CORM-3-derived CO modulates polymorphonuclear leukocyte migration across the vascular endothelium by reducing levels of cell surface-bound elastase

Shinjiro Mizuguchi, Jancy Stephen, Relka Bihari, Nevena Markovic, Shigefumi Suehiro, Alfredo Capretta, Richard F. Potter, Gediminas Cepinskas

Abstract

Recently, it has been shown that carbon monoxide (CO)-releasing molecule (CORM)-released CO can suppress inflammation. In this study, we assessed the effects and potential mechanisms of a ruthenium-based water-soluble CO carrier [tricarbonylchloroglycinate-ruthenium(II) (CORM-3)] in the modulation of polymorphonuclear leukocyte (PMN) inflammatory responses in an experimental model of sepsis. Sepsis in mice was induced by cecal ligation and puncture. CORM-3 (3 mg/kg iv) was administered 15 min after the induction of cecal ligation and puncture. PMN accumulation in the lung (myeloperoxidase assay), bronchoalveolar lavage (BAL) fluid, and lung vascular permeability (protein content in BAL fluid) were assessed 6 h later. In in vitro experiments, human PMNs were primed with LPS (10 ng/ml) and subsequently stimulated with formyl-methionyl-leucylphenylalanine (fMLP; 100 nM). PMN production of ROS (L-012/dihydrorhodamine-123 oxidation), degranulation (release of elastase), and PMN rolling, adhesion, and migration to/across human umbilical vein endothelial cells (HUVECs) were assessed in the presence or absence of CORM-3 (1–100 μM). The obtained results indicated that systemically administered CORM-3 attenuates PMN accumulation and vascular permeability in the septic lung. Surprisingly, in in vitro experiments, treatment of PMNs with CORM-3 further augmented LPS/fMLP-induced ROS production and the release of elastase. The latter effects, however, were accompanied by an inability of PMNs to mobilize elastase to the cell surface (plasma membrane), an event required for efficient PMN transendothelial migration. The CORM-3-induced decrease in cell surface levels of elastase was followed by decreased PMN rolling/adhesion to HUVECs and complete prevention of PMN migration across HUVECs. In contrast, treatment of HUVECs with CORM-3 had no effect on PMN transendothelial migration. Taken together, these findings indicate that, in sepsis, CORM3-released CO, while further amplifying ROS production and degranulation of PMNs, concurrently reduces the levels of cell surface-bound elastase, which contributes to suppressed PMN transendothelial migration.

  • sepsis
  • systemic inflammation
  • oxidative stress
  • leukocyte adhesion/migration
  • proteolytic enzymes
  • carbon monoxide
  • carbon monoxide-releasing molecules
  • tricarbonylchloroglycinate-ruthenium(II)

carbon monoxide (CO) is one of the natural end products of heme oxygenase (HO) activity in mammalian tissues, which degrades heme to produce biliverdin/bilirubin, ferrous iron, and CO (35, 41). The affinity between CO and hemoglobin is ∼220 times stronger than that of hemoglobin for oxygen. Therefore, the higher levels of circulating CO, usually achieved through CO inhalation, result in the formation of toxic levels of carboxyhemoglobin (29). On the other hand, it has been reported that low concentrations of CO, administered either by inhalation (34) or through the upregulation of inducible HO-1 activity (33), offer cytoprotective effects and therefore could be involved in the regulation of physiological responses, such as inflammation.

Recently, a new approach [transitional metal carbonyls, CO-releasing molecules (CORMs)] has been used to systemically deliver CO in a more controlled manner without altering carboxyhemoglobin levels (29). In regard to the latter, it has been demonstrated that CORM-derived CO exhibits vasoactive, anti-inflammatory features and offers protection against cellular and tissue damage (29) in numerous models of injury, including ischemia-reperfusion (16), pulmonary hypertension (52), transplantation (44), hemorrhagic shock (53), and sepsis/endotoxemia (5, 10). The latter effects of CORM-derived CO closely resemble the effects of low-dose inhaled CO, although without altering carboxyhemoglobin levels, as mentioned previously. However, the mechanisms of CORM-derived CO with respect to the modulation of the inflammatory response(s) have not yet been thoroughly investigated.

In general, one of the hallmarks of systemic inflammation [as a consequence of mechanical trauma, transplantation, or bacterial infection (i.e., sepsis)] is an increase in polymorphonuclear leukocyte (PMN) accumulation in the affected systemic organs, such as the lung, heart, kidney, and liver. The mechanism(s) of PMN recruitment to the affected organs is a complex multistep process and involves the activation of both the vascular endothelium and leukocytes, with a subsequent upregulation of the proadhesive phenotype, which, in turn, results in the initiation of adhesive interactions (rolling, firm adhesion, and migration) between PMNs and vascular endothelial cells (ECs) (8, 23). An inflammation-relevant transcription factor (NF-κB) plays a key role in the induction of the proinflammatory/proadhesive phenotype in numerous cell types, including vascular ECs and leukocytes (24, 25). While the adhesion of leukocytes (e.g., PMNs) to vascular ECs has been extensively studied in the past and is quite well understood, the process of PMN migration across the vascular endothelium remains poorly investigated and largely controversial.

