Acute lung injury (ALI) still poses a major challenge in critical care medicine. Neutrophils, platelets, and chemokines are all considered key components in the development of ALI. The Duffy antigen receptor for chemokines (DARC) is thought to be involved in scavenging, transendothelial transport, and presentation of neutrophil-specific chemokines. DARC is expressed on endothelial cells and erythrocytes but not on leukocytes. Here, we show that DARC is crucial for chemokine-mediated leukocyte recruitment in vivo. However, we also demonstrate that changes in chemokine and chemokine receptor homeostasis, associated with Darc gene deficiency, exert strong anti-inflammatory effects. Neutrophils from Darc gene-deficient (Darc−/−) mice display a more prolonged downregulation of CXCR2 during severe inflammation than neutrophils from wild-type mice. In a CXCR2-dependent model of acid-induced ALI, Darc gene deficiency prevents ALI. Darc−/− mice demonstrate fully preserved oxygenation, only a small increase in vascular permeability, and a complete lack of pulmonary neutrophil recruitment. Further analysis reveals that only neutrophils but neither endothelial cells nor erythrocytes from Darc−/− mice confer protection from ALI. The protection appears to be due to abolished pulmonary recruitment of neutrophils from Darc−/− mice. The generation of neutrophil-platelet aggregates, a key mechanism in both pulmonary neutrophil recruitment and thrombus formation, is also affected by altered CXCR2 homeostasis in Darc−/− mice. CXCR2 blockade enhances the formation of platelet-neutrophil aggregates and thereby corrects a formerly unknown bleeding defect in Darc−/− mice. In summary, our study suggests that chemokine/chemokine receptor homeostasis plays a previously unrecognized and crucial role in severe ALI.
- endothelial cells
acute lung injury (ALI) is a devastating syndrome and remains associated with a mortality of up to 38%, despite the use of state-of-the-art treatment (33, 41). ALI can be caused by extrapulmonary causes, e.g., sepsis, massive transfusion, and trauma, as well as intrapulmonary causes, including pneumonia and acid aspiration (41). Aspiration may occur in unconscious and critically ill patients and is a well-known complication of general anesthesia (22, 27, 42). Aspiration of acid in mice and humans induces damage of the alveolar-capillary membrane, leading to neutrophil recruitment, pulmonary edema, and impairment of gas exchange (4, 5, 26). Neutrophil recruitment into the lung is a key event in the development of ALI (45, 47), but the molecular requirements are incompletely defined. Both chemokine receptors and platelets play a crucial role in neutrophil recruitment during ALI (43, 47). The formation of neutrophil-platelet aggregates enhances neutrophil recruitment and is pivotal in the pathogenesis of ALI (46). Several adhesion molecules as well as chemokines mediate these cell-cell interactions and promote neutrophil and platelet activation. Chemokines (chemoattractive cytokines) are small (4–8 kDa), basic proteins that can be classified on the basis of cysteine residue positioning at NH2 terminus (1, 25). Presentation of chemokines to rolling leukocytes on the luminal side of endothelial cells is crucial for leukocyte arrest as well as subsequent firm adhesion and transmigration (16, 17). CXC-chemokines, which have the first two cysteine residues separated by another amino acid, can be further subdivided based on the presence of a Glu-Leu-Arg (ELR) amino acid motif. ELR+ chemokines, including CXCL1/KC and CXCL2/MIP2, bind to the chemokine receptor CXCR2 and activate primarily neutrophils (1, 25). Chemokine receptors belong to the family of G protein-coupled receptors (25). Blockade of both CXC-chemokines and CXCR2 have attenuated the development of ALI in various animal models (29, 43).
DARC, a so-called silent chemokine receptor, binds inflammatory CC- and CXC-chemokines with high affinity and thereby participates in the regulation of chemokine homeostasis (6). As the intracellular domain of DARC lacks the Asp-Arg-Tyr consensus motif, it cannot interact with G protein-coupled receptors and consequently cannot activate intracellular signaling pathways (6). DARC is selectively expressed on erythrocytes and venular endothelial cells but not on leukocytes (9, 32). It is currently assumed that DARC on erythrocytes functions as a “chemokine sink” and regulates chemokine levels in the blood; the biological relevance of the sink function remains unknown (3, 7, 30). Endothelial DARC is thought to mediate transcellular transport of chemokines from the basal to the luminal side and to present them on the endothelial surface (23, 28, 30).
