreased in the skin of Sdc-1–/– mice (Fig. 6A). In this study, and in agreement with our previous studies (29), the only cytokines from those examined that were detected in the livers of burned and infected mice were MIF, TGF-, G-CSF, and GM-CSF. However, Sdc-1–/– mice displayed significantly lower levels of MIF, TGF-, and G-CSF (Fig. 6B and C). Taken together, these data indicate that despite the absence of Sdc-1, Sdc-1–/– mice display a significant local inflammatory cytokine response; however, some of the cytokines that are characteristic of PAO1 infection are reduced. Conversely, the systemic inflammatory response, as illustrated by cytokine expression in the livers, was almost completely absent in Sdc-1–/– mice (Fig. 6B and C). This is consistent with the lack of systemic spread observed in Sdc-1–/– mice.
DISCUSSION
Serious thermal injuries result in dysregulation of immune responses and a hypermetabolic state that can cause multiple-organ failure and predispose patients to bacterial infections (5). P. aeruginosa is one of the most common and pathogenic bacteria encountered in burn wound infections (26, 39). P. aeruginosa can rapidly colonize burn wounds, resulting in extensive inflammation, bacteremia, and high mortality. As thermal injury causes a complicated clinical scenario, it is likely that there are many host and bacterial factors that contribute to P. aeruginosa pathogenesis in burn wounds. The goal of this study was to examine the contribution of one such host factor, Sdc-1, in P. aeruginosa infection of burn wounds. Taken together, our data indicate that Sdc-1 increases the susceptibility of thermally injured mice to P. aeruginosa infection. We show that Sdc-1–/– mice display reduced inflammatory cytokine expression, sepsis, and mortality upon a P. aeruginosa infection of a burn wound in comparison to WT mice.
Tissue injury activates shedding of Sdc-1, and high levels of shed Sdc-1 have been documented in human dermal wound fluids (37). In this study we demonstrate for the first time in vivo that Sdc-1 is shed by the trauma of thermal injury, and shedding is increased further by P. aeruginosa infection (Fig. 1). Although the LasA protease produced by P. aeruginosa can promote Sdc-1 shedding (23), we did not observe a reduction in mortality in thermally injured mice infected with a PAO1::lasA isogenic mutant (unpublished data). This indicates to us that the levels of shed Sdc-1, which are already increased due to the thermal injury, are not enhanced by the action of LasA. However, it is possible that in a P. aeruginosa lung infection where there are no reservoirs of shed Sdc-1, a LasA mutant would display reduced virulence. Consistent with this hypothesis, we detected >10 ng/ml of Sdc-1 ectodomains in tissues that were gently vortexed for 10 s, suggesting that a significantly larger concentration of Sdc-1 is present in thermally injured tissue.
Our results are consistent with previous observations suggesting that P. aeruginosa exploits shed Sdc-1 to enhance its virulence in the lung (22). Similar to data reported by Park etal., we see a significant decrease in mortality of P. aeruginosa-infected, thermally injured Sdc-1–/– mice, and this reduction can be compensated for by the addition of HS. Park et al. reported that P. aeruginosa was reduced in the primary (lung) and secondary (spleen and liver) infection sites in their pneumonia model (22). However, in the thermally injured mouse, we observed that P. aeruginosa was not inhibited from efficiently colonizing the primary infection site (burn wound) but could not spread systemically. It is possible that this inconsistency is due to the devascularized state of the burn wound that allows an uncontrolled proliferation of P. aeruginosa. The infected lung maintains a sufficient blood supply that provides access for immune system mediators to the area. It is possible that more subtle contributions of Sdc-1 to the actions of these mediators are more easily observed in the lung infection model. Despite differences in the local colonization of P. aeruginosa, both studies observed significant decreases in the systemic spread of P. aeruginosa in Sdc-1–/– mice (this study and reference 22). Furthermore, in both studies systemic spread was returned to WT levels by the addition of HS. Taken together, these observations provide strong evidence that the mechanism by which shed Sdc-1 enhances P. aeruginosa pathogenesis is related to the ability of P. aeruginosa to disseminate and cause sepsis.
We observed prevalent PVC of P. aeruginosa in skin sections from WT mice, but no PVC was seen in sections from Sdc-1–/– mice (Fig. 5A and B). As PVC is generally accompanied by blood vessel invasion (17, 34), it is possible that Sdc-1 shedding facilitates blood vessel invasion in thermally injured mice. This possibility is consistent with the lack of P. aeruginosa systemic spread in Sdc-1–/– mice and could explain the reduction in mortality. However, the mechanisms by which shed Sdc-1 enhances blood vessel invasion are unknown at this time. One possible explanation is that the transmembrane core protein that remains in the vascular endothelial cell membrane after shedding acts as an attachment site for P. aeruginosa. Some bacteria, including P. aeruginosa, use HSPGs for adherence and invasion; however, these interactions typically rely on bacterial adherence to the anionic HS side chains and not to the core protein (24, 28). Furthermore, Park et al. determined that P. aeruginosa did not bind to 125I- or 25S-labeled Sdc-1 ectodomains (22). Interestingly, the phenomenon of PVC is almost exclusive to the Pseudomonas and Aeromonas species (17) and has been relatively understudied. The correlation of PVC with sepsis and mortality implicates it as an important mechanism for P. aeruginosa pathogenesis, especially in wounds. The striking absence of PVC in Sdc-1–/– mice leads us to propose that there are important interactions between P. aeruginosa and Sdc-1 itself or an Sdc-1-regulated factor on vascular endothelial cells that deserve further investigation.
Upon thermal injury and P. aeruginosa infection, we detected significant reductions in the mRNA levels for G-CSF, LIF, and IL-6 in the dermal tissue and G-CSF, TGF-3, and MIF in the livers of Sdc-1–/– mice in comparison to WT mice (Fig. 6). However, despite these reductions, Sdc-1–/– mice displayed a significant cytokine response to the thermal injury and subsequent P. aeruginosa infection. Most of the 19 murine cytokine and chemokine genes we examined were detected in both WT and Sdc-1–/– mice, including the important proinflammatory cytokines IFN- and TNF-. The cytokine response elicited in thermally injured Sdc-1–/– mice by P. aeruginosa may help explain why some mice succumbed to the infection in the absence of systemic spread of P. aeruginosa. High levels of circulating cytokines are associated with poor clinical outcomes in trauma patients (4). Circulating proinflammatory cytokines can cause dysfunction of the renal, cardiovascular, respiratory, nervous, and musculoskeletal systems, resulting in systemic inflammatory response syndrome and eventual multiple-organ failure (4).
Significant P. aeruginosa CFU were detected in the livers of WT but not Sdc-1 mice (Fig. 4). Therefore, the most likely explanation for the difference in the systemic inflammatory response to PAO1 in WT versus Sdc-1–/– mice is the absence of significant numbers of bacteria within the livers of Sdc-1–/– mice. However, the differences in the local inflammatory response are less apparent. Similar numbers of P. aeruginosa were detected in skin sections from WT and Sdc-1–/– mice (Fig. 4), therefore the reductions in IL-6, LIF, and G-CSF message are not likely to be due to P. aeruginosa-derived factors. Sdc-1 is known to induce production of IL-6 from dendritic cells and to induce IL-6 and TGF- from macrophages (7). However, in our model, acute inflammation is observed within a few hours following burn and infection, sepsis occurs in the first 24 h, and death is noted by 48 h (29, 33). Macrophages and dendritic cells, which are not usually detected at high numbers in burned tissue for several days following thermal injury, are less likely to be responsible for local cytokine production (16). While many types of cells produce these cytokines, likely sources in our model include activated neutrophils, fibroblasts, and/or endothelial cells within the dermal tissue and hepatocytes within the liver. Sdc-1–/– mice display increased vascular permeability and neutrophil extravasation (8). Therefore, we attempted to determine if the neutrophil response of Sdc-1–/– mice to P. aeruginosa was more prominent than that in WT mice. In order to investigate this possibility, we examined skin sections from WT and Sdc-1–/– mice burned and infected with P. aeruginosa. We did not detect differences in the numbers of neutrophils present upon visual examination of H&E-stained sections or sections immunostained with anti-GR-1 (differentiation marker for mouse neutrophils). Sections from both WT and Sdc-1–/– burn wounds displayed a large neutrophil infiltrate (data not shown). In order to determine if the neutrophils present within the tissue were activated, we performed immunohistochemistry with antibody to CD11b, a cell surface antigen present on activated neutrophils. No differences in staining patterns were observed between WT and Sdc-1–/– mice (data not shown). Therefore, both WT and Sdc-1–/– mice elicit a substantial neutrophil response to thermal injury and P. aeruginosa infection. Following extravasation into infected tissue, activated neutrophils release oxidants and hydrolytic enzymes that are intended to aid in eradication of infecting bacteria (36). However, shed Sdc-1 has been shown to bind the neutrophil serine proteases elastase and cathepsin G (12). Therefore, it is possible that shed Sdc-1 binds and inhibits the activity of one or more neutrophil-derived antimicrobial products, thereby negating the protective effects of activated neutrophils. However, this possibility is less likely, considering that we see no difference in local colonization of P. aeruginosa in WT versus Sdc-1–/– mice.