Our recent study (5) indicated that CO derived from DMSO-soluble tricarbonylchloro-glycinate-ruthenium(II) (CORM-2) can interfere with the upregulation of the inflammatory response in the vascular endothelium by suppressing ROS generation, NF-κB activation, and ICAM-1 expression. In contrast, the potential effects of CORM-derived CO on the modulation of the inflammatory response(s) in leukocytes (i.e., PMNs) have not yet been investigated in more detail. Some evidence, however, has indicated that CORM-derived CO can inhibit the expression of proinflammatory cytokines and increase the expression of anti-inflammatory cytokines in LPS-stimulated macrophages (33). It is important to note that while some studies (26, 45) have demonstrated attenuating effects of CO with respect to ROS production in stimulated PMNs, other studies (46, 51) have also confirmed increased ROS production in PMNs in the presence of CORM-released CO. With regard to the latter, it also has been shown that HO-1-derived CO increases the phagocytic rate (2.3-fold) of macrophages (10).

Therefore, in this study, we used a novel ruthenium-based water-soluble CORM [tricarbonylchloroglycinate-ruthenium(II) (CORM-3)] to assess the effects and potential mechanisms of CORM-3-released CO in the modulation of inflammatory responses in PMNs and ECs and the subsequent functional consequences (adhesion/migration) using cecal ligation and puncture (CLP)-induced peritonitis as a clinically relevant model of systemic inflammation (polymicrobial sepsis) in vivo and sepsis-relevant proinflammatory mediators [LPS and formyl- methionyl-leucylphenylalanine (fMLP)] in vitro.

MATERIALS AND METHODS

Reagents

CORM-3 (molecular weigh: 512.18) was synthesized by Dr. A. Capretta from a commercially available compound (CORM-2 dimer, Sigma Aldrich, St. Louis, MO) in accordance with the previously published methods of Clark et al. (11). CORM-3 was solubilized in double-distilled water (10 mM stock) and stored at −20°C. Inactive CORM-3 (iCORM-3) was prepared by leaving CORM-3 in Dulbecco's PBS (DPBS; pH 7.4) buffer overnight at room temperature to liberate all CO from the molecule, as previously described (47). Medium-199, FCS, penicillin, and streptomycin were purchased from Wisent (St-Bruno, QC, Canada). LPS (Escherichia coli serotype 055:B5), SOD, and fMLP were obtained from Sigma-Aldrich. Human elastase and Ala-Ala-Pro-Val-p-nitroanilide (AAPV-pNA; elastase substrate) were purchased from Calbiochem. L-012 was purchased from Wako Pure Chemical (Osaka, Japan), and dihydrorhodamine (DHR)-123 was obtained from Molecular Probes (Eugene, OR).

Animals

C57BL/6 mice (6 wk old, Jackson Laboratories) were used in the experiments. All animal experiments were performed in accordance with University of Western Ontario Animal Care and Use Committee-approved protocols.

CLP

CLP-induced peritonitis (a clinically relevant model of polymicrobial sepsis) was induced as previously described by us (5). Briefly, mice were anesthetized with 2% isoflurane in oxygen via a face mask. A midline incision was made through the abdominal wall to expose the cecum. Cecum was ligated with a 3-0 silk tie 1 cm from the tip and punctured once with an 18-gauge needle. The cecum was lightly squeezed to express a small amount of stool from the puncture site to assure a full thickness perforation. Care was taken to preserve the continuity of flow between the small and large bowels. The cecum was returned to the abdominal cavity, and the incision was closed. Sham-operated (sham) mice were given anesthesia, and a midline laparotomy was performed. CORM-3 or iCORM-3 (3.0 mg/kg iv, penile vein) was administered in 0.1 ml of normal saline immediately after the induction of CLP. Six hours after the induction of CLP, mice were euthanized by cervical dislocation. After perfusion of the circulatory system with ice-cold saline, the lungs were collected and processed for the assessment of myeloperoxidase (MPO) activity.

MPO Activity Assay

MPO activity, as an index of neutrophil influx, was assessed as previously described by us (7). In brief, lung tissue was homogenized in 0.5 ml of 50 mM potassium phosphate buffer (pH 7.4) and centrifuged at 10,000 g at 4°C for 30 min. The remaining pellet was resuspended in 0.5 ml of 50 mM potassium buffer (pH 6.0) containing 0.5% hexadecyltrimethylammonium bromide, sonicated, and centrifuged at 12,000 g at 4°C for 10 min. MPO activity in the supernatants was assessed spectrophotometrically at a wavelength of 460 nm (Bio-Rad 680 microplate reader) with o-dianisidine as a substrate. MPO activity is expressed as units per gram of tissue.

Bronchoalveolar Lavage Fluid

Mice were challenged with sham operation or CLP as described above. Six hours later, mice were euthanized, and the lungs and trachea were surgically exposed. The trachea was cannulated using an 18-gauge angiocatheter (Becton Dickinson). Subsequently, the brachioalveolar (BAL) fluid was collected by instilling and withdrawing a lavage solution (1.0 ml of 0.9% saline) three times via the tracheal cannula using a 1-ml syringe (Becton Dickinson). The recovery of the lavage solution ranged from 70% to 90% of the instilled volume. There were no differences in the volume recoveries between the experimental groups. The lavage samples were centrifuged at 150 g for 10 min at 4°C. Cells were counted and stained with Wright-Giemsa stain for differential analysis. The protein concentration in cell-free supernatants was measured using Bradford reagent (Bio-Rad).

Cells

Human umbilical vein ECs (HUVECs) were harvested from human umbilical cord veins by collagenase treatment (Worthington Biochem, Freehold, NJ) as previously described by us (8). For the assay involving the “parallel flow chamber,” HUVECs were grown to confluence on fibronectin-coated glass coverslips. HUVECs at passages 1–3 were used for the experiments.