A recently published study convincingly demonstrated that DARC on endothelial cells mediates transcytosis and presentation of chemokines in vitro, which subsequently leads to increased leukocyte migration (28). Chemokine-induced leukocyte extravasation and contact-hypersensitivity reactions were also increased in mice overexpressing DARC on blood vessel endothelium. In line with these findings, we have recently shown that Darc gene-deficient (Darc−/−) mice are protected from LPS- and ischemia-reperfusion-induced acute kidney injury (44). Darc−/− mice completely lacked chemokine presentation on renal endothelial cells and displayed impaired recruitment of neutrophils. Darc−/− mice also exhibited lower plasma CXCL1/KC concentrations both at baseline and during severe systemic inflammation (44). However, Darc gene deficiency did not prevent the development of renal infiltrates in a mouse model of chronic renal inflammation (40).
Our study aimed to investigate leukocyte recruitment in Darc−/− mice, using both intravital microscopy of the cremaster muscle and a clinically relevant model of acid-induced ALI. We have further assessed platelet-neutrophil aggregates, as these are pivotal in the development of ALI. In conducting these in vivo experiments, we unexpectedly discovered a previously unrecognized bleeding defect in Darc−/− mice.
MATERIALS AND METHODS
The Animal Care and Use Committees of the University of Muenster, University of Virginia, and University of Pittsburgh approved all animal experiments, respectively. We used 8- to 12-wk-old wild-type C57BL/6 mice (WT; The Jackson Laboratory) and Darc−/− mice (back-crossed into C57BL/6 for at least 10 generations, Ref. 44). Mice were housed in barrier facilities under specific pathogen-free conditions.
Chemokine-induced arrest following local injection of chemokines was investigated by intravital microscopy of the cremaster muscle as described previously (15). Briefly, mice were anesthetized and placed on a heating pad. After cannulation of the carotid artery and tracheal intubation, the cremaster muscle was prepared for intravital microscopy. An intravital microscope (Axioskop; Zeiss, Thornwood, NY) with a saline immersion objective (SW 40/0.75 numerical aperture) was used for the observation of chemokine-induced arrest in postcapillary venules (diameter 20–40 μm). A charge-coupled device camera (model VE-1000CD; Dage-MTI, Michigan City, IN) was used for recording. Leukocyte arrest was investigated before and 1 min after intravenous injection of 600 ng of CXCL1/KC (PeproTech, Rocky Hill, NJ). Small venules (20- to 40-μm diameter) were used for microapplication of CXCL1/KC into the interstitial tissue next to the vessel wall. Self-made glass micropipettes (beveled tip with a diameter of 7–10 μm) were filled with ∼15 μl of a CXCL1/KC-containing solution (6 mg/ml). A small cushion (∼100 nl or 600 ng CXCL1/KC) of CXCL1/KC solution was injected under microscopic control and visibly expanded the interstitial tissue. Leukocyte arrest was investigated before and 1 min after injection of CXCL1 (PeproTech) as described previously (15). The hemodynamic parameters and surface area were measured and calculated as previously described (15).
LPS injection to induce severe systemic inflammation.
Mice received an intraperitoneal injection of LPS (10 μg/g body wt; Escherichia coli O111:B4) as well as subcutaneous injection of 20 μl/g lactated Ringer solution for fluid resuscitation. This approach represents a model of severe systemic inflammation, as it leads to organ dysfunction, such as acute kidney injury, within 24 h (36, 44). Here, mice were anesthetized to collect blood samples 4 h after LPS injections. Untreated, genotype-matched mice served as controls.
Plasma chemokine concentrations.
Using commercially available sandwich enzyme-linked immunoabsorbent assays (R&D Systems, Wiesbaden-Nordenstadt, Germany), we measured concentrations of the neutrophil-specific chemokine CXCL1/KC in plasma. Whole blood was obtained via cardiac puncture. Plasma was removed and stored at −80° until further use. Enzyme-linked immunoabsorbent assays were carried out according to the manufacturer's instructions. We have already reported absolute plasma concentration for these experiments (44).