Our data as well as data from other investigators indicate that Sdc-1 shedding is likely to play an important role in the innate immune response following wound and lung infections. Staphylococcus aureus, another important opportunistic pathogen, can also promote Sdc-1 shedding through the action of alpha and beta toxins (21). Interestingly, P. aeruginosa and S. aureus are two of the major pathogens responsible for both lung and skin wound infections (22). Another commonality of these two pathogens is their increasing antibiotic resistance. Two recent studies determined that 14% of P. aeruginosa (38) and 56% of S. aureus (20) bloodstream infections were caused by multidrug-resistant strains. Therefore, the identification of Sdc-1 shedding as a common mechanism of pathogenesis could be exploited to develop alternative treatments for both P. aeruginosa and S. aureus infections, which often occur simultaneously. We have recently shown that treatment with protamine sulfate, which neutralizes the activity moiety of Sdc-1 (HS), had beneficial effects on the survival of thermally injured P. aeruginosa-infected mice (9). In fact, thermally injured, P. aeruginosa-infected WT mice that were treated with protamine sulfate displayed mortality levels consistent with those seen for Sdc-1–/– mice in this study. Protamine-treated mice also displayed significantly lower levels of sepsis and proinflammatory cytokine expression, also consistent with our observations in Sdc-1–/– mice. Protamine sulfate has also successfully been used in treatment of P. aeruginosa infection in a mouse pneumonia model (22). It will be extremely interesting to know if protamine sulfate is effective in inhibiting the pathogenesis of S. aureus in lung and/or burn wound infections. A therapeutic agent effective against two major multidrug-resistant pathogens that targets a component of the host response and not the bacteria itself could be a very valuable clinical tool.
ACKNOWLEDGMENTS
We thank Mary Catherine Hastert for her excellent technical assistance with electron microscopy and David Latham for assisting in immunohistochemistry.
This study was supported by a fellowship from the American Burn Association and Western Medical, Ltd. (A.H.), and HL69050 from the NIH (P.W.P.).
Present address: University of Washington Health Sciences Center, Seattle, WA 98195.
REFERENCES
1. Ausubel, F., R. Brent, R. Kingston, D. Moor, J. Seidman, J. Smith, and K. Stauhle. 1988. Current protocols in molecular biology. Wiley Intersciences, New York, N.Y.
2. Bergan, T. 1981. Pathogenetic factors of Pseudomonas aeruginosa. Scand. J. Infect. Dis. Suppl. 29:7-12.
3. Bernfield, M., M. Gotte, P. W. Park, O. Reizes, M. L. Fitzgerald, J. Lincecum, and M. Zako. 1999. Functions of cell surface heparan sulfate proteoglycans. Annu. Rev. Biochem. 68:729-777.
4. Bown, M. J., M. L. Nicholson, P. R. Bell, and R. D. Sayers. 2001. Cytokines and inflammatory pathways in the pathogenesis of multiple organ failure following abdominal aortic aneurysm repair. Eur. J. Vasc. Endovasc. Surg. 22:485-495.
5. Cakir, B., and B. C. Yegen. 2004. Systemic responses to burn injury. Turkish J. Med. Sci. 34:215-226.
6. Faist, E., and C. Kim. 1998. Therapeutic immunomodulatory approaches for the control of systemic inflammatory response syndrome and the prevention of sepsis. New Horiz. 6:S97-102.
7. Gotte, M. 2003. Syndecans in inflammation. FASEB J. 17:575-591.
8. Gotte, M., A. M. Joussen, C. Klein, P. Andre, D. D. Wagner, M. T. Hinkes, B. Kirchhof, A. P. Adamis, and M. Bernfield. 2002. Role of syndecan 1 in leukocyte-endothelial interactions in the ocular vasculature. Investig. Ophthalmol. Vis. Sci. 43:1135-1141.
9. Haynes, A., III, K. P. Rumbaugh, P. W. Park, A. N. Hamood, and J. A. Griswold. 2005. Protamine sulfate reduces the susceptibility of thermally injured mice to Pseudomonas aeruginosa infection. J. Surg. Res. 123:109-117.
10. Holder, I. A. 1983. Experimental studies of the pathogenesis of infections due to Pseudomonas aeruginosa: effect of treatment with protease inhibitors. Rev. Infect. Dis. 5(Suppl. 5):S914-S921.
11. Holder, I. A. 1993. Pseudomonas aeruginosa virulence-associated factors and their role in burn wound infections, p. 235-245. In R. B. Fick (ed.), Pseudomonas aeruginosa: the oppurtunist. CRC Press, Boca Raton, Fla.
12. Kainulainen, V., H. Wang, C. Schick, and M. Bernfield. 1998. Syndecans, heparan sulfate proteoglycans, maintain the proteolytic balance of acute wound fluids. J. Biol. Chem. 273:11563-11569.
13. Kato, M., H. Wang, V. Kainulainen, M. L. Fitzgerald, S. Ledbetter, D. M. Ornitz, and M. Bernfield. 1998. Physiological degradation converts the soluble syndecan-1 ectodomain from an inhibitor to a potent activator of FGF-2. Nat. Med. 4:691-697.
14. Lentsch, A. B., and P. A. Ward. 2000. Regulation of inflammatory vascular damage. J. Pathol. 190:343-348.
15. Li, Q., P. W. Park, C. L. Wilson, and W. C. Parks. 2002. Matrilysin shedding of syndecan-1 regulates chemokine mobilization and transepithelial efflux of neutrophils in acute lung injury. Cell 111:635-646.
16. McWilliam, A. S., D. Nelson, J. A. Thomas, and P. G. Holt. 1994. Rapid dendritic cell recruitment is a hallmark of the acute inflammatory response at mucosal surfaces. J. Exp. Med. 179:1331-1336.
17. Musher, D. M. 1989. Cutaneous manifestations of bacterial sepsis. Hosp. Pract. 24:71-92.
18. Oliver, J. W., and T. E. Debowski. 1998. Pseudomonas aeruginosa septicemia and associated cutaneous lesions. ASCP, Chicago, Ill.
19. Opal, S. M., and C. E. Huber. 2002. Bench-to-bedside review: Toll-like receptors and their role in septic shock. Crit. Care 6:125-136.
20. Osmon, S., S. Ward, V. J. Fraser, and M. H. Kollef. 2004. Hospital mortality for patients with bacteremia due to Staphylococcus aureus or Pseudomonas aeruginosa. Chest 125:607-616.
21. Park, P. W., T. J. Foster, E. Nishi, S. J. Duncan, M. Klagsbrun, and Y. Chen. 2004. Activation of syndecan-1 ectodomain shedding by Staphylococcus aureus alpha-toxin and beta-toxin. J. Biol. Chem. 279:251-258.
22. Park, P. W., G. B. Pier, M. T. Hinkes, and M. Bernfield. 2001. Exploitation of syndecan-1 shedding by Pseudomonas aeruginosa enhances virulence. Nature 411:98-102.
23. Park, P. W., G. B. Pier, M. J. Preston, O. Goldberger, M. L. Fitzgerald, and M. Bernfield. 2000. Syndecan-1 shedding is enhanced by LasA, a secreted virulence factor of Pseudomonas aeruginosa. J. Biol. Chem. 275:3057-3064.
24. Plotkowski, M. C., A. O. Costa, V. Morandi, H. S. Barbosa, H. B. Nader, S. de Bentzmann, and E. Puchelle. 2001. Role of heparan sulphate proteoglycans as potential receptors for non-piliated Pseudomonas aeruginosa adherence to non-polarised airway epithelial cells. J. Med. Microbiol. 50:183-190.