Human neutrophilic leukocytes (PMNs) were isolated from the venous blood of healthy adults by 1% dextran sedimentation and gradient separation on Histopaque-1077 (Sigma). This procedure yields a PMN population that is 95–98% viable (trypan blue exclusion) and 98% pure (acetic acid-crystal violet staining) (7).

ROS Production

Oxidative stress in PMNs was assessed by measuring the intracellular oxidation of DHR-123 (Molecular Probes), an oxidant-sensitive fluorochrome, as previously described (5). To this end, PMNs (1 × 106 cells) in suspension were loaded with DHR-123 (5 μM) for 45 min and stimulated in the presence or absence of CORM-3. After stimulation, cells were washed with PBS, lysed in 0.5% CHAPS buffer, and analyzed spectrofluorometrically (RF-1501 spectrofluorometer, Shimadzu) at excitation/emission wavelengths of 495/523 nm. ROS production is expressed as DHR-123 fluorescence emission per 1 × 106 cells.

In parallel, levels of extracellular ROS produced by PMNs were assessed using L-012, an O2-sensitive chemiluminescence probe, as previously described (19). Briefly, isolated PMNs (1 × 106 cells) were resuspended in 190 μl DPBS containing 5.5 mM glucose and 100 μM L-012 and placed in Lumitrac 96-well plates (Greiner Bio-One). Subsequently, PMNs were stimulated in the absence or presence of CORM-3, and the chemiluminescence intensity was continuously recorded for 30 min at 37°C in a Victor-3 Multilabel counter (Perkin-Elmer). ROS production is expressed as L-012 oxidation per 1 × 106 PMNs.

PMN Elastase

PMN elastase activity in cell supernatants (secreted elastase), the whole cell pellet, and PMN plasma membranes (cell-associated elastase) was assessed as previously described by us (7) with some modifications using AAPV-pNA (Calbiochem) as the elastase substrate. Briefly, PMNs (1 × 106 cells) were primed with LPS for 45 min in 200 μl DPBS containing 5.5 mM glucose in the absence or presence of CORM-3 and subsequently challenged with fMLP for an additional 15 min. Next, PMNs were centrifuged at 400 g for 5 min at room temperature. For the detection of secreted elastase, 50 μl PMN supernatant was mixed with 150 μl Tris·HCl (pH 7.4) containing 1 mM AAPV-pNA, and elastase activity was detected spectrophotometrically (415 nm, Bio-Rad microplate reader) 60 min later. In parallel, elastase activity in the remaining PMNs (whole cell pellet) was assessed after a wash of PMNs in PBS and sonication of the cells in 0.2 ml PBS for 10 s. Elastase activity is expressed as optical densitiy units per 1 × 106 PMNs.

For the detection of plasma membrane-associated elastase, PMNs (2 × 107 cells/ml) were treated as described above and sonicated for 1 min at 30% power output in a lysis buffer containing 100 mM KCl, 5 mM MgCl2, and 25 mM Tris·HCl (pH 9.6). The obtained sonicates were centrifuged at 1,000 g for 10 min to sediment nuclei and undisrupted cells, and supernatants were collected. Plasma membranes were isolated on a self-generating Percoll gradient. To this end, 2 ml of supernatants from PMN sonicates were loaded on top of the mixture of 5.5 ml Percoll with 1.1 ml H2O buffered with 2.4 ml of 400 mM KCl, 20 mM MgCl2, and 400 mM Tris·HCl (pH 9.6). Subsequently, samples were centrifuged at 160,000 g for 15 min, and fractions of 1.5 ml were harvested from the top of the Percoll gradient. Fractions were diluted with 3 ml of 100 mM KCl, 5 mM MgCl2, and 50 mM Tris·HCl (pH 7.4) buffer and centrifuged at 200,000 g for 60 min. The obtained pellet was resuspended in 100 μl PBS and analyzed for elastase activity.

PMN Rolling/Adhesion Assay

PMN-EC adhesive interactions under conditions of “flow” were performed using a parallel flow chamber as previously described (17). In brief, HUVECs grown on the glass coverslips were mounted onto a Plexiglas parallel flow chamber and placed into an air-heated (37°C) chamber affixed to an inverted microscope (Diaphot 300, Nikon). At the beginning of the experiment, HUVECs were perfused with medium 199 for 10 min at a constant rate of 1 dyn/cm2 using a syringe pump (Harvard Apparatus). Subsequently, PMNs (0.5 × 106 cells/ml) were added to the medium and perfused over the endothelial monolayers at a constant rate of 1 dyn/cm2. After 3 min of perfusion, 15 random fields (400 μm2 each) were recorded for 10 s with a digital camera (DXC-107A, Sony) connected to a videotape recorder (JVC). Fluxes of PMN rolling and adhesion (cells that remained stationary for at least 10 s) were analyzed after the experiment.

PMN Transendothelial Migration

PMN transendothelial migration was assessed as previously described by us (7, 8). Briefly, HUVECs were grown to confluence on fibronectin (25 μg/ml)-coated Falcon cell culture inserts (3-μm-diameter pores). Na51CrO4-labeled PMNs (1 × 106 cells) were added to the apical aspect of naïve or stimulated HUVECs and allowed to migrate for 60 min in response to a fMLP (10−7 M, placed into the basal compartment of the inserts) chemotactic gradient. Subsequently, the amount of 51Cr radioactivity (in counts/min) in the basal compartment was assessed using a γ-counter (Wallac, Turku, Finland), and percentage of PMN migration was calculated.

Experimental Protocols

ROS/elastase.