Surface staining and flow cytometry for chemokine receptor CXCR2.
To characterize surface CXCR2 expression on neutrophils, saturating concentrations of anti-CXCR2 (PE), anti-CD45 (PerCP), and anti-Gr-1 (FITC) were added to 100 μl of heparin anti-coagulated blood (antibodies were obtained from R&D Systems, Minneapolis, MN). After RBC lysis with a 1.5 M NH4Cl solution, all samples were run on a FACScan flow cytometer (BD Biosciences, San Jose, CA). Data analysis was performed using FCS Express software (DeNovo Software, Los Angeles, CA).
In vitro chemotaxis assay.
We performed in vitro chemotaxis assays (under-agarose assays) to assess neutrophil response towards chemokines (8). Briefly, a 1.2% agarose solution with 50% H2CO3-buffered HBSS and 50% RPMI 1640 culture medium with 20% fetal calf serum was poured into petri dishes containing a placeholder for three wells (3.5 mm in diameter, 2.4 mm apart, forming a straight line). After solidification, the gels were kept in a 37°C/5% CO2 incubator for 1 h to equilibrate. The center well was then loaded with ∼105 isolated neutrophils in 10 μl of PBS solution (see below for neutrophil isolation). One of the outer wells was loaded with 10 μl of 1.0 μM CXCL2/MIP-2 in PBS; the other outer well was loaded with 10 μl of PBS (control). The gels were then incubated in a 37°C/5% CO2 incubator for 4 h to allow sufficient time for neutrophils to migrate. After 4 h, we counted the number of cells migrating towards CXCL2/MIP-2 and the number of cells migrating towards PBS control to calculate the chemotaxis ratio.
Acid-induced ALI was induced as previously described (47, 45). Briefly, mice were anesthetized and placed on a heating pad to maintain body temperature. Following tracheostomy and cannulation of the carotid artery, mice were ventilated with a respirator (MiniVent, Type 845; Hugo Sachs Elektronik) for 2 h (tidal volume, 10 μl/g body wt; respiratory rate, 140/min; fraction of inspiratory oxygen FiO2 = 0.21). To induce acid-induced ALI, mice received 2 μl/g HCl (pH 1.5) intratracheally, followed by a bolus of air (30 μl/g body wt). Control animals received saline instead of HCl and were also ventilated for 2 h.
Pulmonary function: oxygenation.
To calculate the oxygenation index (PaO2/FiO2, mmHg), arterial blood was taken from the arterial catheter, and the partial pressure of arterial oxygen was measured by a blood gas analyzer (Rapidlab 800 System; Bayer HealthCare).
Pulmonary function: vascular permeability.
Vascular permeability was determined by measuring protein concentration in the bronchoalveolar lavage supernatant (Lowry's method) (47).
PMN recruitment into the lung.
To determine global PMN content in the lung, we measured lung myeloperoxidase according to a previously published protocol (45). Briefly, lungs were homogenized in ice-cold KPO4 buffer. After removal of 17,000 g supernatants, pellets were resuspended in ice-cold KPO4 buffer, followed by two additional spins. Following these steps, 0.5% (wt/vol) hexacyltrimethylammonium bromide-10 mM EDTA in KPO4 was added to the remaining pellet. Suspensions were sonicated, freeze-thawed, and incubated for 20 min at 4°C. Supernatants were used subsequently to determine MPO. Then, assay buffer that comprised 0.2 mg/ml o-dianisidine and 158 mM H2O2 in 50 mM KPO4 was added to supernatants. Changes in absorbance were recorded at 460 nm over 3.5 min. Results were expressed as units of MPO/g of protein in the supernatant as determined by bicinchoninic acid assay (Pierce Chemical, Rockford, IL).
