Partially-desulfated heparin improves survival in Pseudomonas pneumonia by enhancing bacterial clearance and ameliorating lung injury

Abstract

Nosocomial pneumonia (NP, or hospital-acquired pneumonia) is associated with infections originating from hospital-borne pathogens. Persistent microbial presence and acute lung injury are common features of these infections, contributing to the high mortality rates and excessive financial burden for these patients. Pseudomonas aeruginosa (PA), a gram-negative opportunistic pathogen, is one of the prominent pathogens associated with NP. PA pneumonia is characterized by excessive secretion of inflammatory cytokines, neutrophil infiltration, and subsequent lung damage. The persistent presence of PA along with overwhelming inflammatory response is suggestive of impairment in innate immunity. High mobility group box 1 (HMGB1), a recently discovered potent pro-inflammatory cytokine, plays an important role in PA lung infections by compromising innate immunity via impairing phagocyte function through toll-like receptors (TLR) TLR2 and TLR4. ODSH (2-O, 3-O-desulfated heparin), a heparin derivative with significant anti-inflammatory properties but minimal anti-coagulatory effects, has been shown to reduce neutrophilic lung injury in the absence of active microbial infections. This study examined the effects of ODSH on PA pneumonia. This study demonstrates that ODSH not only reduced PA-induced lung injury, but also significantly increased bacterial clearance. The ameliorated lung injury, together with the increased bacterial clearance, resulted in marked improvement in the survival of these animals. The resulting attenuation in lung injury and improvement in bacterial clearance were associated with decreased levels of airway HMGB1. Furthermore, binding of HMGB1 to its receptors TLR2 and TLR4 was blunted in the presence of ODSH. These data suggest that ODSH provides a potential novel approach in the adjunctive treatment of PA pneumonia.

• Heparin
• innate immunity
• ODSH
• Pseudomonas pneumonia

Introduction

Nosocomial pneumonia (NP or hospital-acquired pneumonia) is currently one of the major healthcare challenges (American Thoracic Society and Infectious Diseases Society of America, 2005; Flanders et al., 2006). NP is defined as pneumonia occurring at least 48 h post-hospital admission (American Thoracic Society and Infectious Diseases Society of America, 2005). The prevalence of NP is substantial in mechanically ventilated patients, where it is known as ventilator-associated pneumonia (VAP) (American Thoracic Society and Infectious Diseases Society of America, 2005; Ferrara, 2006). The involvement of multi-drug resistant bacteria in patients with NP makes its management complex and difficult, resulting in high rates of morbidity and mortality (Torres & Rello, 2010). Additional hospital stays of 7–9 days are often required for patients developing NP, resulting in additional expenses of $40,000 per patient (American Thoracic Society and Infectious Diseases Society of America, 2005; Ferrara, 2006). Furthermore, the International Nosocomial Infection Control Consortium (INCC) reported that the prevalence of NP in the developing world is even higher compared to the US (15.8 versus 3.3 per 1000 ventilator days) (Rosenthal et al., 2011), posing NP as a global challenge.

Pseudomonas aeruginosa (PA) is one of the most common gram-negative pathogens in NP and associated with nearly 22% of NP cases worldwide (Jones, 2010). PA is an opportunistic bacteria, mostly affecting immunocompromised subjects (Lyczak et al., 2000). PA lung infection is a major cause of mortality and morbidity in cystic fibrosis patients, who often suffer from lifelong PA infections of the airways (Govan & Deretic, 1996). Although antibiotic therapy is routinely administered to patients infected with PA, resistance often renders these treatments ineffective (El Solh & Alhajhusain, 2009). Alternative approaches, which aim at improving host immunity against invading pathogens, might be beneficial in the treatment of PA infections.