For the measurement of ROS production, PMNs were either placed into Lumitrac 96-well plates (Greiner Bio-One, for L-012 chemiluminescence detection) or into 1.5-ml microcentrifuge tubes (Axygen, for DHR-123 oxidation) and pretreated with CORM-3 (1, 10, or 100 μM) or iCORM-3 (100 μM) for 2 min at 37°C. Subsequently, PMNs were stimulated with fMLP (1 × 10−7 M) in the presence of CORM-3 for 1–30 min (for L-012 chemiluminescence detection) or 15 min (for DHR-123 oxidation).

For the detection of elastase release/translocation, PMNs suspended in 1.5-ml microcentrifuge tubes (Axygen) were first pretreated with CORM-3 (10 or 100 μM) or iCORM-3 (100 μM) for 2 min at 37°C. Next, LPS (10 ng/ml) was added to the cell suspension for 45 min to prime PMNs [a condition required for PMN degranulation in vitro (36)]. Subsequently, PMNs were stimulated with fMLP (1 × 10−7 M) for an additional 15 min, and elastase activity in obtained the supernatants, cell pellet, and plasma membranes was assessed as described above.

PMN-HUVEC adhesive interactions.

For the assessment of PMN rolling and adhesion under flow conditions, PMNs were primed with LPS (10 ng/ml) for 45 min and stimulated with fMLP (1 × 10−7 M) for an additional 15 min in the absence or presence of CORM-3 (100 μM) or iCORM-3 (100 μM).

For the analysis of PMN transendothelial migration (cell culture inserts), HUVECs, PMNs, or both were stimulated with LPS (10 ng/ml) for 4 h (for HUVECs) or 45 min (for PMNs) in the absence or presence of CORM-3 (100 μM) or iCORM3 (100 μM). In these experiments, PMNs were labeled with Na51CrO4 before treatment with LPS/CORM-3. fMLP (10 −7 M) was used as a chemotactic factor.

Statistical Analysis

All values are presented as means ± SE. Statistical analysis was performed using ANOVA with Bonferroni's correction for multiple comparisons. P values of <0.05 were considered to be statistically significant. Statistical analysis was performed with the StatView 5.0J software package (SAS Institute, Cary, NC).

RESULTS

CORM-3-Derived CO Attenuates PMN Infiltration in the Septic Lung

To determine whether CLP-induced PMN recruitment to the lung could be reduced/prevented by CORM-3, the activity of MPO (a marker of PMN accumulation) in lung tissue was assessed. Levels of MPO were markedly increased in the lungs of CLP-challenged mice compared with sham mice (Fig. 1). The systemic (intravenous) administration of CORM-3 significantly decreased MPO activity in the septic lung, whereas iCORM-3 failed to interfere with MPO activity in the lung of CLP-challenged mice.

Fig. 1.

Effects of carbon monoxide (CO)-releasing molecule (CORM) tricarbonylchloroglycinate-ruthenium(II) (CORM-3)-derived CO on myeloperoxidase (MPO) activity in the lung of cecal ligation and perforation (CLP)-challenged mice. Mice were challenged with CLP and treated with CORM-3 (3 mg/kg iv) immediately after the induction of CLP. Sham-operated (sham) mice were injected with vehicle (0.1 ml saline). MPO activity [an index of polymorphonuclear leukocyte (PMN) accumulation] in the lung was assessed 6 h after the induction of CLP. iCORM-3, inactive CORM-3. Results are means ± SE of 3 experiments (4 mice/group). #P < 0.05 compared with sham mice; *P < 0.05 compared with CLP mice.

In parallel, PMN accumulation in the alveolar space (BAL fluid) and lung vascular permeability were also assessed. The obtained results indicated that the induction of sepsis (CLP) in mice resulted in an increase in the total cell count in BAL fluid, with PMNs being the predominant cell type infiltrating the alveolar space (Fig. 2A). In addition, CLP-challenged mice had a significantly higher concentration of protein in BAL fluid compared with sham mice (Fig. 2B), indicating increased lung vascular permeability in septic animals. The systemic administration of CORM-3 (but not its inactive counterpart iCORM-3) effectively reduced PMN accumulation in BAL fluid and completely prevented the CLP-induced increase in vascular permeability (Fig. 2, A and B, respectively).

Fig. 2.

Effects of CORM-3-derived CO on cell counts and protein concentrations in the bronchoalveolar lavage (BAL) fluid of CLP-challenged mice. Mice were challenged with CLP and treated with CORM-3 (3 mg/kg iv) immediately after the induction of CLP. Sham mice were injected with vehicle (0.1 ml saline). Six hours after CLP induction/CORM-3 treatment, BAL fluid was collected, and total/differential cell counts (A) and protein concentrations (Bradford assay; B) were assessed. Results are means ± SE of 4 experiments (5 mice/group). #P < 0.05 compared with sham mice; *P < 0.05 compared with CLP mice.

CORM-3-Derived CO Increases ROS Production by fMLP-Stimulated PMNs

Next, we assessed the effects of CORM-3-liberated CO on the modulation of the PMN inflammatory response (production of ROS) induced by a sepsis-relevant stimulus (fMLP). To this end, we first assessed the production of extracellular ROS (L-012 oxidation) after the stimulation of PMNs with fMLP (10−7 M) in the presence or absence of CORM-3 (1–100 μM). As shown in Fig. 3A, stimulation of PMNs with fMLP in the presence of CORM-3 resulted in a dose-dependent increase in extracellular ROS production, as assessed by the oxidation of the O2-sensitive probe L-012. The most profound increase in ROS production was observed 2 min after the stimulation of PMNs. It is important to note that the latter effect could be seen as early as 1 min after the addition of fMLP to CORM-3-treated PMNs and persisted for as long as 30 min of coincubation (data not shown). In contrast, iCORM-3 failed to increase fMLP-induced ROS production by PMNs.