25. Pollack, M. 2000. Pseudomonas aeruginosa, p. 2310-2327. In J. E. B. G. L. Mandell and R. Dolin (ed.), Mandell, Douglas, and Bennett's principles and practice of infectious diseases. Churchill Livingstone, Philadelphia, Pa.
26. Pruitt, B. A., Jr., A. T. McManus, S. H. Kim, and C. W. Goodwin. 1998. Burn wound infections: current status. World J. Surg. 22:135-145.
27. Reich, H. L., D. Williams Fadeyi, N. S. Naik, P. J. Honig, and A. C. Yan. 2004. Nonpseudomonal ecthyma gangrenosum. J. Am. Acad. Dermatol. 50:S114-S117.
28. Rostand, K. S., and J. D. Esko. 1997. Microbial adherence to and invasion through proteoglycans. Infect. Immun. 65:1-8.
29. Rumbaugh, K. P., J. A. Colmer, J. A. Griswold, and A. N. Hamood. 2001. The effects of infection of thermal injury by Pseudomonas aeruginosa PAO1 on the murine cytokine response. Cytokine 16:160-168.
30. Rumbaugh, K. P., J. A. Griswold, and A. N. Hamood. 1999. Contribution of the regulatory gene lasR to the pathogenesis of Pseudomonas aeruginosa infection of burned mice. J. Burn Care Rehab. 20:42-49.
31. Rumbaugh, K. P., J. A. Griswold, and A. N. Hamood. 2000. The role of quorum sensing in the in vivo virulence of Pseudomonas aeruginosa. Microbes Infect. 2:1721-1731.
32. Rumbaugh, K. P., J. A. Griswold, B. H. Iglewski, and A. N. Hamood. 1999. Contribution of quorum sensing to the virulence of Pseudomonas aeruginosa in burn wound infections. Infect. Immun. 67:5854-5862.
33. Rumbaugh, K. P., A. N. Hamood, and J. A. Griswold. 2004. Cytokine induction by the P. aeruginosa quorum sensing system during thermal injury. J. Surg. Res. 116:137-144.
34. Soave, R., H. W. Murray, and M. M. Litrenta. 1978. Bacterial invasion of pulmonary vessels. Pseudomonas bacteremia mimicking pulmonary thromboembolism with infarction. Am. J. Med. 65:864-867.
35. Stieritz, D. D., and I. A. Holder. 1975. Experimental studies of the pathogenesis of infections due to Pseudomonas aeruginosa: description of a burned mouse model. J. Infect. Dis. 131:688-691.
36. Stockley, R. A. 2004. Chronic obstructive pulmonary disease, neutrophils and bacteria: from science to integrated care pathways. Clin. Med. 4:567-572.
37. Subramanian, S. V., M. L. Fitzgerald, and M. Bernfield. 1997. Regulated shedding of syndecan-1 and -4 ectodomains by thrombin and growth factor receptor activation. J. Biol. Chem. 272:14713-14720.
38. Tacconelli, E., M. Tumbarello, S. Bertagnolio, R. Citton, T. Spanu, G. Fadda, and R. Cauda. 2002. Multidrug-resistant Pseudomonas aeruginosa bloodstream infections: analysis of trends in prevalence and epidemiology. Emerg. Infect. Dis. 8:220-221.
39. Tredget, E. E., H. A. Shankowsky, R. Rennie, R. E. Burrell, and S. Logsetty. 2004. Pseudomonas infections in the thermally injured patient. Burns 30:3-26.
40. Ward, P. A., and A. B. Lentsch. 1999. The acute inflammatory response and its regulation. Arch. Surg. 134:666-669.(Allan Haynes III, Frank R)
Department of Medicine, Baylor College of Medicine, Houston, Texas 77030
ABSTRACT
The innate immune system is comprised of many components that function coordinately to prevent bacterial sepsis. However, thermal injury suppresses many of these factors, and the opportunistic pathogen Pseudomonas aeruginosa takes advantage of this condition, making it one of the leading causes of morbidity and mortality in the setting of thermal injury. P. aeruginosa is extremely efficient at colonizing burn wounds, spreading systemically, and causing sepsis, which often results in a systemic inflammatory response, multiple-organ failure, and death. The pathogenicity of P. aeruginosa is due to the arsenal of virulence factors produced by the pathogen and the immunocompromised state of the host. Syndecan 1 is a major heparan sulfate proteoglycan present on many host cells involved in thermal injury. Syndecan 1 anchored to the cell surface can be cleaved in a process termed ectodomain shedding. Syndecan 1 shedding results in the release of intact, soluble proteoglycan ectodomains that have diverse roles in innate immunity. Here we show for the first time that thermal injury results in shedding of syndecan 1 from host tissue. Our data show that syndecan 1 null mice are significantly less susceptible to P. aeruginosa infection than their wild-type counterparts, as demonstrated by (i) significantly lower mortality; (ii) absence of systemic spread of P. aeruginosa; and (iii) significant reductions in some proinflammatory cytokines. These results suggest that shed syndecan 1 plays an important role in the pathogenesis of P. aeruginosa infection of thermal injury and that syndecan 1-neutralizing agents may be effective supplements to current P. aeruginosa treatments.
INTRODUCTION
The gram-negative opportunistic pathogen Pseudomonas aeruginosa is particularly prevalent in burn wound infections. P. aeruginosa bacteremia is one of the leading causes of morbidity and mortality in the setting of thermal injury. It often causes multiorgan failure and results in mortality rates of up to 50% (2, 25, 26, 39). The extensive damage induced during P. aeruginosa infection is due to the ability of the microorganism to produce a large array of virulence factors (25) and the immunocompromised state of the individual. However, the increased ability of P. aeruginosa over many other pathogens to cause sepsis in thermally injured individuals is not fully understood.
Beyond the initial barrier of the skin, several components of the host's innate immune system are designed to act as sentinels that guard against systemic bacterial attacks. Toll-like receptors (TLRs) are transmembrane receptors found on host cells that recognize pathogen-associated molecular patterns (19). Upon recognition of a pathogen-associated molecular pattern, TLRs activate downstream signaling events, which induce an immediate inflammatory response (19). Acute inflammation, as seen in thermal injury, is characterized by vasodilatation and neutrophil infiltration. Consequences of inflammation include the production of cytokines and chemokines. Chemokines, such as interleukin-8 (IL-8), recruit neutrophils into the damaged area, which normally engulf the infecting bacteria before high numbers are reached. Proinflammatory cytokines, such as IL-1, IL-6, and tumor necrosis factor alpha (TNF-), are produced predominately by activated macrophages. Among their many roles, cytokines regulate neutrophil functions and perpetuate the inflammatory response. Although these first lines of defense are normally effective, the highly devascularized environment of a burn wound and overall dysregulation of immunity in burned individuals provide an ideal setting for the proliferation and spread of bacteria within the injured tissue and subsequent dissemination throughout the immunocompromised host. In this environment, inflammation can shift from a protective process to a destructive one. The massive neutrophil infiltration, characteristic of thermal injury, may actually have a negative biological effect. Activated neutrophils release oxidants and hydrolytic enzymes that, when combined with bacterial degradative enzymes, can result in increased proteolysis within the wound and may actually aid spread and eventual dissemination of bacteria (14, 33). This neutrophil paradox has been implicated in the pathology of many chronic inflammatory conditions (36).
Dozens of inflammatory mediators play important roles in balancing the protective and destructive effects of inflammation. One important group of inflammatory mediators are heparan sulfate proteoglycans (HSPGs) (7). HSPGs are ubiquitous molecules expressed by all adherent cells and are composed of a core protein and one or more covalently attached heparan sulfate (HS) chains (7). The HS chains are made up of repeating units of hexuronic acid and N-substituted glucosamine disaccharide units. One family of HSPGs implicated in inflammation is the syndecans. Four syndecan family members (syndecan 1 to 4) have been cloned in mammals and are implicated in mediating a wide range of effects, including leukocyte rolling on endothelial cells, wound repair, angiogenesis, vascular permeability, and modulation of chemokine activity (7). Syndecans attached to the cell surface can be cleaved in a process termed shedding. Syndecan shedding results in the release of intact, soluble proteoglycan ectodomains into the host cell environment that can mediate cellular physiology at distant sites (7). Although syndecan ectodomains are shed at low levels as part of normal cell turnover, tissue injury can markedly activate syndecan shedding (12). Thus, shed syndecan 1 (Sdc-1) is not found at detectable levels in the biological fluids of healthy individuals, but high levels of shed Sdc-1 and Sdc-4 have been documented in human dermal wound fluids (37) and in tracheal aspirates of ventilated preterm infants (13).