As described previously (44), total RNA was isolated from kidneys using TRIzol/chloroform-protocol (Invitrogen, Karlsruhe, Germany). Total RNA was eluted in diethylpyrocarbonate-treated water. Reverse transcription was accomplished using a Darc-specific primer (5′CCAGTAGCCCAGGTTGCATA′3, Invitrogen SuperScript II Protocol). The detection PCR was carried out with gene-specific primers (upstream primer 5′TGTCTGTATCCGGTGGAAACC′3 and downstream primer 5′CCAGTAGCCCAGGTTGCATA′3) gaining a specific product of 389 bp. PCR was performed using the Platinum Taq DNA polymerase kit (Invitrogen). RNA extraction and Darc-RT-PCR were normalized against glyceraldehyde-3-phosphate dehydrogenase-specific RT-PCR (upstream primer 5′CGGAGTCAACGGATTTGGTCGTAT′3 and downstream primer 5′AGCCTTCTCCATGGTGGTGAAGAC′3, gaining a specific product of 307 bp).
To further delineate the role of neutrophils vs. endothelial cells and erythrocytes in the development of ALI, we conducted an adoptive cell transfer, as described previously (45). Bone marrow-derived neutrophils from WT mice and Darc−/− mice were isolated using a Ficoll gradient. After depletion of endogenous neutrophils with a MAb directed against murine PMN (clone RB6-8C5; BD Biosciences, Heidelberg, Germany), WT and Darc−/− received 1.4 to 1.6 × 106 bone marrow-derived neutrophils from either WT or Darc−/− mice via tail vein injection (Fig. 1A). ALI was induced by intratracheal HCl application 2 h after reconstitution.
Mice were anesthetized and placed on a heating pad to maintain body temperature. A 3-mm segment of the tail tip was cut off with a scalpel. Bleeding was monitored with the tail submerged into isotonic NaCl solution (37°C). Bleeding time was recorded at the time after which bleeding had stopped for 5 s. Experiments were stopped after 10 min, and bleeding time recorded as 600 s, regardless of ongoing bleeding.
Test samples were analyzed using the native ROTEM technique based on the rotational thrombelastograph system (Pentapharm, Munich, Germany), as described elsewhere (11, 18). The parameters of ROTEM analyses included clot formation time (CFT, s), clotting time (CT, s), the α angle and maximum clot firmness (MCF, mm). We resorted to ROTEM because standard coagulation assays, including thrombin formation time and platelet aggregation assays, did not reveal any differences between WT and Darc−/− mice (data not shown) despite an abnormal bleeding time in Darc−/− mice.
To investigate platelet-neutrophil interactions, saturating concentrations of anti-CD41 (FITC), anti-CD45 (PerCP), anti-Gr-1 (Alexa Fluor 633), and anti-CD11b (PE) were added to 100 μl of acid-citrate dextrose (ACD) anticoagulated blood (47). After incubation for 10 min at room temperature, cells were fixed using paraformaldehyde. Neutrophil-platelet interactions were measured by flow cytometry (FACSCalibur; BD). Data analysis was performed using FCS Express software (DeNovo Software, Los Angeles, CA). In preliminary studies, we identified two populations of CD41+ neutrophils, i.e., two different populations of neutrophil platelet aggregates: platelet-rich (CD41high) and platelet-poor aggregates (CD41low) (Fig. 1B).
All data are presented as median (interquartile range). Statistical analysis included Shapiro-Wilks test for normality, one-way ANOVA, Kruskal-Wallis one-way analysis of variance, post hoc Student-Newman-Keuls test, t-test, and Mann-Whitney U-test (P < 0.05 was considered statistically significant).
There were no apparent phenotypic differences between WT and Darc−/− mice. However, Darc−/− mice displayed a small but statistically significant decrease in total WBC and increase in neutrophil counts (Table 1).
DARC is important for chemokine-mediated leukocyte recruitment in vivo.
To investigate whether DARC participates in chemokine transport across the endothelial layer in vivo, we conducted intravital microscopy of the cremaster muscle (Fig. 2). Venule diameters, wall shear rates, and leukocyte rolling all were similar between Darc−/− and WT mice (data not shown). Local extravascular application of KC via micropipette close to a venule significantly increased leukocyte adhesion in WT mice but not in Darc−/− mice (Fig. 2, A–C). Representative video micrographs of WT mice before (Fig. 2A) and after (Fig. 2B) local extravascular injection of CXCL1/KC are shown. Paravascular injections of FMLP resulted in similar increases in leukocyte adhesion in WT and Darc−/− mice (Fig. 2D). These data support the hypothesis that DARC is involved in leukocyte recruitment by means of chemokine transport across the endothelium. Intravascular injection of a moderate dose of CXCL1/KC resulted in an increase in adherent leukocytes that was only significant in WT but not in Darc−/− mice (Fig. 2E), indicating a role for DARC in chemokine presentation as well.