Fig. 3.

Effects of CORM-3 on the production of ROS in formyl-methionyl-leucylphenylalanine (fMLP)-stimulated PMNs. ROS production was assessed by measuring the extracellular oxidation of L-012 (chemiluminescence; A) and intracellular oxidation of dihydrorhodamine (DHR)-123 (B). For extracellular ROS production (A), PMNs were suspended in a buffer containing L-012 (100 μM) and stimulated with fMLP in the absence or presence of CORM-3 (1, 10, or 100 μM) or iCORM-3 (100 μM). L-012 oxidation (chemiluminescence) was assessed after 2 min after the stimulation with fMLP and is expressed as chemiluminescence per 1 × 106 PMNs. For DHR-123 oxidation (B), PMNs were loaded with DHR-123 (5 μM) for 45 min and stimulated with fMLP for 15 min in the absence or presence of CORM-3 or iCORM-3 (100 μM). Oxidation of DHR-123 is expressed as fluorescence emission per 1 × 106 PMNs. Results are means ± SE; n = 6 (in duplicate). #P < 0.05 compared with control (Ctrl; unstimulated) cells; *P < 0.05 compared with fMLP only-stimulated cells.

Similarly, the levels of intracellular ROS, as assessed by the oxidation of DHR-123 after 15 min of stimulation of PMNs with fMLP, were also significantly increased in CORM-3-treated but not iCORM-3 (100 μM)-treated PMNs (Fig. 3B).

To exclude the potential oxidation of DHR-123 or L-012 by CORM-3 or iCORM-3, the above-mentioned compounds were interacted in a test tube in the absence of PMNs. The obtained results indicated that neither CORM-3 nor iCORM-3 induced the oxidation of DHR-123/L-012 (n = 3, data not shown).

It is important to note that similar results with regard to ROS production (both intracellular and extracellular) were obtained in the experiments in which PMNs were first primed with LPS (10 ng/ml for 45 min) and subsequently challenged with fMLP (10−7M) (data not shown), indicating that CORM-3-derived CO further amplifies ROS production in LPS/fMLP-stimulated PMNs.

CORM-3-Derived CO Modulates Elastase Release and Cell Surface Binding in LPS-Primed fMLP-Stimulated PMNs

Previous studies, including our own, have indicated that PMN proteolytic (elastase) activity is an important component involved in PMN migration across the vascular endothelial barrier (7, 36). It appears that, upon activation, PMNs mobilize elastase to the cell surface, where it is used for the proteolytic degradation of EC junctional components (7).

Therefore, in this series of experiments, we assessed whether CORM-3-liberated CO could modulate PMN elastase release/cell surface binding under septic conditions. As shown in Fig. 4, priming PMNs with LPS (10 ng/ml for 45 min) and subsequently challenging them with fMLP (1 × 10−7M) for 15 min (a condition necessary for PMN degranulation in vitro) resulted in increased elastase secretion into cell supernatants (Fig. 4A) and a concomitant decrease in total elastase activity in PMNs (cell pellet; Fig. 4B). The latter was associated with increased elastase binding/mobilization to the cell surface (as assessed by elastase activity in PMN plasma membranes; Fig. 5) obtained from LPS/fMLP-stimulated PMNs.

Fig. 4.

Modulation of PMN elastase secretion by CORM-3-derived CO. PMNs (1 × 106 cells) were primed with LPS (10 ng/ml) for 45 min in the absence or presence of CORM-3 (10 and 100 μM) or iCORM-3 (100 μM) and stimulated with fMLP for an additional 15 min. Elastase activity in cell supernatants (A) and the cell pellet (B) was assessed using the elastase substrate Ala-Ala-Pro-Val-p-nitroanilide (AAPV-pNA; 1 mM). Elastase activity is expressed as optical density units (OD) per 1 × 106 PMNs. Results are means ± SE; n = 5 (in duplicate). #P < 0.05 compared with Ctrl (unstimulated) cells; *P < 0.05 compared with LPS-primed, fMLP-stimulated cells.

Fig. 5.

Modulation of elastase cell surface binding by CORM-3-derived CO. PMNs (1 × 106 cells) were primed with LPS (10 ng/ml) for 45 min in the absence or presence of CORM-3 or iCORM-3 (100 μM) and stimulated with fMLP for an additional 15 min. Elastase activity in PMN plasma membranes was assessed using the elastase substrate AAPV-pNA (1 mM). Elastase activity was expressed as OD per milligram of protein. Results are means ± SE; n = 3. #P < 0.05 compared with Ctrl (unstimulated) cells; *P < 0.05 compared with LPS-primed, fMLP-stimulated cells.

The administration of CORM-3 to LPS/fMLP-challenged PMNs further increased elastase release in a dose-dependent manner (Fig. 4A), indicating that CORM-3-liberated CO further activates PMNs under experimental conditions of sepsis. However, while the levels of secreted elastase were higher in the supernatants obtained from CORM-3-treated PMNs (Fig. 4A), concurrently the levels of plasma membrane-bound (cell surface) elastase were significantly reduced by CORM-3-derived CO (Fig. 5).

To exclude potential direct effects of CORM-3/iCORM-3 in the modulation of elastase activity, purified human elastase was coincubated with CORM-3 or iCORM-3 under identical experimental conditions. The obtained results indicated that neither CORM-3 nor iCORM-3 affected elastase activity in vitro (n = 2, data not shown).