Syndecan shedding during tissue injury is presumably due to mechanical shearing of the ectodomains and/or cleavage by host cell-derived proteases (12). However, it has also been shown that the LasA protease produced by P. aeruginosa activates syndecan 1 shedding (23). Shed Sdc-1 contributed to P. aeruginosa pathogenesis in a mouse model of lung infection (22). Sdc-1 knockout (Sdc-1–/–) mice were significantly more resistant to P. aeruginosa lung infection than their wild-type (WT) counterparts, as demonstrated by decreased mortality and bacteremia (22). Furthermore, treating P. aeruginosa-infected WT mice with specific inhibitors of Sdc-1 shedding and HS was therapeutic (22). Considering the demonstrated importance of shed Sdc-1 to P. aeruginosa pathogenesis in the lung, and assuming that the significant mechanical damage that is inflicted on host tissue during thermal injury results in high levels of shed Sdc-1, we hypothesized that shed Sdc-1 is a decisive host factor contributing to P. aeruginosa pathogenesis in burn wound infections. Here we report that Sdc-1–/– mice are significantly less susceptible to P. aeruginosa burn wound infections than WT mice and that shed Sdc-1 may play a role in the blood vessel invasion and subsequent systemic spread of P. aeruginosa. These data provide support for the prospect of using Sdc-1 neutralizing agents in the treatment of burn wounds.
MATERIALS AND METHODS
Mice. Sdc-1–/– mice were generated and bred as previously described (22). Sdc-1–/– mice used in this study were backcrossed eight times onto the C57BL/6 background and were maintained by interbreeding at Baylor College of Medicine (Houston, Tex.). WT Sdc-1+/+ C57BL/6 mice were obtained from Charles River Laboratories (Wilmington, MA) or were derived from crossing mice heterozygous for the syndecan 1 gene. Mice used in experiments were 6 to 8 weeks old and weighed 17 to 20 g. Mice were housed and studied under protocols approved by the Institutional Animal Care and Use Committee in the animal facility of Texas Tech University Health Sciences Center (Lubbock, TX).
Bacterial growth and inoculum. P. aeruginosa strain PAO1 was grown in Luria-Bertani (LB) medium (1). Aliquots (50 μl) of overnight cultures of PAO1 were subcultured in fresh LB broth and grown at 37°C for 4 h to an optical density at 540 nm of approximately 0.9. A 100-μl aliquot of each culture was then pelleted, washed in phosphate-buffered saline (PBS), and serially diluted (10-fold serial dilutions) in PBS. A 100-μl aliquot of the 10–4 dilution was injected into each animal. As we have previously determined (9), this dilution contains approximately 2 x 102 to 2 x 103 CFU of P. aeruginosa. We have also shown previously that this dose of PAO1 produces 90 to 100% lethality in WT C57BL/6 mice by 48 h postburn infection (9). The exact inoculum of each strain was determined by plating serial dilutions of the inoculum on LB agar plates.
Thermally injured mouse model. The virulence of PAO1 was examined by using the modified burned-mouse model of Stieritz and Holder (35). In this modified model, a scald burn is induced. Mice were anesthetized by intraperitoneal injection of 0.4 ml of Nembutal/20 g of body weight at 5 mg/ml (5% sodium pentobarbital; Abbott Laboratories, North Chicago, Ill.), and their backs were shaved. The mice were securely placed into a template with an opening (4.5 by 1.8 cm) exposing their shaved backs. About 15% of the total surface area of the mouse was exposed through the opening on the template. The thermal injury was induced by placing the exposed area of the shaved skin in 90°C water for 10 s. Such an injury is nonlethal but causes a third-degree (full-thickness) burn. Fluid replacement therapy consisting of a subcutaneous injection of 0.8 ml of a 0.9% NaCl solution was administered immediately following the burn. Mice were challenged by the subcutaneous inoculation of 100 μl of the bacterial inoculum (see above) directly under the burn. Control mice were subcutaneously injected with 100 μl of sterile phosphate-buffered saline (PBS) directly under the burn. In some experiments, porcine mucosal heparan sulfate (Sigma, St. Louis, Mo.) dissolved in PBS was simultaneously injected under the burn with the bacterial inoculum. During recovery, the mice were continuously observed under warming lights.
Mortality among infected mice was recorded at 24 h intervals, up to 5 days postburn and postinfection (postburn/infection). Animals were treated humanely and in accordance with the protocol approved by the Animal Care and Use Committee at Texas Tech University Health Sciences Center (Lubbock, TX).
Quantitation of bacteria within the skin and livers. At 24 h postburn/infection, mice were euthanized by intracardial injection of 0.2 ml of Sleepaway (sodium pentobarbital-7.8% isopropyl alcohol euthanasia solution; Fort Dodge Laboratories, Inc., Fort Dodge, Iowa). Skin sections of approximately 5 by 5 mm were obtained from the burned skin of both the control and challenged mice. Simultaneously the entire liver of each animal was obtained. Tissues were weighed, suspended in 2 ml of PBS, and homogenized (overhead stirrer; Wheaton Instruments, Millville, N.J.). Homogenates were serially diluted, and a 100-μl aliquot of each diluent was plated on LB agar plates to determine the number of CFU, which was calculated per gram of tissue.
Immunoblotting experiments. Skin sections from control (nonburned) mice, mice that had been administered a third-degree scald burn, and mice that were burned and infected with P. aeruginosa (PAO1) were harvested at 24 h postburn. The tissue was placed in phosphate-buffered saline (PBS) and gently vortexed for 10 s. Five-microliter samples of the wound fluid extracts were then dot blotted in triplicate onto a cationic nylon membrane (Immobilon Ny+; Millipore) and quantified by dot immunoblotting with the 281-2 rat anti-mouse Sdc-1 ectodomain monoclonal antibody as described previously (21, 23). The same blot was stripped and then reprobed with an antibody directed against the cytoplasmic domain of Sdc-1 (21) to confirm that Sdc-1 detected by the 281-2 antibody was shed ectodomain and not intact Sdc-1 released from cells damaged by vortexing.
Hematoxylin and eosin (H&E) staining. Mouse skin was harvested and immediately placed into 4% paraformaldehyde in phosphate buffer, pH 7.3, and fixed overnight at 4°C. The tissue was then processed through a graded series of ethanol, cleared in xylene, and infiltrated with paraffin. The tissue was embedded on edge in paraffin, and blocks were sectioned with a Cut 4055 Olympus microtome. Five-micrometer sections of the blocks were collected on Superfrost/Plus slides (Fisher Scientific) and H&E stained using standard laboratory techniques. Slides were visualized by light microscopy at 400x with an Olympus BH-2 microscope and imaged with an Olympus camera model C-35AD-4 (Tokyo, Japan).
Scanning electron microscopy. Mouse skin that had been burned and infected with P. aeruginosa was harvested, cut into 1-mm-thick cross-sections, and placed into fixative containing 1.5% glutaraldehyde and 3% paraformaldehyde in 0.1 M Millonig's buffer with dextrose. Mouse skin that had been burned but not infected was treated in a similar manner to use as a control. The tissues were fixed overnight at 4°C. Tissue was then frozen in liquid nitrogen and fractured with a frozen blade to expose more surface. Tissues were then washed in buffer, postfixed in 1% osmium tetroxide, and taken through a graded series of alcohols. The tissue was critically point dried with a Tousimis critical point dryer (Rockville, MD). The sections were mounted on edge on stubs coated with carbon tape. Some of the infected skin tissue samples were carved using a razor blade under a dissecting scope to reveal more of the inner structure. All the tissues were gold coated using a Polaron sputter coater. The sections were then scanned for bacteria, and pictures were taken using a Hitachi S-500 scanning electron microscope (Hitachi America, Ltd., Brisbane, CA).
