Cell-cell interactions and bronchoconstrictor eicosanoid reduction with inhaled carbon monoxide and resolvin D1

I/R induced second-organ reflow injury.

Male FVB mice (6–8 wk; Charles River; LabDiet with essential fatty acids from supplier) were used in accordance with the Harvard Medical Area Standing Committee on Animals (no. 02570 to C. Serhan and no. 04435 to A. Choi). Bilateral hind-limb ischemia was initiated using tourniquets consisting of a rubber band placed on each hind limb. Mice were subjected to hind-limb ischemia for 60 min, after which the tourniquets were removed to initiate reperfusion. CO inhalation was carried out by Laura Fredenburgh, M.D. and Emeka Ifedigbo under an NIH program project (P01-HL108801, A. Choi). Mice were kept in the CO chamber (125–250 ppm) for 60 min before the induction of hind-limb ischemia. RvD1 (250–500 ng with 100 μl of 1% ethanol containing physiological saline; Cayman Chemical) was intravenously administered to the tail vein 5 min before the start of the reperfusion period. Following the reperfusion period (120 min), the mice were euthanized, and their lungs were harvested and stored at −80 C° or in 10% (vol:vol) buffered formalin to be processed for histology assessment by the Histology Core of Boston Children's Hospital. PMN infiltrations into lungs were quantified by lung myeloperoxidase (MPO) as before (29). Briefly, the frozen lungs were homogenized and centrifuged, and the tissue levels of MPO and cytokines were determined with ELISA (R&D Systems and multiplex ELISA). To investigate PMN-platelet aggregates in murine whole blood after I/R, heparinized blood was collected 120 min postreperfusion and incubated with rat anti-mouse CD41 (BD Biosciences, clone MWReg30) and rat anti-mouse Ly-6G (BD Biosciences, clone 1A8) for 30 min at 4°C, followed by red blood cell lysis with 1× RBC lysis buffer (BioLegend) and fixation with 1% (vol:vol) formalin-PBS before flow cytometry with a BD Cant II (BD Biosciences). Data were analyzed using FlowJo (Tree Star).

LM metabololipidomics.

LM metabololipidomics was performed as before (8). Deuterium (d)-labeled internal standards d₈-5S-hydoxy-eicosatetraenoic acid (HETE), d₄-LTB₄, d₅-lipoxin (LX) A₄, d₄-PGE₂, d₅-RvD2, and d₅-LTC₄ (500 pg each) in ice-cold methanol were added to facilitate quantification of sample recovery. All samples for LM metabololipidomics were extracted using C18 solid-phase extraction (SPE) columns and subjected to liquid chromatography-tandem mass spectrometry (LC-MS-MS). The LC-MS-MS system, QTrap 6500 (AB Sciex), was equipped with a Shimadzu Prominence LC-20AD HPLC and a Shimadzu SIL-20AC autoinjector (Shimadzu). An Agilent Eclipse Plus C18 column (100 mm × 4.6 mm × 1.8 μm) was used with a gradient of methanol:water:acetic acid of 55:45:0.01 (vol:vol:vol) to 98:2:0.01 at a 0.4 ml/min flow rate. To monitor and quantify the levels of targeted LM, a multiple-reaction monitoring (MRM) method was devised with signature ion fragments for each molecule. Identification was conducted using published criteria (as summarized in Ref. 8) using LC retention times, specific fragmentation patterns, and at least six diagnostic fragmentation ions. Quantification was carried out based on the peak area of the MRM transition, and the linear calibration curves were obtained with an authentic standard for each compound.

Immunofluorescent detection of HIF-1α.

Granulocyte-macrophage colony-stimulating factor-differentiated human macrophages were adhered onto four-well chamber slides overnight (0.1 × 10⁶ cells/well) and incubated with or without 500 ppm CO for 4 h. Cells were then fixed with 1% buffered formaldehyde and permeabilized with 0.1% Triton X-100 for 15 min. After being blocked with 5% FBS containing PBS for 15 min, hypoxia-inducible factor 1α (HIF-1α) was detected by incubation with anti HIF-1α antibody (NOVUS Biologicals, clone H1α67, 1:100 dilution for 1 h at 4°C) followed by incubation with Alexa Fluor 488 goat anti-mouse IgG antibody (Molecular Probes, 1:1,000 dilution for 30 min at 4°C). Nucleus was stained with DAPI (Sigma, Fluoroshield with DAPI). HIF-1α expression and translocation were analyzed using a fluorescent microscope system (Olympus).

Human whole blood incubation.