CORM-3-Derived CO Differentially Modulates the Adhesive Properties of HUVECs and PMNs

We (5) have previously reported that exposing HUVECs to CO released from DMSO-soluble CORM-2 attenuates PMN adhesion to LPS-stimulated HUVECs.

In this study, we assessed the effects of CORM-3-released CO on the modulation of PMN-HUVEC adhesive interactions (rolling and adhesion) and migration across HUVECs using parallel flow chamber and cell culture insert approaches, respectively.

As shown in Fig. 6, stimulation of PMNs with LPS/fMLP significantly increased both PMN rolling (A) and PMN adhesion (B) to naïve HUVECs, a phenomenon that was significantly reduced by treating PMNs with CORM-3. Similar results were obtained in the experiments in which PMN rolling/adhesion to LPS-stimulated HUVECs were assessed (data not shown).

Fig. 6.

Modulation of PMN rolling and adhesion to human umbilical vein endothelial cells (HUVECs) by CORM-3-derived CO. Freshly isolated PMNs were primed with LPS (10 ng/ml) for 45 min in the absence or presence of CORM-3 (100 μM) or iCORM-3 (100 μM) and stimulated with fMLP for an additional 15 min. Subsequently, PMNs were interacted with naïve (unstimulated) HUVECs under a constant shear rate of 1 dyn/cm2. Numbers of rolling PMNs (A) and adhered PMNs (B) were determined after 3 min of perfusion of PMNs over HUVEC monolayers. Results are means ± SE of 3 independent experiments. #P < 0.05 compared with unstimulated cells; *P < 0.05 compared with LPS only-stimulated cells.

It is important to note that while CORM-3-derived CO under the conditions described above was effective in reducing overall PMN rolling and adhesion, at the same time, treatment of PMNs with CORM-3 (but not iCORM-3) resulted in the complete inhibition of fMLP-induced PMN migration across naïve (unstimulated) HUVECs (Fig. 7A). Surprisingly, treatment of LPS-stimulated HUVECs with CORM-3 had no effect on the transendothelial migration of naïve PMNs (Fig. 7B). Moreover, interactions of PMNs and HUVECs under conditions when both cells (i.e., PMNs and HUVECs) were stimulated with LPS/fMLP in the presence of CORM-3 resulted in the complete inhibition of PMN migration (Fig. 7C), a phenomenon that resembles the scenario of when only PMNs were treated with CORM-3 (Fig. 7C).

Fig. 7.

Modulation of PMN migration across HUVECs by CORM-3-derived CO. A: 51Cr-labeled PMNs were stimulated with LPS (10 ng/ml) for 45 min in the absence or presence of CORM-3 or iCORM-3 (100 μM), washed, and then added to naïve (unstimulated) HUVECs grown on permeable supports. B: confluent HUVEC monolayers grown on permeable supports were stimulated with LPS (10 ng/ml) for 4 h in the absence or presence of CORM-3 or iCORM-3 (100 μM). Subsequently, HUVECs were washed, and naïve (unstimulated) 51Cr-labeled PMNs were added to the apical aspect of HUVEC monolayers. C: 51Cr-labeled PMNs, HUVECs, or both (i.e., PMNs and HUVECs) were stimulated with LPS (10 ng/ml) for 45 min and 4 h, respectively, in the presence or absence of CORM-3 (100 μM). Under all experimental conditions, PMN migration across HUVECs in response to a fMLP (1 × 10−7 M, placed into the basal compartment of the inserts) chemotactic gradient was assessed 60 min later. Results are means ± SE; n = 6 (in duplicate). #P < 0.05 compared with unstimulated cells; *P < 0.05 compared with LPS only-stimulated cells.

DISCUSSION

Systemic inflammatory response syndrome (SIRS), as a consequence of mechanical or surgical trauma, organ transplantation, or bacterial infection per se (i.e., sepsis), is a common clinical problem leading to substantial morbidity and mortality in intensive care units worldwide. One of the key features of SIRS/sepsis is an overwhelming production of proinflammatory mediators (e.g., cytokines) in the circulation and an accumulation of PMNs (neutrophils) in affected systemic organs. While PMN recruitment to the afflicted sites is an entirely normal host response to remove pathogens or dead tissue, the overwhelming accumulation of PMNs and subsequent production of cytotoxic ROS and the release of proteolytic enzymes contribute significantly to the development of multiple organ dysfunction syndrome (MODS) (9).

Given the complexity and acute nature of SIRS and sepsis, it is perhaps not surprising that little progress has been made in improving the overall outcome. Efforts to block one or another component of the sepsis-associated inflammatory pathway(s) or to target a single cell response(s) have had little or no impact on patient survival. Of the many pathways/drugs tested, few have demonstrated efficacy (9).

Recent findings in the field have indicated that one of the potent regulators of the inflammatory response is inducible HO-1, an enzyme that catalyzes the formation of CO, biliverdin/bilirubin, and ferrous iron and exhibits anti-inflammatory properties (32) that are beneficial for the resolution of inflammation (35, 42, 50).

In regard to the above, several studies (27, 28, 32, 42) have demonstrated beneficial anti-inflammatory effects of CO in preventing microvascular perfusion deficits and cellular injury in different organs during SIRS. In addition, it has been demonstrated that CO suppresses LPS-induced proinflammatory cytokine production by macrophages and interferes with the upregulation of the proadhesive phenotype in vascular ECs (5, 33, 43). In support of the above, it has been shown that both mice and humans deficient in HO-1 expression have a phenotype of an increased inflammatory state (21).