Cytokine expression. Total RNA from mouse skin sections was extracted using TRI Reagent-RNA/DNA/Protein isolation reagent per the manufacturer's guidelines (Molecular Research Center, Inc., Cincinnati, OH). RNA was treated with RQ1 RNase-Free DNase (Promega, Madison, WI) to remove any contaminating genomic DNA. A total of 20 μg of DNA-free RNA was used to detect the expression of murine cytokine genes using the mck4 and mck3b Multi-Probe Template Sets and the Riboquant RNase Protection Assay System (PharMingen, San Diego, CA). The probes were hybridized overnight at 56°C with 20 μg of total RNA from either the liver or the skin. The mixtures were then RNase treated, purified, and resolved on a 6% acrylamide-8 M urea gel per the manufacturer's guidelines. Gels were dried, and protected bands representing different cytokines were detected after exposure of the gels to a phosphorimager screen. Radioactive signals were imaged using a Typhoon variable mode imager (Amersham Biosciences, Piscataway, NJ). Signals were quantitated using Image-Quant TL (Amersham Biosciences, Piscataway, NJ). The signal of the band representing a cytokine of interest was normalized to the signal of the L32 (ribosomal gene) band from the same sample.
Statistical analysis. The Student's t test and the Fisher's exact test (Statview; Abacus Concepts, Inc., Berkeley, CA) were used to determine significant differences between the numbers of CFU/gram of tissue in bacterial colonization experiments and between groups for the mortality experiments, respectively. Significant differences between cytokine mRNA levels in the RNase protection assays were determined by the Student's t test (Microsoft Excel 2002, Austin, TX).
RESULTS
Thermally injured syndecan 1 knockout mice are less susceptible to P. aeruginosa infection. Tissue injury activates syndecan shedding (12), and shed Sdc-1 is found at high levels in human dermal wound fluids (37) and in tracheal aspirates of ventilated preterm infants (13). However, Sdc-1 has not been demonstrated in burn wounds. To determine if thermal injury results in Sdc-1 shedding, Sdc-1 ectodomain levels in burn fluid were measured by dot immunoblotting. Skin sections from control (nonburned) mice, mice that had been administered a third-degree scald burn, and mice that were burned and infected with P. aeruginosa (PAO1) were harvested at 24 h postburn. The tissue was placed in phosphate-buffered saline (PBS) and gently vortexed for 10 s. Samples were then dot blotted onto a cationic nylon membrane and quantified by immunoblotting with an Sdc-1 ectodomain antibody. Sdc-1 was detected in the fluid extract of burned, but not nonburned, skin; however, increased levels of Sdc-1 were detected in the wound fluid extracts of mice that were burned and infected with PAO1 (Fig. 1). Sdc-1 ectodomains were not detected when the membrane was reprobed with an antibody against the cytoplasmic domain of Sdc-1, confirming that Sdc-1 was shed and not released intact from cells (data not shown).
To determine if the increased levels of shed Sdc-1 that are associated with thermal injury correlated with the severity of P. aeruginosa infection, we utilized Sdc-1–/– mice, which carry a deletion in the gene encoding Sdc-1, and the thermally injured mouse model. Sdc-1–/– mice are fertile and develop normally (8). The thermally injured mouse model of P. aeruginosa infection was originally described by Stieritz and Holder and closely resembles the pathology of human thermal injury (35). We have previously determined that the P. aeruginosa strain PAO1 causes approximately 90% mortality to thermally injured mice by 48 h postburn/infection (30, 31).
Groups of WT or Sdc-1–/– mice on the C57BL/6 background were given full-thickness scald burns and infected with PAO1. The mice were observed for 120 h (5 days) postburn/infection, and percent mortality was recorded for each group (Fig. 2). Eighty-two percent (18 out of 22) of the burned and infected C57Bl/6 WT mice died in the first 48 h, and no further mortality was observed. However, the percent mortality of Sdc-1–/– mice was significantly reduced at all time points (Fig. 2; 13% [2 out of 15] at 48 h [P = 0.00005] and 40% [6 out of 15] at 120 [P = 0.014]).
Shed Sdc-1, via its HS chains, can function as a soluble effector to regulate various cellular processes, including proliferation, adhesion, and differentiation (3). In a P. aeruginosa lung infection model, HS was as effective as purified shed Sdc-1 at increasing the susceptibility of Sdc-1–/– mice to lung infection from P. aeruginosa (22). To determine if the absence of shed Sdc-1 was responsible for the lower susceptibility of burned Sdc-1–/– mice to P. aeruginosa infection, we tested whether HS could increase mortality. Sdc-1–/– mice were burned and infected with PAO1 as before; however, some groups of mice were also administered a single 20-μg, 50-μg, or 100-μg dose of HS locally at the injury site. These doses were chosen because we have previously determined that 100 μg of protamine sulfate is required to neutralize syndecan 1 in vivo (9). As before, the percent mortality at 48 h in Sdc-1–/– mice infected with PAO1 was relatively low 25% (2 out of 8 mice). However, the percent mortality was significantly higher (83%, 5 out of 6 mice; P = 0.05; for 100 μg) in mice that received a simultaneous HS injection and increased to 100% by the 72-h time point (Fig. 3). The percent mortality among HS-injected mice was similar to that seen for WT mice that possess Sdc-1 (compare Fig. 2 to 3). No mortality was observed in nonburned mice that were given 100 μg HS or infected with PAO1 plus 100 μg HS (data not shown). Taken together, these results indicate that the high levels of Sdc-1 shed due to thermal injury exacerbate the virulence of P. aeruginosa.
Shed syndecan 1 facilitates the systemic spread of P. aeruginosa. P. aeruginosa infection in the murine thermal injury model can be divided into two phases (10, 11). The first phase involves colonization of the burned tissue by P. aeruginosa. The devascularized state of the burn eschar (burn wound area) provides an environment ideally suited for bacterial growth, and P. aeruginosa proliferates extremely quickly. Once a threshold concentration of bacteria is reached in the eschar (approximately 109 CFU/g tissue), P. aeruginosa spreads systemically, causing sepsis followed by multiple-organ failure and eventually death (10, 11). We have previously seen that by 24 h postburn/infection, P. aeruginosa CFU in the eschar increase from 102 CFU/g tissue (infecting dose) to approximately 109 CFU/g tissue and can also be isolated from the liver and spleen (29, 32). By 48 h postburn/infection, the majority of mice succumb to the infection (29, 32).
To determine if Sdc-1 influenced either stage of P. aeruginosa infection, we examined both the local colonization within the eschar and the systemic spread of P. aeruginosa in burned and infected WT and Sdc-1–/– mice. We examined the local colonization of bacteria within the eschar by determining the number of PAO1 CFU in epithelial sections taken at two specific sites within the area of thermal injury, the inoculation site and a site 15 mm distal to the inoculation site. Using this approach, we have previously shown that PAO1 spreads efficiently from the inoculation site to the distal site by 24 h postburn (32). As in the mortality experiments, mice were thermally injured and inoculated with PAO1. At 24 h postburn/infection, the mice were euthanized. Epithelial sections were isolated and homogenized, and the number of PAO1 CFU was determined as described in Materials and Methods. WT and Sdc-1–/– mice had similar P. aeruginosa CFU counts within both the "inoculation" and "distal" epithelial sections (Fig. 4), and administration of HS did not affect local spread. These results indicate that within 24 h, P. aeruginosa is able to efficiently multiply and colonize the burn eschar despite the lack of Sdc-1.
To examine the systemic spread of P. aeruginosa, mice were thermally injured and inoculated with PAO1 as described above. At 24 h postinfection, the mice were euthanized, the livers were harvested, and the CFU/gram of tissue was determined. On average, 3.4 x 105 ± 9 x 104 CFU were recovered from livers of WT mice (5 mice/group), while no bacteria were recovered from livers of Sdc-1–/– mice (5 mice/group). Administration of heparan sulfate, as described above, resulted in similar CFU counts as seen in WT mice (9 x 103 ± 5 x 103). Taken together, these results indicate that P. aeruginosa can efficiently colonize the burn eschar and multiply to threshold levels despite the absence of Sdc-1. However, Sdc-1 appears to be needed for dissemination of PAO1 into the bloodstream and secondary colonization of internal organs.