Inflammation, characterized by the recruitment of leukocytes, is critical for host defense against invading pathogens. However, dysregulated inflammation can result in tissue injury (Smith, 1994). Excessive inflammation in response to infections can also hamper host defense, further aggravating the existing lung injury by uncleared pathogens. Thus, reducing excessive inflammation may help to ameliorate inflammatory lung injury and improve host defense against PA infection. High Mobility Group Box 1 protein (HMGB1) is a highly conserved nuclear protein involved in nucleosome stabilization, DNA repair, and regulation of transcription (Lotze & Tracey, 2005). However, HMGB1 exhibits characteristics of potent pro-inflammatory cytokines once released into the extracellular milieu (Abraham et al., 2000; Lotze & Tracey, 2005). It has been shown that extracellular HMGB1 mediates neutrophilic lung injury (Abraham et al., 2000). Our recent studies showed that extracellular HMGB1 in airways can directly compromise host ability to clear PA infections, exacerbating PA-induced lung injury in animal models of cystic fibrosis and ventilator-associated pneumonia (Entezari et al., 2012; Patel et al., 2013).

ODSH, 2-O, 3-O desulfated heparin is a selectively desulfated heparin derivative which retains the potent anti-inflammatory properties of heparin, but is devoid of potent anti-coagulant effects of the parent compound (Fryer et al., 1997). ODSH has been shown to inhibit the secretion of pro-inflammatory cytokines including tumor necrosis factor (TNF)-α, interleukin (IL)-6, and IL-1β upon stimulation or injury, but has minimal effects on coagulation (Rao et al., 2010; Lakshmi et al., 2011). In addition, inflammatory lung injury mediated by either human leukocyte elastase or HMGB1 can be inhibited by treatment with ODSH (Fryer et al., 1997; Rao et al., 2010).

In this study, we sought to test the effects of ODSH on bacterial clearance in a mouse model of PA pneumonia. We hypothesized that ODSH will improve the clinical outcomes in PA pneumonia by inhibiting acute lung injury and enhancing bacterial clearance, in part, by attenuating the accumulation of airway HMGB1 induced by the PA infection.

Materials and methods

Special reagents

ODSH (2-O, 3-O-desulfated heparin) was prepared by Scientific Protein Laboratories (Wanaukee, WI) as described previously (Fryer et al., 1997) using unfractionated USP sodium heparin as starting material. Isotonic stock solutions were used to treat mice. For binding studies, human TLR2, TLR4, HMGB1, anti-HMGB1 monoclonal antibody and horseradish peroxidase-donkey anti-mouse antibody were purchased from R&D Systems (Minneapolis, MN). 3,3′,5,5;-tetramethylbenzidine (TIB) chromogen solution from Zymed (San Francisco, CA) and all other chemicals that are not specifically mentioned were from Sigma (St. Louis, MO).

Animal studies

C57BL/6J mice (male, 8–12-weeks-old; The Jackson Labs, Bar Harbor, ME) were used in this study in accordance with the Institutional Animal Care and Use Committees of St. John’s University. The mice were housed in a specific pathogen-free environment maintained at 22 °C in ≈50% relative humidity and with a 12-h light/dark cycle. All mice had ad libitum access to standard rodent chow and water.

For the studies involving assessment of the impact of ODSH on host resistance to Pseudomonas aeruginosa (PA), mice were anesthetized with isoflurane or sodium pentobarbital (60 mg/kg) as described previously (Patel et al., 2013) and then intranasally inoculated with PA (as indicated in figure legends) that had been prepared as outlined below. ODSH was then administered by subcutaneous injection at various timepoints post-infection. Mice were then either monitored for morbidity/mortality (see below) or euthanized at fixed timepoints post-infection by intraperitoneal injection of an overdose of sodium pentobarbital (120 mg/kg) to permit harvest of bronchoalveolar lavage (BAL) fluid (see below) and lung tissues for analyses of PA burden. For the survival and time-course studies, mice were inoculated intratracheally as described below.

Survival study

To ascertain the potential for ODSH to impact on host mortality/morbidity from PA infection, a separate set of mice were anesthetized with intraperitoneal sodium pentobarbital. Thereafter, with a 1–2-cm incision, the trachea of each mouse was dissected and then inoculated with 0.5 × 10⁸ CFU of PAO1. PA was intratracheally inoculated to induce more severe injury and significant lethality in mice in order to determine the effect of ODSH on host survival. Mice were then administrated 75 mg ODSH/kg or saline (control) subcutaneously every 12 h, starting at the time of inoculation. All mice were observed for up to 48 h post-inoculation for indices of morbidity or for mortality.