However, most of the studies addressing the role of CO in the modulation of inflammatory responses were performed using systems that either artificially induce the overexpression of HO-1 or provide CO in a gaseous form (organ superfusion by CO gas or CO gas inhalation). Both of the approaches mentioned above have limitations with respect to the control of temporal and spatial CO distributions and the amount of CO produced at the cell/tissue level (13, 33). Moreover, the administration of exogenous CO via inhalation results in an increase in the carboxyhemoglobin concentration, thus presenting a potential threat to the host and limiting the use of the latter approach (13, 32).

Recently, transitional metal carbonyls (CORMs) have been used to deliver CO in a more controlled manner without altering carboxyhemoglobin levels. As a consequence, these molecules have received increasing attention for potential pharmaceutical applications (30, 31). With regard to the latter, CORMs have been shown to act pharmacologically in rat aortic and cardiac tissue, where the liberation of CO produced vasorelaxant, antihypertensive, and antirejection effects (14, 30, 31) and reduced myocardial ischemia-reperfusion damage (16). In addition, the anti-inflammatory activity of CORM-2, a DMSO-soluble CORM, has also been demonstrated in experimental models of sepsis in vivo (CLP-induced peritonitis) (5) and in vitro (LPS-stimulated murine macrophages) (41).

Therefore, in this study, we used a novel ruthenium-based water-soluble CORM, CORM-3, to assess the effects and potential mechanisms of CORM-3-released CO in the modulation of inflammatory responses in PMNs and ECs using a clinically relevant model of systemic inflammation (sepsis) in vivo and sepsis-relevant proinflammatory stimuli (LPS and fMLP) in vitro. The obtained results indicated that CORM-3-derived CO offers anti-inflammatory effects with respect to decreased PMN recruitment to the septic lung and sepsis-induced vascular permeability in vivo (Figs. 1 and 2). These findings are in line with our previously published results (5) addressing the anti-inflammatory effects of another DMSO-soluble CORM, CORM-2.

However, in contrast to our previous and recent observations indicating that CO released from CORM-2 (5) or CORM-3 (data not shown) decreases oxidant production in LPS-stimulated HUVECs, the results of the present study indicate that CORM-3-liberated CO in a dose-dependent manner augments fMLP-induced ROS production by neutrophilic leukocytes (PMNs).

O2 has been shown to be the key radical contributing to overall ROS production by PMNs (1). Although multiple cellular sources [e.g., mitochondrial respiration, NAD(P)H oxidase, and cytochrome P-450] may individually or in concert contribute to the overall induction of oxidative stress, it is generally accepted that the predominant source of O2 in PMNs is the phagocytic NADPH oxidase system (1, 41).

The results of the present study indicate that O2 is the prime ROS produced by fMLP-stimulated PMNs in the presence of CORM-3 (as assessed by the oxidation of the O2-specific probe L-012) (19). The predominant production of O2 was confirmed using SOD, which completely inhibited L-012 oxidation under the experimental conditions used in this study (data not shown). Which redox system [e.g., NAD(P)H oxidase or mitochondrial respiration] is involved in CORM-3-mediated O2 production remains to be determined.

Although our present results indicate opposing and, therefore, controversial effects of CORM-3-derived CO in the modulation of ROS production by ECs (e.g., HUVECs; data not shown) and PMNs, there is sufficient evidence supporting both scenarios. Previous studies (26, 45) in the field have indicated that CORM-3-liberated CO can suppress O2 overproduction in LPS-stimulated macrophages or fMLP-challenged PMNs. Meanwhile, some studies (46, 51) have indicated that CO can amplify ROS production by airway smooth muscle cells and PMNs. The diverse and often opposing effects of CO appear to be dependent on the CO concentration, temporal and spatial CO distributions, proximity of the heme proteins, and specific redox pathway(s) that CO gets involved in (13, 33, 40). One possible explanation of why in our study CORM-3-derived CO amplified O2 production by LPS/fMLP-stimulated PMNs [as opposed to the inhibitory effects of CORM-3 with respect to the above (26)] could be the difference in time points chosen to assess ROS (i.e., O2) production and also the different experimental approaches being used. While in the studies mentioned above (26, 45) O2 production was assessed after the activation of PMNs with fMLP for prolonged periods of time (up to 2 h) and in the presence of cytochalasin B [a cytoskeleton-destabilizing and degranulation- potentiating agent (26)], in the present study O2 production was assessed under acute conditions (1–30 min) and under physiological conditions (i.e., in the absence of cytochalazin B).

Despite of the fact that overwhelming production of ROS during inflammation can cause direct cytotoxicity, moderate production of ROS has also been recognized as a part of the intracellular signaling involved in the upregulation of the proinflammatory/proadhesive phenotype in both vascular ECs and PMNs, resulting in increased PMN recruitment to affected organs/tissues (6, 12).

PMN recruitment to inflamed tissues involves a series of complex, yet well-coordinated, PMN-EC adhesive interactions (i.e., PMN rolling and firm adhesion) and migration across the endothelial barrier within the microvasculature (23). The available literature indicates that there is a general consensus on the molecular determinants of PMN-EC adhesive interactions, i.e., extensive characterization of the adhesion molecules (e.g., P-selectin, E-selectin, ICAM-1, and VCAM-1 on the vascular endothelium) and their ligands (e.g., L-selectin, sialyl Lewis X, and β2-integrins on PMNs), the kinetics of their expression after an induction with various proinflammatory mediators (e.g., TNF-α or LPS), and their relative roles in rolling and the development of strong adhesive interactions (23).