Syndecan 1 is involved in P. aeruginosa perivascular cuffing. Two distinctive clinical features of P. aeruginosa bacteremia are invasion and necrosis of blood vessels (34). In humans, blood vessel invasion by P. aeruginosa correlates with the appearance of skin lesions termed ecthyma gangrenosum (17). Ecthyma gangrenosum is classically considered a pathognomonic sign of sepsis by P. aeruginosa (17, 18). However, a few other bacterial species have now been associated with ecthyma gangrenosum (27). Upon microscopic examination, blood vessel invasion by P. aeruginosa can be recognized by the presence of perivascular cuffing (PVC) (34). Bacilli form a circumferential pattern surrounding the vessel and are often aligned single file or in stacks between muscle cells of the venous walls (34). The mechanisms of PVC by P. aeruginosa and its role in pathogenesis are not fully understood. However, in the thermally injured mouse model, we have consistently seen that PVC accompanies sepsis and subsequent death from P. aeruginosa infection (K. P. Rumbaugh, unpublished observations).
As Sdc-1 is a component of normal vascular endothelium and Sdc-1–/– mice displayed significantly less sepsis from P. aeruginosa infection, we wanted to determine if Sdc-1 played a role in PVC. Groups of mice were thermally injured and infected with PAO1 as described in Materials and Methods. At 24 h postburn/infection, mice were euthanized and skin sections were harvested for histological evaluation. Sections measuring approximately 1 by 1.5 cm were extracted from the burn margin (the area where burned and healthy tissue merge). PVC was visualized in hematoxylin and eosin (H&E)-stained sections by the appearance of circumferentially arranged gram-negative bacilli around one or more blood vessels (Fig. 5A). PVC was usually seen in association with vessel thrombosis or blood clotting within the vessel, as seen in Fig. 5A. PVC was observed in skin sections from 4 of 5 thermally injured, P. aeruginosa-infected WT mice. However, none of the five sections obtained from Sdc-1–/– mice displayed PVC. Although bacilli could be seen in tissue sections from Sdc-1–/– mice, they were dispersed throughout the section and the blood vessels appeared normal without thrombosis (Fig. 5B). Sdc-1–/– mice that were administered HS in addition to PAO1 also displayed extensive PVC (data not shown). Skin tissue from thermally injured WT mice infected with PAO1 was also visualized by electron microscopy. Scanning electron microscopy (SEM) revealed organized arrangements of P. aeruginosa bacilli in areas surrounding blood vessels (Fig. 5C).
Syndecan 1 modulates the host cytokine response to P. aeruginosa infection in thermally injured mice. Extensive inflammation is an important initiating event in thermal injury that predisposes the host to sepsis, multiorgan failure, and death (6, 40). We have previously shown that the combination of thermal injury and P. aeruginosa infection results in an acute and exaggerated inflammatory response that is characterized by elevated mRNA levels of several cytokine and chemokine genes (29). This "cytokine storm" can be detected locally in the skin and systemically within the livers of burned and PAO1-infected mice by 24 h postburn/infection (29). Shed Sdc-1 is known to regulate the expression and function of many cytokines (7). Sdc-1 stimulates the release of TNF-, IL-1, and IL-6 from dendritic cells and IL-1, IL-6, TNF-, IL-12, transforming growth factor- (TGF-), and prostaglandin E2 (PGE2) from macrophages (7), and it is known to play a key role in IL-8-mediated chemotaxis of neutrophils to sites of tissue injury (15). As Sdc-1 shedding on epithelial cells accompanies thermal tissue injury, we wanted to determine if Sdc-1 was a mediator in the extreme upregulation of cytokines seen at the onset of thermal injury. Therefore, we examined the mRNA levels of 19 murine cytokine and chemokine genes in both the skin and livers of thermally injured, PAO1-infected WT or Sdc-1–/– mice. The murine genes examined were TNF-, TNF-, IL-3, IL-6, IL-7, IL-11, TGF-1, TGF-2, TGF-3, gamma interferon (IFN-) and IFN-, lymphotoxin (LT), macrophage migration inhibitory factor (MIF), granulocyte-macrophage colony-stimulating factor (GM-CSF), macrophage colony-stimulating factor (M-CSF), granulocyte colony-stimulating factor (G-CSF), leukocyte inhibitory factor (LIF), and stem cell factor (SCF).
WT and Sdc-1–/– mice were thermally injured and infected with PAO1. Mice were euthanized at 24 h postburn/infection, and RNA was harvested from skin and liver sections as described in Materials and Methods. RNase protection assays (RPA) were run using the Multi-Probe RPA System (BD RiboQuant). Transcripts for many of the cytokines and chemokines examined were detected in the skin of both WT and Sdc-1–/– mice. As we have previously reported (29), MIF, M-CSF, TGF-, TNF-, IL-6, IFN-, IFN-, LIF, and G-CSF were all expressed in thermally injured, PAO1-infected skin (data not shown). However, the mRNA levels of G-CSF, LIF, and IL-6 were significantly decreased in the skin of Sdc-1–/– mice (Fig. 6A). In this study, and in agreement with our previous studies (29), the only cytokines from those examined that were detected in the livers of burned and infected mice were MIF, TGF-, G-CSF, and GM-CSF. However, Sdc-1–/– mice displayed significantly lower levels of MIF, TGF-, and G-CSF (Fig. 6B and C). Taken together, these data indicate that despite the absence of Sdc-1, Sdc-1–/– mice display a significant local inflammatory cytokine response; however, some of the cytokines that are characteristic of PAO1 infection are reduced. Conversely, the systemic inflammatory response, as illustrated by cytokine expression in the livers, was almost completely absent in Sdc-1–/– mice (Fig. 6B and C). This is consistent with the lack of systemic spread observed in Sdc-1–/– mice.
DISCUSSION
Serious thermal injuries result in dysregulation of immune responses and a hypermetabolic state that can cause multiple-organ failure and predispose patients to bacterial infections (5). P. aeruginosa is one of the most common and pathogenic bacteria encountered in burn wound infections (26, 39). P. aeruginosa can rapidly colonize burn wounds, resulting in extensive inflammation, bacteremia, and high mortality. As thermal injury causes a complicated clinical scenario, it is likely that there are many host and bacterial factors that contribute to P. aeruginosa pathogenesis in burn wounds. The goal of this study was to examine the contribution of one such host factor, Sdc-1, in P. aeruginosa infection of burn wounds. Taken together, our data indicate that Sdc-1 increases the susceptibility of thermally injured mice to P. aeruginosa infection. We show that Sdc-1–/– mice display reduced inflammatory cytokine expression, sepsis, and mortality upon a P. aeruginosa infection of a burn wound in comparison to WT mice.
Tissue injury activates shedding of Sdc-1, and high levels of shed Sdc-1 have been documented in human dermal wound fluids (37). In this study we demonstrate for the first time in vivo that Sdc-1 is shed by the trauma of thermal injury, and shedding is increased further by P. aeruginosa infection (Fig. 1). Although the LasA protease produced by P. aeruginosa can promote Sdc-1 shedding (23), we did not observe a reduction in mortality in thermally injured mice infected with a PAO1::lasA isogenic mutant (unpublished data). This indicates to us that the levels of shed Sdc-1, which are already increased due to the thermal injury, are not enhanced by the action of LasA. However, it is possible that in a P. aeruginosa lung infection where there are no reservoirs of shed Sdc-1, a LasA mutant would display reduced virulence. Consistent with this hypothesis, we detected >10 ng/ml of Sdc-1 ectodomains in tissues that were gently vortexed for 10 s, suggesting that a significantly larger concentration of Sdc-1 is present in thermally injured tissue.
Our results are consistent with previous observations suggesting that P. aeruginosa exploits shed Sdc-1 to enhance its virulence in the lung (22). Similar to data reported by Park etal., we see a significant decrease in mortality of P. aeruginosa-infected, thermally injured Sdc-1–/– mice, and this reduction can be compensated for by the addition of HS. Park et al. reported that P. aeruginosa was reduced in the primary (lung) and secondary (spleen and liver) infection sites in their pneumonia model (22). However, in the thermally injured mouse, we observed that P. aeruginosa was not inhibited from efficiently colonizing the primary infection site (burn wound) but could not spread systemically. It is possible that this inconsistency is due to the devascularized state of the burn wound that allows an uncontrolled proliferation of P. aeruginosa. The infected lung maintains a sufficient blood supply that provides access for immune system mediators to the area. It is possible that more subtle contributions of Sdc-1 to the actions of these mediators are more easily observed in the lung infection model. Despite differences in the local colonization of P. aeruginosa, both studies observed significant decreases in the systemic spread of P. aeruginosa in Sdc-1–/– mice (this study and reference 22). Furthermore, in both studies systemic spread was returned to WT levels by the addition of HS. Taken together, these observations provide strong evidence that the mechanism by which shed Sdc-1 enhances P. aeruginosa pathogenesis is related to the ability of P. aeruginosa to disseminate and cause sepsis.