Bronchoalveolar lavage fluids (BAL)

BAL samples were prepared as described previously (Entezari et al. 2012; Patel et al. 2013). Briefly, after mice were euthanized, the trachea was exposed and dissected, and a 20-Gauge (×1¼ in) catheter was inserted. The lungs were then gently lavaged twice with two 1-ml volumes of sterile non-pyrogenic phosphate-buffered saline solution (PBS, pH 7.4) (Mediatech, Inc., Hendon, VA). The BAL samples from each given mouse were combined and centrifuged (1600 rpm, 10 min, 4 °C). The resultant supernatants were collected and stored in a −80 °C freezer for later analyses of levels of HMGB1 and total protein content via Western blot analysis and a bicinchoninic acid assay (BCA), respectively. After the lavage, the lungs were harvested and homogenized in 1 ml PBS.

Bacterial culture

Pseudomonas aeruginosa strain PAO1 was a kind gift of Dr Alice Prince. The bacteria were grown overnight in 2% LB Broth (Acros Organics, Fair Lawn, NJ). Bacteria were collected by centrifuging at 3000 rpm at 25 °C, washed once with PBS, and the OD of the suspension measured at 590 nm in a Spectronic 20+ spectrophotometer (Thermo Scientific, Madison, WI). The concentration of the viable bacteria in the solution was extrapolated from a standard curve prepared earlier relating OD to numbers of bacteria; the value in the sample itself was further confirmed by plating serial dilutions of the inoculum.

Quantitative bacteriology

Viable bacterial counts in the airways and lungs were determined as described previously (Entezari et al. 2012; Patel et al. 2013). Briefly, viable bacterial counts in the airways and lungs were determined using by plating serial dilutions of the BAL and lung homogenates, respectively, onto Pseudomonas Isolation Agar (PIA, Difco; Sparks, MD) and culturing at 37 °C. Each dilution was plated in duplicates. After ≈16 h, the numbers of colonies on each plate were enumerated and total colony forming units per milliliter homogenate or BAL calculated. After the lavage steps, the lungs were fixed with 4% formaldehyde solution and stored in formaldehyde. The sections were subsequently stained with hematoxylin and eosin.

Measurement of HMGB1 and total protein content

BAL total protein concentrations were determined using a colorimetric BCA assay. BAL HMGB1 levels were assessed via immunoblot analysis using anti-HMGB1 antibodies as noted previously in Patel et al. (2013). In brief, samples were separated on SDS-PAGE. Proteins were electrotransferred to a PVDF membrane and then blocked in 5% non-fat dry milk in Tris-buffered saline (pH 7.6). The membrane was then incubated with anti-HMGB1 antibody (1:1000 dilution, Sigma) and then with anti-rabbit horseradish peroxidase–coupled secondary antibodies (1:5000 dilution, GE Healthcare, Princeton, NJ). After washing, antibody binding was detected using enhanced chemiluminescence plus Western blotting detection reagents (Thermo Scientific, West Palm Beach, FL). The blots were then scanned with a UVP Biospectrum 600 Imaging System (Vision Works LS, Upland, CA) and band intensities quantified using Vision Works image acquisition and analysis software (Version 6.8).

Cell counts and cell differentials

Total cells in each BAL sample were collected by centrifugation. After the sample’s supernatant was removed, the cell pellet was re-suspended into 300 µl of PBS and total cell counts determined using standard hemocytometer procedures. For differential cell counts, cytospin preparations were made and cells were stained with HEMA-3 stain (Fisher Scientific, Kalamazoo, MI). Neutrophils were identified by their size and polymorphonuclear nucleus.