However, unlike the situation with PMN-EC adhesive interactions, the mechanisms by which PMNs traverse the endothelium and enter the interstitium are not entirely clear and are rather controversial. This is in part due to the existence of very dynamic interactions between vascular ECs and PMNs during the course of paracellular PMN migration (22). The latter event is driven by complex homophilic interactions between the adhesion molecules [e.g., PECAM-1 (CD31), CD99, and junction adhesion molecules] expressed on both ECs and PMNs and heterophilic interactions between β2-integrins (e.g., lymphocyte function-associated antigen-1 and Mac1) expressed on PMNs and IgG superfamily members (e.g., ICAM-1 and ICAM-2) expressed on activated ECs (23). In addition, the proteolytic activity of PMNs (e.g., elastase, a serine proteinase that is contained at millimolar concentrations in the azurophilic granules of PMNs) has been shown to play a key role in PMN migration across the vascular EC barrier (7, 36, 39, 48, 49).

Although some previous findings with regard to the above have indicated that PMN transendothelial migration is not dependent on proteases (15, 18), the vast majority of recent studies (including our own) have indicated that proteases, particularly PMN-derived elastase, play critical roles in disrupting the integrity of interendothelial cell junctions and therefore contribute to the PMN migration across the vascular endothelium both in vitro and in vivo (4, 7, 20, 36, 48). Moreover, it has been demonstrated that in response to inflammatory mediators [e.g., TNF-α or platelet-activating factor (PAF)] or bacterial product-relevant (e.g., LPS and fMLP) stimulation, PMNs not only secrete elastase into the surrounding milieu (2, 3739) but also induce an up to 20-fold increase in elastase expression on the cell surface (i.e., upregulate the levels of plasma membrane-bound elastase), where it is mobilized to the leading front of migrating PMNs (7, 37, 39). The latter event has a profound physiological significance since membrane-bound elastase, as opposed to its secreted counterpart, appears to be catalytically active and less susceptible to inhibition with serine proteinase inhibitors present in the circulation (3, 7, 37). In addition, recent findings have indicated the existence of a cooperative interaction between α6β1-integrin and cell surface-bound neutrophil elastase in the regulation of PMN transmigration in vivo (48, 49). Therefore, membrane-bound elastase presents PMNs with a perfect tool for the degradation of intercellular junction proteins and, later on, extracellular matrix components, thus playing a central role in the process of PMN diapedesis.

The results of the present study indicate that stimulation of PMNs with a sepsis-relevant stimulus (LPS/fMLP) increases PMN rolling/adhesion to HUVECs under experimental conditions of flow, an effect that could be substantially reduced by treatment of PMNs with CORM-3. The latter findings are in line with recently published data indicating that CORM-3-derived CO suppresses the PMN proadhesive phenotype by inhibiting PAF-induced expression of CD11b (47).

However, the most profound effects of CORM-3-released CO in the modulation of PMN-HUVEC adhesive interactions were found in experiments addressing PMN transendothelial migration. In these experiments, treatment of PMNs with CORM-3 completely prevented PMN migration across HUVECs in response to a fMLP chemotactic gradient (Fig. 7A). Surprisingly, treatment of HUVECs with CORM-3 (Fig. 7B) had no effect on PMN migration, suggesting that the migration-suppressing effects of CORM-3-derived CO are primarily associated with the modulation of PMN function.

The exact mechanism(s) with respect to the above is not readily available; however, it could be attributed, at least in part, to the ability of CORM-3-derived CO to modulate the levels of cell surface-bound elastase in PMNs. As demonstrated in this study for the first time, CORM-3-derived CO, while promoting the overall secretion of elastase into the surrounding milieu in response to PMN stimulation with LPS/fMLP, at the same time completely prevents the recruitment of elastase [the key factor involved in PMN transendothelial migration (7)] to the plasma membranes (Fig. 5). The latter provides with a potentially new mechanistic insight related to the anti-inflammatory activity of CORM-derived CO.

It is worthwhile to note, however, that the latter findings do not preclude the potential role of CORM-derived CO in modulating the secretion/translocation of other PMN-derived proteolytic enzymes (e.g., cathepsin G or matrix metalloproteinases), which can individually or in concert with elastase contribute to the overall process of PMN diapedesis. The latter scenario may be vital in light of the existing redundancy with regard to the mechanisms of PMN transendothelial migration assuring appropriate migration to the sites of inflammation, a process that is critical to the survival of the host.

Taken together, the findings of the present study indicate that modulation of the inflammatory response in neutrophilic leukocytes (and vascular ECs) by CORM-3-derived CO under experimental conditions of sepsis in vitro is a complex and multifactorial phenomenon that involves the amplification of PMN activation (increased ROS production) and degranulation (increased secretion of elastase). However, the CO-dependent increase in PMN activation results in suppressed PMN adhesion to ECs and an inability of PMNs to mobilize elastase to the cell surface, resulting in the complete inhibition of PMN migration across EC monolayers. In addition, it appears that the effects of CORM-3-released CO are more potent in modulating PMN inflammatory response(s) compared with ECs, predisposing the functional outcome of EC-PMN adhesive interactions in favor of the latter. These findings could be of high importance in designing new approaches for the clinical treatment of SIRS/sepsis using a new class of CO-releasing compounds, CORMs.

GRANTS

This work was supported by Heart and Stroke Foundation of Ontario Grant HSFO-NA6171 (to G. Cepinskas), Canadian Institutes for Health Research Grant MOP-68848 (to R. F. Potter), and Lawson Health Research Institute Internal Research Fund IRF-043-06 (to G. Cepinskas).

REFERENCES

View Abstract