We observed prevalent PVC of P. aeruginosa in skin sections from WT mice, but no PVC was seen in sections from Sdc-1–/– mice (Fig. 5A and B). As PVC is generally accompanied by blood vessel invasion (17, 34), it is possible that Sdc-1 shedding facilitates blood vessel invasion in thermally injured mice. This possibility is consistent with the lack of P. aeruginosa systemic spread in Sdc-1–/– mice and could explain the reduction in mortality. However, the mechanisms by which shed Sdc-1 enhances blood vessel invasion are unknown at this time. One possible explanation is that the transmembrane core protein that remains in the vascular endothelial cell membrane after shedding acts as an attachment site for P. aeruginosa. Some bacteria, including P. aeruginosa, use HSPGs for adherence and invasion; however, these interactions typically rely on bacterial adherence to the anionic HS side chains and not to the core protein (24, 28). Furthermore, Park et al. determined that P. aeruginosa did not bind to 125I- or 25S-labeled Sdc-1 ectodomains (22). Interestingly, the phenomenon of PVC is almost exclusive to the Pseudomonas and Aeromonas species (17) and has been relatively understudied. The correlation of PVC with sepsis and mortality implicates it as an important mechanism for P. aeruginosa pathogenesis, especially in wounds. The striking absence of PVC in Sdc-1–/– mice leads us to propose that there are important interactions between P. aeruginosa and Sdc-1 itself or an Sdc-1-regulated factor on vascular endothelial cells that deserve further investigation.
Upon thermal injury and P. aeruginosa infection, we detected significant reductions in the mRNA levels for G-CSF, LIF, and IL-6 in the dermal tissue and G-CSF, TGF-3, and MIF in the livers of Sdc-1–/– mice in comparison to WT mice (Fig. 6). However, despite these reductions, Sdc-1–/– mice displayed a significant cytokine response to the thermal injury and subsequent P. aeruginosa infection. Most of the 19 murine cytokine and chemokine genes we examined were detected in both WT and Sdc-1–/– mice, including the important proinflammatory cytokines IFN- and TNF-. The cytokine response elicited in thermally injured Sdc-1–/– mice by P. aeruginosa may help explain why some mice succumbed to the infection in the absence of systemic spread of P. aeruginosa. High levels of circulating cytokines are associated with poor clinical outcomes in trauma patients (4). Circulating proinflammatory cytokines can cause dysfunction of the renal, cardiovascular, respiratory, nervous, and musculoskeletal systems, resulting in systemic inflammatory response syndrome and eventual multiple-organ failure (4).
Significant P. aeruginosa CFU were detected in the livers of WT but not Sdc-1 mice (Fig. 4). Therefore, the most likely explanation for the difference in the systemic inflammatory response to PAO1 in WT versus Sdc-1–/– mice is the absence of significant numbers of bacteria within the livers of Sdc-1–/– mice. However, the differences in the local inflammatory response are less apparent. Similar numbers of P. aeruginosa were detected in skin sections from WT and Sdc-1–/– mice (Fig. 4), therefore the reductions in IL-6, LIF, and G-CSF message are not likely to be due to P. aeruginosa-derived factors. Sdc-1 is known to induce production of IL-6 from dendritic cells and to induce IL-6 and TGF- from macrophages (7). However, in our model, acute inflammation is observed within a few hours following burn and infection, sepsis occurs in the first 24 h, and death is noted by 48 h (29, 33). Macrophages and dendritic cells, which are not usually detected at high numbers in burned tissue for several days following thermal injury, are less likely to be responsible for local cytokine production (16). While many types of cells produce these cytokines, likely sources in our model include activated neutrophils, fibroblasts, and/or endothelial cells within the dermal tissue and hepatocytes within the liver. Sdc-1–/– mice display increased vascular permeability and neutrophil extravasation (8). Therefore, we attempted to determine if the neutrophil response of Sdc-1–/– mice to P. aeruginosa was more prominent than that in WT mice. In order to investigate this possibility, we examined skin sections from WT and Sdc-1–/– mice burned and infected with P. aeruginosa. We did not detect differences in the numbers of neutrophils present upon visual examination of H&E-stained sections or sections immunostained with anti-GR-1 (differentiation marker for mouse neutrophils). Sections from both WT and Sdc-1–/– burn wounds displayed a large neutrophil infiltrate (data not shown). In order to determine if the neutrophils present within the tissue were activated, we performed immunohistochemistry with antibody to CD11b, a cell surface antigen present on activated neutrophils. No differences in staining patterns were observed between WT and Sdc-1–/– mice (data not shown). Therefore, both WT and Sdc-1–/– mice elicit a substantial neutrophil response to thermal injury and P. aeruginosa infection. Following extravasation into infected tissue, activated neutrophils release oxidants and hydrolytic enzymes that are intended to aid in eradication of infecting bacteria (36). However, shed Sdc-1 has been shown to bind the neutrophil serine proteases elastase and cathepsin G (12). Therefore, it is possible that shed Sdc-1 binds and inhibits the activity of one or more neutrophil-derived antimicrobial products, thereby negating the protective effects of activated neutrophils. However, this possibility is less likely, considering that we see no difference in local colonization of P. aeruginosa in WT versus Sdc-1–/– mice.
Our data as well as data from other investigators indicate that Sdc-1 shedding is likely to play an important role in the innate immune response following wound and lung infections. Staphylococcus aureus, another important opportunistic pathogen, can also promote Sdc-1 shedding through the action of alpha and beta toxins (21). Interestingly, P. aeruginosa and S. aureus are two of the major pathogens responsible for both lung and skin wound infections (22). Another commonality of these two pathogens is their increasing antibiotic resistance. Two recent studies determined that 14% of P. aeruginosa (38) and 56% of S. aureus (20) bloodstream infections were caused by multidrug-resistant strains. Therefore, the identification of Sdc-1 shedding as a common mechanism of pathogenesis could be exploited to develop alternative treatments for both P. aeruginosa and S. aureus infections, which often occur simultaneously. We have recently shown that treatment with protamine sulfate, which neutralizes the activity moiety of Sdc-1 (HS), had beneficial effects on the survival of thermally injured P. aeruginosa-infected mice (9). In fact, thermally injured, P. aeruginosa-infected WT mice that were treated with protamine sulfate displayed mortality levels consistent with those seen for Sdc-1–/– mice in this study. Protamine-treated mice also displayed significantly lower levels of sepsis and proinflammatory cytokine expression, also consistent with our observations in Sdc-1–/– mice. Protamine sulfate has also successfully been used in treatment of P. aeruginosa infection in a mouse pneumonia model (22). It will be extremely interesting to know if protamine sulfate is effective in inhibiting the pathogenesis of S. aureus in lung and/or burn wound infections. A therapeutic agent effective against two major multidrug-resistant pathogens that targets a component of the host response and not the bacteria itself could be a very valuable clinical tool.
ACKNOWLEDGMENTS
We thank Mary Catherine Hastert for her excellent technical assistance with electron microscopy and David Latham for assisting in immunohistochemistry.
This study was supported by a fellowship from the American Burn Association and Western Medical, Ltd. (A.H.), and HL69050 from the NIH (P.W.P.).
Present address: University of Washington Health Sciences Center, Seattle, WA 98195.
REFERENCES
1. Ausubel, F., R. Brent, R. Kingston, D. Moor, J. Seidman, J. Smith, and K. Stauhle. 1988. Current protocols in molecular biology. Wiley Intersciences, New York, N.Y.
2. Bergan, T. 1981. Pathogenetic factors of Pseudomonas aeruginosa. Scand. J. Infect. Dis. Suppl. 29:7-12.
3. Bernfield, M., M. Gotte, P. W. Park, O. Reizes, M. L. Fitzgerald, J. Lincecum, and M. Zako. 1999. Functions of cell surface heparan sulfate proteoglycans. Annu. Rev. Biochem. 68:729-777.
4. Bown, M. J., M. L. Nicholson, P. R. Bell, and R. D. Sayers. 2001. Cytokines and inflammatory pathways in the pathogenesis of multiple organ failure following abdominal aortic aneurysm repair. Eur. J. Vasc. Endovasc. Surg. 22:485-495.