Prolonged CXCR2 downregulation in Darc−/− mice during systemic inflammation impairs neutrophil chemotaxis.
We have previously reported that compared with WT mice Darc−/− mice display lower absolute concentrations of circulating CXCL1/KC both at baseline and after high-dose LPS administration (44). However, the relative increase in CXCL1/KC is significantly greater in Darc−/− mice than in WT mice. Darc−/− mice exhibit a 1,000-fold CXCL1/KC release during systemic inflammation compared with a 200-fold increase in WT mice (Fig. 3A). The relatively greater release of CXCL1/KC during systemic inflammation in Darc−/− mice was associated with altered CXCR2 expression on neutrophils (Fig. 3B). Darc−/− mice displayed a significantly longer downregulation of CXCR2 on neutrophils than WT mice. CXCR2 expression on WT neutrophils returned to baseline within 4 h after LPS injection. However, CXCR2 expression on neutrophils from Darc−/− mice remained reduced after 4 h and was significantly lower than that in WT mice. Correspondingly, neutrophil chemotaxis towards CXCL2/MIP-2 4 h after LPS injection was significantly less in Darc−/− mice than in WT mice (Fig. 2E).
Darc gene deficiency protects from HCl-induced ALI.
We next compared Darc−/− mice and WT mice regarding the development of HCl-induced ALI (Fig. 4, A–D). Two hours after induction of ALI, WT mice showed a reduced oxygenation index (PaO2/FiO2) (Fig. 4A), increased recruitment of neutrophils into the lungs (Fig. 4B) and elevated protein concentrations in the BAL (Fig. 4C). Darc−/− mice, on the other hand, experienced a significant protection from ALI as demonstrated by preserved oxygenation index (Fig. 4A), lack of neutrophil recruitment (Fig. 4B), and lower protein levels in the BAL fluid (Fig. 4C). WT but not Darc−/− mice showed Darc mRNA expression 2 h after intratracheal HCl instillation (Fig. 4D). There was no detectable Darc mRNA expression in lungs of WT or Darc−/− mice. We have presented MPO data as changes relative to baseline, as absolute baseline neutrophil counts, and consequently MPO values (data not shown) were different between WT and Darc−/− mice.
Only neutrophils but neither endothelial cells nor erythrocytes from Darc−/− confer protection from HCl-induced ALI.
Reconstitution experiments with isolated neutrophils in the setting of HCl-induced ALI (Fig. 1A) served to further explore the protection from ALI seen in Darc−/− mice. Our data (Fig. 5, A and B) showed that only mice reconstituted with neutrophils from Darc−/− mice were protected from ALI. Independent of DARC expression on endothelial cells or erythrocytes in the recipient mice, injected neutrophils from Darc−/− mice preserved normal oxygenation (Fig. 5A) and lacked recruitment into the lung (Fig. 5B) after intratracheal instillation of HCl.
Differential effect of Darc deficiency on the formation of neutrophil-platelet aggregates.
The formation of aggregates between neutrophils (GR-1+/CD11b+) and platelets (CD41+) is crucial for neutrophil recruitment during ALI and for thrombus formation. Flow cytometry gated for neutrophils revealed two types of neutrophil-platelet aggregates (GR-1+/CD11b+/CD41+) in both WT and Darc−/− mice (Fig. 1B): platelet-rich aggregates (CD41high) and platelet-poor aggregates (CD41low).
At baseline, there were significantly more circulating neutrophil-platelet aggregates (CD41+) in Darc−/− mice than in WT mice (Fig. 6A). This was due to a greater number of CD41low aggregates in Darc−/− mice (Fig. 6E); the amount of CD41high aggregates was similar between WT and Darc−/− mice under baseline conditions (Fig. 6C). Severe systemic inflammation increased the overall formation of circulating neutrophil-platelet aggregates (CD41+) in WT but not in Darc−/− mice (Fig. 6A). Both WT and Darc−/− mice demonstrated a significant increase in CD41high aggregates after LPS administration; this increase was more pronounced in Darc−/− mice (Fig. 6B). Severe systemic inflammation significantly raised the formation of CD41low aggregates in WT mice but not in Darc−/−. By contrast, CD41low aggregates significantly declined in Darc−/− mice after LPS injection (Fig. 6E). Blockade of CXCR2 with a monoclonal antibody increased the formation of neutrophil-platelet aggregates in both WT and Darc−/− mice (Fig. 6, B, D, and F), in particular the formation of CD41high aggregates (Fig. 6D).