HMGB1 binding to TLR2 and TLR4

For studies of the effect of ODSH on HMGB1 binding to TLRs, polyvinyl 96-well plates were coated with each TLR of interest (0.5 µg/well). Separately, a constant amount of HMGB1 (0.1 µg HMGB1 protein in 100 µl PBS containing 0.01% Trtion X-100 [PBST] and 0.1% BSA) was incubated with an equal volume of serially diluted ODSH (0.000085–8.5 µM in PBST-0.1% BSA) overnight at 4 °C. The following day, 50 µl of HMGB1-ODSH mix was transferred to each TLR-coated well and incubated at 37 °C for 2 h. Wells were then washed 4-times with PBST. To detect bound HMGB1, 50 µl of monoclonal HMGB1-Ab (25 ng/well) was added to each well, the mixture was incubated for 1 h at room temperature and the wells were washed again 4-times with PBST. Horseradish peroxidase-conjugated secondary antibody (50 µl/well, 1:2000 dilution) was then added to each well and the plates incubated for 1 h at room temperature. After the wells were washed once with PBST, a colorimetric reaction was initiated by addition of 50 µl TIB solution and terminated after 15 min by addition of 50 µl 1 N HCl. Absorbance at 450 nm was then measured using a Spectromax M2 microplate reader (Molecular Devices, Sunnyvale, CA) and plotted against ODSH concentration. The data was analysed using SoftMax Pro (Molecular Devices) software by fitting the data in a 4-parameter logistic non-linear regression equation to obtain the IC₅₀ values.

Statistical analysis

Results are presented as means (± SEM) from at least three independent experiments. The data were analyzed for statistical significance according to paired and unpaired t tests, analysis of variance (ANOVA), or a Kaplan-Meier analysis, using Microsoft Excel (Microsoft Corp., Seattle, WA). A p value ≤0.05 was considered statistically significant.

Results

ODSH reduces pulmonary bacterial burden in mice with PA pneumonia

Pulmonary infections with PA are difficult to eradicate. To test the effect of ODSH on the clearance of this problematic pathogen, mice were treated with ODSH or saline at 0 and 12 h post-intranasal inoculation with 5 × 10⁸ CFU/mouse of PAO1, a non-mucoid strain of PA. The number of viable bacteria in lung homogenates and BAL were assessed to determine bacterial burden in the lungs and airways. Mice treated with ODSH at 25 and 75 mg/kg showed a significant decrease in bacterial burden in lungs (5.14 [± 0.26] versus 4.16 [± 0.18] and 4.13 [± 0.22] log CFU/ml lung homogenate, respectively [p ≤ 0.01]) (Figure 1a). Similarly, the bacterial burden in BAL was reduced upon ODSH treatment (4.68 [± 0.29] versus 4.06 [± 0.19] and 3.96 [± 0.18] log CFU/ml, at 25 and 75 mg/kg, respectively [p ≤ 0.05]) (Figure 1b), indicating the potential application of ODSH in improving host defense against PA lung infection.

Figure 1. ODSH reduces pulmonary bacterial burden in mice with PA pneumonia. Male C57BL/6 J mice were intranasally inoculated with 5 × 10⁸ CFU PA and treated with either different doses of ODSH or saline (control) at 0 and 12 h. BAL and lungs were harvested 24 h post-inoculation. Number of viable bacteria in airways and lungs were quantified from serial dilutions of either BAL or lung homogenates on Pseudomonas isolation agar plates. Bacterial burden is presented as log CFU/ml of (a) lung homogenate and (b) BAL. Data represent mean (± SEM) of five independent experiments, n ≥ 10 for each group. *Value significantly different compared to the control group (*p ≤ 0.05 and **p ≤ 0.01).

ODSH ameliorates PA-induced lung injury

The inflammatory response accompanying PA infection can lead to severe lung injury (Sadikot et al., 2005), characterized by increased total protein contents and cell infiltration in airways (Entezari et al., 2012; Suntres et al., 2002). Figure 2a shows that mice inoculated with PA had acute lung injury indicated by a time-dependent increase of total protein content in airways and marked lung damage at 24 h post-inoculation (Figure 2c). ODSH has been shown to attenuate inflammatory lung injury induced by intratracheal instillation of human leukocyte elastase (Fryer et al., 1997).