5. Cakir, B., and B. C. Yegen. 2004. Systemic responses to burn injury. Turkish J. Med. Sci. 34:215-226.
6. Faist, E., and C. Kim. 1998. Therapeutic immunomodulatory approaches for the control of systemic inflammatory response syndrome and the prevention of sepsis. New Horiz. 6:S97-102.
7. Gotte, M. 2003. Syndecans in inflammation. FASEB J. 17:575-591.
8. Gotte, M., A. M. Joussen, C. Klein, P. Andre, D. D. Wagner, M. T. Hinkes, B. Kirchhof, A. P. Adamis, and M. Bernfield. 2002. Role of syndecan 1 in leukocyte-endothelial interactions in the ocular vasculature. Investig. Ophthalmol. Vis. Sci. 43:1135-1141.
9. Haynes, A., III, K. P. Rumbaugh, P. W. Park, A. N. Hamood, and J. A. Griswold. 2005. Protamine sulfate reduces the susceptibility of thermally injured mice to Pseudomonas aeruginosa infection. J. Surg. Res. 123:109-117.
10. Holder, I. A. 1983. Experimental studies of the pathogenesis of infections due to Pseudomonas aeruginosa: effect of treatment with protease inhibitors. Rev. Infect. Dis. 5(Suppl. 5):S914-S921.
11. Holder, I. A. 1993. Pseudomonas aeruginosa virulence-associated factors and their role in burn wound infections, p. 235-245. In R. B. Fick (ed.), Pseudomonas aeruginosa: the oppurtunist. CRC Press, Boca Raton, Fla.
12. Kainulainen, V., H. Wang, C. Schick, and M. Bernfield. 1998. Syndecans, heparan sulfate proteoglycans, maintain the proteolytic balance of acute wound fluids. J. Biol. Chem. 273:11563-11569.
13. Kato, M., H. Wang, V. Kainulainen, M. L. Fitzgerald, S. Ledbetter, D. M. Ornitz, and M. Bernfield. 1998. Physiological degradation converts the soluble syndecan-1 ectodomain from an inhibitor to a potent activator of FGF-2. Nat. Med. 4:691-697.
14. Lentsch, A. B., and P. A. Ward. 2000. Regulation of inflammatory vascular damage. J. Pathol. 190:343-348.
15. Li, Q., P. W. Park, C. L. Wilson, and W. C. Parks. 2002. Matrilysin shedding of syndecan-1 regulates chemokine mobilization and transepithelial efflux of neutrophils in acute lung injury. Cell 111:635-646.
16. McWilliam, A. S., D. Nelson, J. A. Thomas, and P. G. Holt. 1994. Rapid dendritic cell recruitment is a hallmark of the acute inflammatory response at mucosal surfaces. J. Exp. Med. 179:1331-1336.
17. Musher, D. M. 1989. Cutaneous manifestations of bacterial sepsis. Hosp. Pract. 24:71-92.
18. Oliver, J. W., and T. E. Debowski. 1998. Pseudomonas aeruginosa septicemia and associated cutaneous lesions. ASCP, Chicago, Ill.
19. Opal, S. M., and C. E. Huber. 2002. Bench-to-bedside review: Toll-like receptors and their role in septic shock. Crit. Care 6:125-136.
20. Osmon, S., S. Ward, V. J. Fraser, and M. H. Kollef. 2004. Hospital mortality for patients with bacteremia due to Staphylococcus aureus or Pseudomonas aeruginosa. Chest 125:607-616.
21. Park, P. W., T. J. Foster, E. Nishi, S. J. Duncan, M. Klagsbrun, and Y. Chen. 2004. Activation of syndecan-1 ectodomain shedding by Staphylococcus aureus alpha-toxin and beta-toxin. J. Biol. Chem. 279:251-258.
22. Park, P. W., G. B. Pier, M. T. Hinkes, and M. Bernfield. 2001. Exploitation of syndecan-1 shedding by Pseudomonas aeruginosa enhances virulence. Nature 411:98-102.
23. Park, P. W., G. B. Pier, M. J. Preston, O. Goldberger, M. L. Fitzgerald, and M. Bernfield. 2000. Syndecan-1 shedding is enhanced by LasA, a secreted virulence factor of Pseudomonas aeruginosa. J. Biol. Chem. 275:3057-3064.
24. Plotkowski, M. C., A. O. Costa, V. Morandi, H. S. Barbosa, H. B. Nader, S. de Bentzmann, and E. Puchelle. 2001. Role of heparan sulphate proteoglycans as potential receptors for non-piliated Pseudomonas aeruginosa adherence to non-polarised airway epithelial cells. J. Med. Microbiol. 50:183-190.
25. Pollack, M. 2000. Pseudomonas aeruginosa, p. 2310-2327. In J. E. B. G. L. Mandell and R. Dolin (ed.), Mandell, Douglas, and Bennett's principles and practice of infectious diseases. Churchill Livingstone, Philadelphia, Pa.
26. Pruitt, B. A., Jr., A. T. McManus, S. H. Kim, and C. W. Goodwin. 1998. Burn wound infections: current status. World J. Surg. 22:135-145.
27. Reich, H. L., D. Williams Fadeyi, N. S. Naik, P. J. Honig, and A. C. Yan. 2004. Nonpseudomonal ecthyma gangrenosum. J. Am. Acad. Dermatol. 50:S114-S117.
28. Rostand, K. S., and J. D. Esko. 1997. Microbial adherence to and invasion through proteoglycans. Infect. Immun. 65:1-8.
29. Rumbaugh, K. P., J. A. Colmer, J. A. Griswold, and A. N. Hamood. 2001. The effects of infection of thermal injury by Pseudomonas aeruginosa PAO1 on the murine cytokine response. Cytokine 16:160-168.
30. Rumbaugh, K. P., J. A. Griswold, and A. N. Hamood. 1999. Contribution of the regulatory gene lasR to the pathogenesis of Pseudomonas aeruginosa infection of burned mice. J. Burn Care Rehab. 20:42-49.
31. Rumbaugh, K. P., J. A. Griswold, and A. N. Hamood. 2000. The role of quorum sensing in the in vivo virulence of Pseudomonas aeruginosa. Microbes Infect. 2:1721-1731.
32. Rumbaugh, K. P., J. A. Griswold, B. H. Iglewski, and A. N. Hamood. 1999. Contribution of quorum sensing to the virulence of Pseudomonas aeruginosa in burn wound infections. Infect. Immun. 67:5854-5862.
33. Rumbaugh, K. P., A. N. Hamood, and J. A. Griswold. 2004. Cytokine induction by the P. aeruginosa quorum sensing system during thermal injury. J. Surg. Res. 116:137-144.
34. Soave, R., H. W. Murray, and M. M. Litrenta. 1978. Bacterial invasion of pulmonary vessels. Pseudomonas bacteremia mimicking pulmonary thromboembolism with infarction. Am. J. Med. 65:864-867.
35. Stieritz, D. D., and I. A. Holder. 1975. Experimental studies of the pathogenesis of infections due to Pseudomonas aeruginosa: description of a burned mouse model. J. Infect. Dis. 131:688-691.
36. Stockley, R. A. 2004. Chronic obstructive pulmonary disease, neutrophils and bacteria: from science to integrated care pathways. Clin. Med. 4:567-572.
37. Subramanian, S. V., M. L. Fitzgerald, and M. Bernfield. 1997. Regulated shedding of syndecan-1 and -4 ectodomains by thrombin and growth factor receptor activation. J. Biol. Chem. 272:14713-14720.
38. Tacconelli, E., M. Tumbarello, S. Bertagnolio, R. Citton, T. Spanu, G. Fadda, and R. Cauda. 2002. Multidrug-resistant Pseudomonas aeruginosa bloodstream infections: analysis of trends in prevalence and epidemiology. Emerg. Infect. Dis. 8:220-221.
39. Tredget, E. E., H. A. Shankowsky, R. Rennie, R. E. Burrell, and S. Logsetty. 2004. Pseudomonas infections in the thermally injured patient. Burns 30:3-26.
40. Ward, P. A., and A. B. Lentsch. 1999. The acute inflammatory response and its regulation. Arch. Surg. 134:666-669.(Allan Haynes III, Frank R)