Human whole blood was drawn from healthy volunteers, who had not taken any medications for 2 wk before donation, by venipuncture in a heparinized syringe (Partners Human Research Committee Protocol No. 1999-P-001297). Human whole blood was incubated with CO (125–250 ppm) at 37°C for 60 min, followed by platelet-activating factor (PAF) stimulation (100 nM, at 37°C for 30 min). Carboxyhemoglobin (CO-Hb) levels were measured by spectrophotometry after reduction with sodium hydrosulfite as before (1). Cell viability was assessed by Trypan blue dye exclusion test after red blood cell lysis with 1× RBC lysis buffer (BioLegend) and by plasma lactate dehydrogenase (LDH) levels (LDH Activity Assay Kit, Sigma) following CO incubation. For evaluation of different agonists, human whole blood was incubated in ambient air or CO (250 ppm) for 60 min (37°C), followed by stimulation at 37°C for 30 min with PAF (100 nM) and U46619 (1 μM) or for 10 min with adenosine diphosphate (ADP) (10 μM) and thrombin (1 U/ml) plus N-Formyl-Met-Leu-Phe (fMLP) (1 μM). To investigate PMN-platelet aggregates and adhesion molecule expressions on PMN and platelet, direct immunofluorescence labeling was performed (at 4°C for 30 min) using anti-human CD41a (Affymetrix eBioscience, clone HIP8), CD18 (BioLegend, clone TS1/18), CD11b (BioLegend, clone CBRM1/5), and CD62P (BD Biosciences, clone AK-4) in combination with the corresponding isotype controls. For LM metabololipidomics, the incubations were stopped with four volumes of ice-cold methanol with deuterium-labeled internal standards and held at −80°C before SPE.

Urinary LTE₄ amount in S. pneumoniae-infected baboon and CO treatment.

Baboon sepsis was initiated as before (19). Adult male colony-bred baboons (Papio cynocephalus) were housed in the Duke University Vivarium (Durham, NC) and handled in accordance with American Association for Accreditation of Laboratory Animal Care guidelines. The experimental protocol was approved by the Duke University Institutional Animal Care and Use Committee (A271-11-10). The baboons were sedated, intubated, and mechanically ventilated. The animals were inoculated with S. pneumoniae (Serotype 19A-7; ATCC, 10⁹ colony-forming units per lung), recovered, and extubated. At 48 h under sepsis, urine samples were collected pre-CO exposure. The animals were again sedated, intubated, and treated with inhaled CO (200 ppm) for 60 min under mechanical ventilation. Urine samples were obtained post-CO exposure at 55 h. Urinary LTE₄ amount was determined with LC-MS-MS-based metabololipidomics after manual SPE. In brief, columns (SEP-PAK C18, 500 mg, Waters) were equilibrated with 12 ml of methanol and 12 ml of ddH₂O. Before extraction, 500 pg of d₅-LTC₄ was added to 2 ml of urine as an internal standard. After centrifugation at 1,200 g for 10 min, supernatant was acidified (pH = ∼3.5) and immediately loaded onto an SPE column. After being loaded, columns were washed with 12 ml of neutral ddH₂O, followed by 10 ml of hexane and 8 ml of methyl formate. Samples were eluted with 8 ml methanol and taken to dryness and resuspended in methanol:water (50:50, vol:vol) for LC-MS-MS-automated injections. Urinary creatinine was measured with creatinine colorimetric assay kit (Cayman Chemical).

Statistical analysis.

Results are expressed as means ± SE. Statistical analysis was performed using the t-test for two-group comparisons and one-way ANOVA test for multiple-group comparisons with post hoc analysis using Dunnett's test (GraphPad Prism). P < 0.05 was considered to be statistically significant.

CO and RvD1 protect lungs from PMN-mediated I/R injury.

We assessed whether inhaled low-dose CO and intravenous RvD1 could each protect from PMN-mediated lung injury following I/R in an established murine model of operating room surgical insults in humans (12, 23). Histology indicated that I/R initiated second-organ lung PMN infiltration (Fig. 1A, top). Low-dose inhaled CO (250 ppm; see Fig. 1) or RvD1 (500 ng/mouse i.v.) each markedly reduced PMN lung infiltration (Fig. 1A, bottom). Combined CO (250 ppm) and RvD1 (500 ng) further decreased lung PMN infiltration compared with each treatment alone. Microscopic examination of the histology demonstrated that mice treated with combined CO and RvD1 showed essentially normal bronchiolar structure. I/R-initiated PMN lung infiltration was quantified by MPO levels (Fig. 1). Inhaled CO dose dependently decreased lung leukocyte influx, giving ∼10% reduction at 125 ppm and ∼55% reduction at 250 ppm (Fig. 1B). For comparison, RvD1 also dampened PMN infiltration, giving ∼70% reduction (Fig. 1C). These results indicated that CO and RvD1 each protected lung from PMN-mediated tissue injury during reflow by reducing PMN infiltration.