Figure 2. ODSH ameliorates PA-induced acute lung injury. In a mouse model of infection-induced acute lung injury, mice were inoculated intratracheally with 0.5 × 10⁸ CFU PA and euthanized at different timepoints post-infection (n = 3–5 mice/group). (a) Total protein contents in BAL samples of non-ODSH-treated mice that were isolated at different timepoints post-inoculation are presented as percentage of total protein content in samples at 24 h post-infection. *p ≤ 0.05 and ***p ≤ 0.001 compared to value at 0 h. (b, c) Mice were treated with different doses of ODSH or saline (control) at both 0 and 12 h post-intra-nasal inoculation with 5 × 10⁸ CFU PA and then euthanized at 24 h to harvest BAL and lung tissues in the same experiments noted in Figure 1. (b) Total protein content was determined in BAL samples from PA-infected mice treated with different doses of ODSH or saline and presented as a percentage of control group value. *p ≤ 0.05 vs control group. (c) Right lungs of mice with PA infection and treated with different levels of ODSH in same experiments as mentioned in Figure 1 were fixed in formalin. Representative images of lung sections (stained with hematoxylin and eosin) from mice treated with different doses of ODSH or saline. Magnification = 10×. Data and images represent at least four independent experiments for each dose; n > 12 for each group.

To determine whether ODSH can ameliorate PA-induced acute lung injury along with the improved bacterial clearance, male C57BL/6J mice were inoculated with PA, and then treated with different doses of ODSH at 0 and 12 h post-inoculation. Compared to mice treated with saline, levels of total protein in BAL from ODSH-treated mice were significantly reduced from 100 [±25.52]% of control levels to 33.36 [±10.92]% in mice that received 25 mg ODSH/kg and to 31.22 [±10.02]% in those given 75 mg ODSH/kg (p < 0.05; Figure 2b), respectively. No significant improvement was observed in mice treated with 8.3 mg ODSH/kg (i.e. 93.34 [±29.05]%, Figure 2b). Histological analysis showed that ODSH at 25 and 75 mg/kg markedly reduced PA-induced cell accumulation in the lung interstitium and alveolar spaces compared to the control animals (Figure 2c). In accordance with histologic images, the number of total cells as well as number of neutrophils was found to be decreased (Table 1) at 24 h post-intra-nasal inoculation in ODSH-treated mice. These data indicate that ODSH administration at 25 and 75 mg/kg helped to maintain alveolar integrity by reducing PA infection-induced lung injury and cell accumulation. In addition, mice treated with ODSH at these levels had improved clinical symptoms compared to the control mice, which exhibited symptoms of severe illness with lethargy and huddling into corners of cages. Together, these results suggest that ODSH may improve clinical outcomes in PA pneumonia as a result of enhanced bacterial clearance and reduced lung injury.

Table 1. ODSH treatment reduces cell accumulation in the airways.

	ODSH (mg/kg)
	0	8.3	25	75
Total cell count (× 10^6)	20.67 ± 7.35	17.04 ± 5.29 (p = 0.66)	5.50 ± 2.20 (p = 0.06)	5.14 ± 3.48 (p = 0.08)
Neutorphil count (× 10^6)	5.63 ± 1.74	3.15 ± 1.02 (p = 0.38)	2.29 ± 0.95 (p = 0.11)	1.77 ± 0.74 (p = 0.06)

Total numbers of cells were counted in BAL samples obtained from mice intra-nasally inoculated with 5 × 10⁸ CFU, as described in Figure 1. Mice treated with ODSH have decreased numbers of total cells and neutrophils compared to mice receiving saline.

Data shown are mean (± SEM) of five independent experiments; n ≥ 10 for each group.

p Values are compared to control mice receiving saline.

ODSH reduces PA infection-induced elevation of airway HMGB1 and decreases binding of HMGB1 to TLR2 and TLR4

We recently demonstrated that high levels of airway HMGB1 could directly compromise host defense against bacterial infection (Entezari et al., 2012; Patel et al., 2013). To explore potential molecular mechanisms underlying these observations, this study determined whether the effect of ODSH on PA pneumonia resulted from modulating levels of airway HMGB1. Using Western blot analysis, levels of extracellular HMGB1 in BAL obtained every 4 h up to 24 h post-inoculation were measured. Figure 3a shows a time-dependent increase in airway levels of HMGB1 in PA-infected mice. Mice treated with 25 and 75 mg/kg of ODSH had markedly reduced airway HMGB1 levels compared to mice treated with saline, while ODSH at 8.3 mg/kg had no observable effect (Figure 3b).