Darc gene deficiency causes a bleeding defect that is independent of endothelial cells.
As interactions between endothelial cells, neutrophils, and platelets are also important for thrombus formation, we measured bleeding time in WT and Darc−/− mice. Darc−/− mice displayed a significantly prolonged bleeding time compared with WT mice (Fig. 7A). Rotational thrombelastography also showed significantly impaired CFT and MCF in Darc−/− mice (Fig. 7, B and C), indicating that the bleeding defect in Darc−/− mice is independent of endothelial cells. Moreover, neutrophil depletion had no effect on bleeding time in either WT or Darc−/− mice, suggesting that mainly platelets from Darc−/− mice are involved in this bleeding defect. However, anti-CXCR2 treatment appeared to correct the bleeding defect.
Our data allow us to draw the following conclusions: Darc gene deficiency protects from severe HCl-induced ALI in mice. Neutrophils from Darc−/− mice, although lacking Darc expression, solely mediate protection from severe ALI. DARC is, nonetheless, critically involved in chemokine-mediated leukocyte recruitment in vivo. DARC enables the transendothelial transport of tissue chemokines as well as the intravascular presentation of circulating chemokines in vivo. However, Darc gene deficiency also alters chemokine homeostasis and consequently affects chemokine receptor homeostasis. A prolonged downregulation of CXCR2 on neutrophils in Darc−/− mice during severe systemic inflammation and subsequently impaired chemotaxis appears to be the consequence of a relatively greater release of CXCL1/KC. Formation of neutrophil-platelet aggregates, which is critical in both ALI and thrombus formation, is also affected by Darc gene deficiency. Further analyses led us to detect a previously unrecognized bleeding defect in Darc−/− mice. Overall, our findings suggest a critical, previously unrecognized role for DARC in chemokine/chemokine receptor homeostasis that is crucial for the development of severe HCl-induced ALI.
Silent chemokine receptors, such as DARC, have been implicated in maintenance of chemokine homeostasis; the knowledge regarding the biological relevance of this effect is incomplete (6, 31). There has been a long-standing controversy about the role of DARC in vivo. Using intravital microscopy and paravascular microinjections, we have provided further evidence for the role of endothelial DARC as a transendothelial transporter and intravascular presenter of chemokines, making it pivotal for leukocyte recruitment in vivo. Darc gene deficiency selectively impairs chemokine-mediated leukocyte arrest, a crucial step in leukocyte recruitment during inflammation (16). FMLP-induced leukocyte arrest is not affected by Darc gene deficiency. Our findings are in agreement with a recent study showing that overexpression of DARC enhanced chemokine-mediated transmigration of leukocytes both in vitro and in vivo. The authors further demonstrated that DARC internalizes and transports tissue-derived inflammatory chemokines onto the luminal endothelial cell surface in vitro (28).
We have previously shown that Darc gene deficiency leads to reduced levels of circulating CXCL1/KC both at baseline and after induction of severe systemic inflammation (44). We have also demonstrated that RBC-bound chemokine concentrations during severe systemic inflammation amount to only 30% of plasma chemokine concentrations, limiting the “sink effect” of erythrocyte DARC during severe inflammation (44). We here present data that, relative to baseline levels, Darc gene deficiency results in significantly higher circulating levels of CXCL1/KC during severe systemic inflammation. The relatively higher concentrations of CXCL1/KC in Darc−/− mice during severe systemic inflammation seem to result in a prolonged downregulation of CXCR2 on neutrophils, to which CXCL1/KC binds. The prolonged CXCR2 downregulation in Darc−/− mice in turn appears to impair neutrophil function, as demonstrated by reduced neutrophil chemotaxis towards CXCL2/MIP-2.