Next, we tested whether CO and RvD1 could additively protect lung tissues from I/R injury. To this end, mice were treated with 125 ppm inhaled CO, 250 ng RvD1, or a combination of both (Fig. 1D). Subthreshold doses of inhaled CO (125 ppm) or RvD1 (250 ng/mouse i.v.) alone did not significantly reduce PMN tissue infiltration. When administered together, they significantly lowered PMN influx (∼20% reduction), giving additive tissue protection.

CO and RvD1 reduce I/R-initiated LTs and Tx.

LM such as LTB₄ and TxB₂ play a crucial role in I/R-initiated lung injury (12). Hence, we sought evidence to determine whether CO or RvD1 each regulate LM in I/R injury. Mice were subjected to 60 min of ischemia followed by 120 min of reperfusion, lungs were collected, and LM metabololipidomics performed. Results in Fig. 2A show MRM chromatographs of arachidonic acid (AA)-derived LTs and TxB₂ identified in mice lungs. Each LM was profiled using MRM and identified by direct comparison with synthetic and authentic standards using matching criteria, including LC retention time, characteristic fragmentation patterns, and diagnostic ions (Fig. 2B). We quantified each LM identified in lung tissue (Fig. 2C, Table 1) and found that inhaled low-dose CO (250 ppm) and RvD1 (500 ng/mouse) as well as both together gave significant reduction in proinflammatory LTs including LTB₄ (∼80% reduction). We also found reduction in TxB₂ (∼65% reduction) and cysteinyl LTs with LTC₄ and LTE₄ (∼85% reductions). These results indicated that CO and RvD1 reduced proinflammatory LTs and Tx in lungs following I/R. Each of these AA-derived bioactive mediators plays crucial roles in pathogenesis of PMN-mediated lung injury (Fig. 2D): LTB₄ is a potent PMN chemoattractant, cysteinyl LTs stimulate tissue edema as well as airway smooth muscle contraction, and Tx contributes to platelet activation (25). We monitored a panel of cytokines in lung tissue following hind-limb I/R (Fig. 3) and found that low-dose inhaled CO (250 ppm), RvD1 (500 ng i.v.), and both led to significant reductions of proinflammatory cytokines including interferon (IFN)-γ (∼40%), interleukin (IL)-1β (∼30%), IL-12p40 (∼60%), IL-6 (∼65%), and monocyte chemoattractant protein 1 (∼50%). These results demonstrated that low-dose inhaled CO and RvD1 each reduced proinflammatory cytokine levels in lungs following I/R.

Inhaled low-dose CO maintains RvD1 in I/R-injured lungs.

SPM are subject to further local enzymatic metabolism; for example, RvD1 is converted to 8-oxo- and 17-oxo-RvD1 by 15-hydroxy PG dehydrogenase (15-PGDH), an eicosanoid oxidoreductase, which gives dramatically reduced bioactivity (27). CO attenuates 15-PGDH-mediated RvD1 conversion (7). Here, we examined whether inhaled CO has an impact on RvD1 amount in I/R-injured lungs. RvD1 in lungs was profiled using LM metabolomics and identified (Fig. 4 and Table 1). We quantified RvD1 in lung tissue and found that combined CO and RvD1 gave significantly higher RvD1 amounts compared with RvD1 administered alone, suggesting that CO enhances the levels of SPM in I/R lungs by preventing their local inactivation.

CO and RvD1 reduce PMN-platelet aggregates: mice whole blood after I/R.

I/R injury is known to induce systemic PMN-platelet aggregates in whole blood (12). Hence, we questioned whether CO and RvD1 could each regulate PMN-platelet aggregates during I/R injury. PMN-platelet aggregates in murine whole blood were monitored by expression levels of mouse platelet marker CD41 in the PMN population (Fig. 5, A and B), giving nearly ∼100% increases by I/R compared with sham mice. Low-dose inhaled CO (250 ppm) and RvD1 each significantly reduced PMN-platelet aggregates. Because HIF-1α plays an important role in attenuating acute lung injury (10), we examined the role of CO in regulating HIF-1α in human macrophages. Compared with normoxia, CO 500 ppm (4 h) increased HIF-1α expression and translocation into the nucleus monitored by immunofluorescent staining (Fig. 6). Thus, along with protection of deleterious impact of PMN-platelet aggregation, CO also activated HIF-1α that contributes to lung protection.

Human translation: CO reduces PMN-platelet aggregates in whole blood.