Figure 3. ODSH reduces PA-induced levels of airway HMGB1 and inhibits the binding of HMGB1 to TLR2 and TLR4. (a) BAL samples, isolated from mice that were inoculated intratracheally with 0.5 × 10⁸ CFU PA and euthanized at different timepoints post-inoculation, as noted in Figure 2a, were probed for levels of HMGB1 by western blot analysis. Representative blot shows levels of airway HMGB1 in PA-infected mice. (b) To test the effect of ODSH on the levels of airway HMGB1, BAL samples were collected at 24 h post-intra-nasal inoculation with 5 × 10⁸ CFU PA (in the same experiment as Figure 1) and probed by Western blot analysis. The image is representative of five independent experiments. (c, d) Competitive binding assay to study effect of ODSH on HMGB1 binding to (c) TLR2 and (d) TLR4. Graphs are representative of typical experiment; data are mean (± SD) of quadruplicate wells for each concentration. Data were analyzed using Soft-Max Pro software by fitting data in the 4-parameter logistic non-linear regression equation to obtain IC₅₀ values.

TLR2 and TLR4, important receptors for HMGB1, play critical roles in mediating HMGB1-impaired host defense. To determine whether ODSH can interfere with binding of HMGB1 to TLR2 and TLR4, competitive ELISA was performed. Figures 3c and d illustrate that increasing ODSH concentrations result in decreased HMGB1 binding to both TLR2 and TLR4, as indicated by reduced optical density. The concentration of ODSH that could reduce binding of HMGB1 to TLR2 and TLR4 to 50% of maximum level (IC₅₀) was 0.069 and 0.09 µM, respectively. Together, these results suggest that ODSH improved bacterial clearance by both reducing levels of airway HMGB1 and inhibiting binding of HMGB1 to TLR2 and TLR4.

ODSH improves survival of mice with PA pneumonia

To determine effects of ODSH on PA infection-induced mortality, survival studies were performed in a mouse model of lethal PA pneumonia. Mice were each intratracheally inoculated with 0.5 × 10⁸ CFU PAO1 and observed for survival up to 48 h post-inoculation after receiving either ODSH (75 mg/kg) or saline every 12 h starting at the time of PA inoculation. No mice received antibiotics. Figure 4 shows that only 25% of mice treated with saline survived at 48 h, while 50% of those treated with ODSH survived to this time. In addition to improved survival, mice in the ODSH group appeared more active compared to mice given saline that showed severe signs of illness including lethargy, huddling, and unwillingness to move upon stimulation. These data affirm the beneficial effects of ODSH in PA pneumonia.

Figure 4. ODSH improves survival of mice with PA pneumonia. Male C57BL/6J mice were inoculated intratracheally with 0.5 × 10⁸ CFU PA and treated with either 75 mg ODSH/kg or saline (control) every 12 h, starting at the time of inoculation. Mice were observed for survival up to 48 h post-inoculation. Mice that survived for up to 48 h were euthanized at 48 h time-point. Data were analyzed using Kaplan-Meier analysis. Mice treated with ODSH had improved survival probability compared to the control group (0.50 versus 0.25) at 48 h (p = 0.069 by Wilcoxon test); n = 16 for each group.

Discussion

Persistent bacterial presence with excessive inflammation is a characteristic of patients with PA pneumonia. In this study, we show that treatment with ODSH not only ameliorates inflammatory lung injury but also improves bacterial clearance in a mouse model of PA pneumonia. The increased bacterial clearance and ameliorated lung injury in ODSH-treated mice resulted in an improved survival in these animals. We further demonstrated that not only did ODSH treatment decrease levels of airway HMGB1 in mice with PA lung infection, it also inhibited binding of HMGB1 to its receptors, TLR2 and TLR4, suggesting potential mechanisms for ODSH-improved innate immunity against PA infection.