We used our recently described model of HCl-induced ALI (45, 47) to test the functional relevance of altered chemokine/chemokine receptor homeostasis in vivo. Development of ALI in this model is dependent on both neutrophils and CXCR2 (43, 47). Similar to our previous studies on the role of DARC in acute kidney injury (44), Darc gene deficiency confers complete protection from ALI, most likely due to abolished recruitment of neutrophils into the lungs. Reconstitution experiments revealed that Darc gene deficiency mediates protection neither via endothelial cells nor via erythrocytes but rather through neutrophils. As neutrophils do not express DARC, they are not primarily affected by Darc gene deficiency. Nonetheless, neutrophils from Darc−/− mice are exposed to a different chemokine environment and subsequently undergo altered CXCR2 expression, which in turn impairs chemotaxis. We conclude that chemokine/chemokine receptor homeostasis plays a greater role in the development of severe HCl-induced ALI than chemokine presentation on endothelial cells or erythrocytes. Also, small clinical studies have linked the degree of CXCR2 downregulation during severe inflammation after major trauma to different clinical courses; mild downregulation was associated with a higher rate of severe ALI, whereas extensive CXCR2 downregulation was associated with a greater incidence of posttraumatic sepsis (2, 37, 38).
Previous studies have used gene-deficient mice and chimeric mice to delineate the role of DARC in mild to moderate pulmonary inflammation (14, 30). Those data suggest that erythrocyte DARC restricts excessive PMN infiltration into the lung, whereas endothelial DARC exerts proinflammatory properties by means of transendothelial chemokine transport and luminal presentation of chemokines (13, 14, 30). Using a model of HCl-induced ALI, our results shed light on the role of DARC during severe pulmonary inflammation. Compared with previous studies with no reported mortality, our model represents a far more severe form of pulmonary inflammation, as 100% of WT mice die within 6 h of HCl instillation (47). Our data do not demonstrate a relevant role for DARC in chemokine transport or presentation during severe pulmonary inflammation but rather a key role for DARC-dependent chemokine homeostasis. We have recently shown that during severe inflammation the erythrocyte-bound chemokine concentrations amount to only ∼30% of plasma concentrations (44). These data suggest an exhaustion of chemokine-binding capacities by erythrocytes and thus a limited role for DARC as a chemokine sink during severe inflammation. By contrast, erythrocyte-bound chemokine concentrations are twofold higher than plasma chemokine concentrations during mild pulmonary inflammation (30), suggesting nonsaturated chemokine-binding capacities on erythrocytes and thus an important role for DARC as a chemokine sink.
Instead of bone marrow transplantation to generate chimeric mice (14, 30), we have performed reconstitution experiments, allowing us to study the acute short-term effects of altered chemokine/chemokine receptor homeostasis. This approach eliminates possible interference of adaptive changes during recovery from bone marrow transplantation. Also, in bone marrow transplantation all hematopoietic cells are replaced, not just neutrophils.
We have confirmed our findings in two models of severe inflammation, ALI and systemic inflammation after LPS injection. Although the timing itself was different for each model, we have chosen the respective time points to account for model-specific effects on maximum neutrophil recruitment, as determined in preliminary studies (data not shown).
Earlier studies have reported an increased neutrophil recruitment into various organs in Darc−/− mice during inflammation (3, 20, 21). A conclusive comparison with the first study (3) is limited because of missing baseline MPO data for Darc−/− mice, which we have found to be different from WT mice (data not shown). A thorough discussion of the second study (20) is complicated by partial retraction of some data (21) without detailed display of new data and by conflicting findings regarding neutrophil recruitment into different organs. In summary, our data do not seem to contradict previous studies but rather expand existing knowledge by focusing on DARC during severe inflammation and its role in maintenance of chemokine/chemokine receptor homeostasis.