Next we questioned whether CO could regulate PMN-platelet interactions in human whole blood (15). Fresh human whole blood was incubated in ambient air or CO (125–250 ppm, 60 min, 37°C), followed by addition of PAF (100 nM, 30 min, 37°C). PAF activates both PMN and platelets and has important roles in I/R-stimulated leukocyte adherence (3). To assess potential toxicity of low-dose CO exposure, we monitored CO-Hb, a well-appreciated marker (1), and cell viability with Trypan blue dye exclusion and with plasma LDH after the CO exposure. CO at 250 ppm, the effective inhaled dose, did not alter either CO-Hb or cell viability (Table 2), indicating that, in these incubations, CO was not cytotoxic.

We next monitored expression of adhesion molecules on PMN and platelets, as well as PMN-platelet aggregates, using flow cytometry (Fig. 7, A–C). CO significantly reduced expression of CD11b and CD18 in PMN, with ∼30% and ∼40% reductions at 125 ppm and 250 ppm, compared with ambient air (Fig. 7A). We also assessed platelet CD62P, which were also significantly decreased by CO, giving ∼50% reduction with 125 ppm and ∼60% reduction by 250 ppm, compared with ambient air (Fig. 7B). PMN-platelet aggregates were monitored by expression of CD41a (Fig. 7C) that was reduced ∼40% and ∼50% with CO at 125 ppm and 250 ppm. In separate experiments, RvD1 also dose dependently decreased PAF (100 nM)-induced CD41a in CD18⁺ PMN populations by 20–25% at 10–100 nM, indicating reduced PMN-platelet interactions (Fig. 7D). In comparison, combination of CO (250 ppm) and RvD1 (10 nM) showed a trend toward further reduction of CD41a compared with each treatment alone but did not prove to be statistically significant (Fig. 7E).

In human whole blood, CO also impacts PMN-platelet aggregates stimulated by specific agonists (Table 3). We selected PAF and thrombin plus fMLP as agonists of both PMN and platelets. The Tx mimetic U46619 and ADP were used as platelet-selective agonists (4, 14). Both PAF (100 nM) and U46619 (1 μM) significantly induced PMN-platelet aggregates (∼220% and ∼110% increases, respectively). ADP (10 μM) and thrombin (1 U/ml) plus fMLP (1 μM) showed ∼80% and ∼75% increases in PMN-platelet aggregates. CO (250 ppm for 60 min) significantly reduced both PAF- and U46619-stimulated PMN-platelet aggregates (∼42% and ∼38% inhibition). In contrast, CO (250–500 ppm for 60 min) did not reduce PMN-platelet aggregates stimulated by either ADP or thrombin plus fMLP (Table 3). We also tested whether CO exposure after stimulation by PAF could reverse PMN-platelet aggregates in human whole blood. To this end, human whole blood was initially stimulated with PAF (100 nM) for 15 min, followed by incubations in ambient air or 250 ppm of CO for 60 min (Table 4). PAF significantly induced PMN-platelet aggregates (∼207% increase) in ambient air incubation. Of interest, in this setting, CO exposure for 60 min after PAF activation of PMN-platelet aggregates in human whole blood showed significant reduction (∼47% inhibition) on these aggregates.

CO regulates cysteinyl LTs and Tx in human whole blood.

Having found that CO reduced PMN-platelet aggregates, we next questioned whether CO could regulate LM production in human whole blood. Human whole blood was incubated at 37°C in ambient air or CO (250 ppm for 60 min), followed by addition of PAF (100 nM for 30 min). LM in whole blood were profiled using LM metabololipidomics. Figure 8A reports MRM chromatographs of AA-derived TxB₂ and cysteinyl LTs identified in human whole blood. Each LM was profiled and identified by direct comparison with synthetic and authentic standards using matching criteria (8) (Fig. 8A). We quantified each LM and found that 250-ppm CO exposure gave 70–80% reduction in PAF-stimulated LTC₄, LTD₄, and TxB₂ (∼90% reduction) in human whole blood (Fig. 8B).

In baboon lung infections, low-dose inhaled CO reduces urinary LTE₄ levels.

Systemic inflammation including sepsis produces potent LM, such as cysteinyl LTs, which provoke airway smooth muscle contraction, edema, and further inflammation (21, 25). To assess the extent to which CO regulates cysteinyl LT levels in primates during lung infections, we employed a nonhuman primate (baboon) model of pneumonia (19). Infection was initiated by direct inoculation of S. pneumoniae into both lungs. Diagnosis of baboon sepsis was validated based on hyperthermia, tachycardia, tachypnea, leukocytosis, and positive bacterial culture from peripheral blood (Table 5). These animals met the established criteria for pneumonia (change in white blood cells, signs/symptoms of respiratory infection, and positive bacterial cultures). Urine samples were collected, and the animals were treated with low-dose inhaled CO (200 ppm for 60 min). After inhaled CO, urine samples were collected, and urinary LTE₄ was profiled (Fig. 9A) and identified (Fig. 9B). Low-dose inhaled CO gave ∼60% reduction in urinary LTE₄.