Inflammatory responses including the recruitment of phagocytes are essential in clearing invading pathogens. PA infection can induce excessive inflammatory responses with increased capillary permeability and neutrophilic infiltration (Suntres et al., 2002) (Figures 2a and c). However, severe infection and pronounced inflammation often co-exist in patients with PA infection. For example, while suffering from lifelong PA infection, patients with cystic fibrosis exhibit overwhelming neutrophilic inflammation in their lungs (Heijerman, 2005). Similarly, either mechanical stretch or hyperoxia during mechanical ventilation can induce significant pulmonary inflammatory responses, and patients undergoing mechanical ventilation have greater susceptibility to infections (Berra et al., 2010; Halbertsma et al., 2005; Patel et al., 2013; Rosenthal et al., 2011). Persistence of infection accompanied with excessive inflammatory responses suggests a defect in host immunity instead of lack in inflammatory response in these patients. The impairment in host defense might contribute to the low success rate in the treatment of PA infection in patients with CF and VAP.

With their large genome containing multiple genes for antibiotic resistance, not only are PA inherently resistant to many of the classic anti-bacterial agents, but also can readily acquire resistance to new classes of antibiotics known to be effective against bacterial infections (Livermore, 2002; Stover et al., 2000). The persistent presence of large numbers of PA due to impaired host defense and ineffectiveness of routine anti-microbial therapy may lead to the use of more potent and/or higher doses of antibiotics for prolonged duration, which may further increase the risk of antibiotic resistance and drug toxicity, establishing a vicious cycle in the management of these patients (Etminan et al., 2012). In this study we show that animals treated with ODSH have significantly reduced bacterial burden in their lungs (Figures 1a and b), suggesting that ODSH can improve host defense against PA infection. By improving host defense against PA infection, ODSH has the potential to decrease the doses and duration of anti-microbial therapy. Subsequently, this can reduce the chances of PA to acquire resistance to antibiotics as well as minimize toxic effects of anti-microbial therapy, leading to decreased hospital stay and therapeutic cost for patients with PA infection.

In the study here, it is shown that mice treated with ODSH had significantly reduced lung injury. Vascular leak, marked by elevation of total protein content in airways of mice infected with PA, was significantly reduced upon treatment with ODSH (Figure 2b), indicating the effectiveness of ODSH in maintaining lung integrity. ODSH-improved lung injury may have resulted from multiple factors, including increased bacterial clearance in the lungs (Figures 1a and b) and inhibited accumulation of leukocytes, especially neutrophils (Table 1) in the lungs of mice with PA infection. Heparin and its derivatives have been shown to possess anti-inflammatory properties in various cell and animal models (Carr, 1979; Young, 2008). These agents inhibit NF-κB activation and secretion of inflammatory cytokines such as IL-8, TNFα, and IL-6 from endotoxin-stimulated monocytes (Hochart et al., 2006). Similarly, heparin and ODSH have been reported to protect against neutrophilic elastase-mediated lung injury (Fryer et al., 1997). In a recent clinical trial, inhalation of heparin has been shown to significantly reduce days on ventilator in critically ill patients with acute lung injury (Dixon et al., 2010). Whether this outcome is due to potential immuno-modulatory effects of heparin, and whether ODSH has a similar effect on the hospital stay of these patients, are matters for further investigation.