The observation of altered formation of neutrophil-platelet aggregates in our study provides further insight into possible mechanisms of reduced pulmonary neutrophil recruitment and subsequent protection from ALI in Darc−/− mice. Formation of neutrophil-platelet aggregates represents a key component in the pathogenesis of acid-induced lung injury (47). Aggregate formation is mainly mediated through proinflammatory mediators, such as platelet P-selectin, PSGL-1 on neutrophils, as well as β2- and β3-integrins (46). Elimination of Darc modulates the number and the composition of neutrophil-platelet aggregates. Darc gene deficiency causes an increased formation of CD41low aggregates, i.e., platelet-poor aggregates. Formation of neutrophil-platelet aggregates, in particular CD41high neutrophil-platelet aggregates, appears to be CXCR2 dependent, as the application of a blocking CXCR2 antibody stimulates the formation of neutrophil-platelet aggregates. The exact mechanism for the effect of CXCR2 blockade is unknown at this time, as platelets themselves do not express CXCR2. However, homotypic neutrophil aggregation, i.e., formation of neutrophil-neutrophil aggregates, occurs in vivo and in vitro and depends on ligands for CXCR2 (19, 35, 39). Severe systemic inflammation gives rise to increased formation of CD41+ aggregates in WT but not in Darc−/− mice. This is largely due to an actual decrease in CD41low aggregates in Darc−/− mice. CD41high aggregates are elevated in both WT and Darc−/− mice after high-dose LPS injections. This process appears, at least partially, to be CXCR2 dependent. LPS induces CXCR2 downregulation on neutrophils, and CXCR2 blockade enhances formation of CD41high aggregates. Although they represent potentially attractive hypotheses, we can currently only speculate about different roles of CD41high and CD41low aggregates in thrombus formation and leukocyte recruitment, respectively.
Darc gene deficiency increases the formation of platelet-poor aggregates but not that of platelet-rich aggregates. Glycoprotein IIb/IIIa inhibition leads to a rise in neutrophil-platelet aggregates and reduces platelet-platelet aggregates (10, 34), resulting in platelet-specific hemostatic defects. Preliminary plasma coagulation and platelet aggregation studies did not show any difference between WT and Darc−/− mice. However, both bleeding time studies and rotational thrombelastography revealed significant hemostatic abnormalities in Darc−/− mice, which have not been recognized previously. The thrombelastography pattern in Darc−/− mice is similar to that observed after glycoprotein IIb/IIIa inhibition, i.e., prolonged clot formation time and reduced maximum clot firmness (11, 12, 24). As neither endothelial cells, nor neutrophils, nor plasma coagulation factors appear to be major factors in this bleeding defect, platelets emerge as likely mediators. The exact mechanisms remain to be identified at this time. CXCR2 blockade does not seem to have a prohemostatic effect under normal conditions. Nonetheless, CXCR2 blockade increases the formation of neutrophil-platelet aggregates, in particular platelet-rich aggregates, and corrects the prolonged bleeding time in Darc−/− mice, reversing the effect of Darc gene deficiency on both the formation of neutrophil-platelet aggregates and bleeding time. One can therefore hypothesize that enhanced formation of neutrophil-platelet aggregates, in particular formation of platelet-rich aggregates, enabled normalization of bleeding time in Darc−/− mice.
In summary, our study shows that DARC participates in inflammatory leukocyte recruitment by transendothelial transport of chemokine and maintenance of chemokine/chemokine receptor homeostasis, in particular CXCL1/KC and CXCR2 homeostasis. Darc gene deficiency leads to impaired leukocyte arrest on endothelial cells and blocks pulmonary neutrophil recruitment during severe ALI. The latter effect is independent of endothelial cells and erythrocytes but dependent on neutrophils, supporting a biologically relevant role for DARC in the maintenance of chemokine/chemokine receptor homeostasis. Moreover, maintenance of chemokine/chemokine receptor homeostasis appears to have a greater impact on neutrophil recruitment and subsequent development of severe ALI than chemokine presentation on endothelial cells or erythrocytes. Darc gene deficiency also affects the generation of neutrophil-platelet aggregates, a key mechanism in both neutrophil recruitment and thrombus formation. CXCR2 blockade differentially enhances the formation of platelet-neutrophil aggregates and thereby corrects a previously unrecognized bleeding defect in Darc−/− mice.
This study was supported by grants from the Else Kröner Fresenius Stiftung (A. Zarbock, K. Singbartl) and the Deutsche Forschungsgemeinschaft (ZA428/3-1 to A. Zarbock).
No conflicts of interest are declared by the author(s).
We thank Dr. Klaus Ley for critical discussion of the manuscript.
- Copyright © 2010 the American Physiological Society