Innate immunity responds to invading pathogens with a well-coordinated recruitment of professional phagocytes to eradicate invading pathogens. Inflammatory responses with excessive phagocytes are observed in patients with CF and VAP (Halbertsma et al., 2005; Heijerman, 2005). However, despite the pronounced inflammatory responses, severe pulmonary infections are hallmarks of these diseases, suggesting that the host defense against invading pathogens is impaired. We have recently demonstrated potential molecular mechanisms underlying the impaired innate immunity against PA infections in animal models of CF and VAP. We previously demonstrated that host defense against PA infection could be compromised via HMGB1-mediated pathways (Entezari et al. 2012; Patel et al., 2013). Levels of HMGB1 in both lung lavage fluids and sputum are elevated in patients with CF and VAP (Entezari et al., 2012; Rowe et al., 2008; van Zoelen et al., 2007). Inhibition of HMGB1 using neutralizing antibodies can improve both bacterial clearance and lung injury in the mouse models of these diseases (Entezari et al., 2012; Patel et al., 2013). In this study, we show that PA infection induced an increase in the levels of airway HMGB1 in a time-dependent manner, with a substantial accumulation at 24 h (Figure 3a). This marked elevation in the levels of airway HMGB1 coincided with substantial mortality (Figures 3a and 4). ODSH treatment resulted in a reduction in HMGB1 levels in PA infected mouse airways (Figure 3b). This reduction can be critical in breaking the vicious cycle where high levels of airway HMGB1 lead to the recruitment of inflammatory cells, impaired phagocytic ability, and compromised bacterial clearance, all of which can further increase the levels of airway HMGB1 from the excessive inflammatory cells (Abraham et al., 2000; Entezari et al., 2012; Liu et al., 2008). The beneficial effects of ODSH in PA lung infections were further illustrated by the improved survival in the mouse model of lethal PA infection (Figure 4). The improved survival in ODSH treated mice is likely a result of enhanced innate immunity and ameliorated lung injury, two critical factors in the pathogenesis of PA pneumonia.

Surface receptors, such as members of the TLR family, play important roles in innate immunity in detecting and responding to invading pathogens (Chaudhuri & Sabroe, 2008). We have previously demonstrated that the interactions of HMGB1 with TLR2 and TLR4 play critical roles in mediating HMGB1-compromised phagocytic function of alveolar macrophages (Entezari et al., 2012). Data presented in the current study show that interactions of HMGB1 with TLR2 and TLR4 can be disturbed dose-dependently by ODSH (Figures 3c and d). This observation is consistent with the study showing that heparin can alter the conformation of HMGB1, leading to reduced binding of HMGB1 to its receptor, RAGE (receptor for advanced glycation end-products) (Ling et al., 2011). The reduced binding of HMGB1 to TLR2 and TLR4 by ODSH can further attenuate the detrimental effects of HMGB1 on pulmonary host defense. Taken together, these data indicate that ODSH affects the HMGB1 pathway at its release and its binding to receptors. This may be important in immune cells like macrophages that are essential components of host defense and are susceptible to HMGB1-mediated impairment of phagocytosis (Entezari et al., 2012). Inhibition of the HMGB1 release and its exclusion from binding to either TLR might explain the improved immunity against PA-infection upon ODSH treatment in this mouse model. HMGB1 has also been recently shown to interact with pathogen associated molecules such as lipopoly-saccharide (LPS) or endogenous cytokines like IL-1β, and potentiate activity of LPS or IL-1β mediated by their respective receptors (Hreggvidsdottir et al., 2009; Sha et al., 2008). Therefore, it is also possible that ODSH can bind to the complex of HMGB1 with other endogenous/exogenous molecules and inhibit binding of HMGB1/LPS or HMGB1/IL-1β complex to respective receptors, leading to ameliorated inflammatory response and immuno-modulation. Although the pharmacokinetics of subcutaneous ODSH (i.e. entry into pulmonary compartment) was not explored here, ODSH has been shown to reach systemic circulation in animals after subcutaneous administration (Rao; unpublished data). It is very likely that ODSH traverses from lung capillaries to alveolar compartments, and is even more likely in a host with increased microvascular permeability during a PA infection.

In summary, results presented in this study show a novel application of ODSH in modulating innate immunity against PA lung infections. Inhibition of HMGB1 release and blocking its downstream pathways are potential mechanisms that contribute, at least in part, to the beneficial effects of ODSH in host defense against bacterial infections. Therefore, ODSH may provide a novel strategy in treating PA pneumonia.

Declaration of interest

Thomas Kennedy is the inventor of the ODSH and ParinGenix holds the patent for ODSH. The authors alone are responsible for the content of this manuscript.

This work was supported by grants (LLM) from ParinGenix Inc., National Heart and Blood Institute (HL093708) and St. John’s University.

Acknowledgements

The authors would like to thank Rohan Parikh and Dr Louis Trombetta for insightful discussions and generous support; Michelle Zur, Nathan Siegelaub, and Nicholas Chen for carefully reading the manuscript.