Aspirin-triggered 15-epi-lipoxin A₄ regulates neutrophil-platelet aggregation and attenuates acute lung injury in mice

The acute respiratory distress syndrome (ARDS) is a common and often fatal syndrome characterized by neutrophilic inflammation. In ARDS and experimental models of acute lung injury (ALI), an exuberant inflammatory response with release of proteases and other toxic mediators from activated leukocytes leads to tissue damage, increased permeability of the alveolar-capillary barrier, and the formation of protein-rich lung edema.¹ 

Several lines of evidence indicate that activated platelets act as important effectors in host defense, influencing pulmonary neutrophil recruitment and contributing to the development of ALI.^2-4  Platelets become activated within the blood circulation at sites of inflammation and can adhere to other platelets, to leukocytes, forming neutrophil-platelet aggregates (NPAs) or monocyte-platelet aggregates (MPAs), and to exposed endothelium.⁵  The presence of circulating leukocyte-platelet aggregates (LPAs) is a sensitive indicator of platelet activation and has been observed in patients with inflammatory conditions including sepsis, coronary diseases, rheumatoid arthritis, inflammatory bowel disease, cystic fibrosis, and ARDS.^6-10 

Activated platelets synthesize thromboxane A₂ from arachidonic acid via a sequential reaction catalyzed by cyclooxygenase-2 (COX-2). Aspirin (acetylsalicylic acid [ASA]) acetylates COX-2, and this conformational change leads to inhibition of prostanoid synthesis.¹¹  Acetylation of COX-2 switches catalytic activity to convert arachidonic acid to 15R-hydroxyeicosatetraenoic acid, which can be subsequently converted to 15(R)-epi-lipoxin A₄ (15[R]-epi-LXA₄), also known as ASA-triggered lipoxin (ATL).¹²  Lipoxins are endogenous lipid mediators generated during inflammation that can block inflammatory cell recruitment, inhibit cytokine release, and decrease vascular permeability, which collectively are anti-inflammatory properties.^13,14  Lipoxins also stimulate epithelial reconstitution, facilitate apoptosis, and promote macrophage efferocytosis, which collectively are proresolution properties.^13,14  Production of lipoxins requires transcellular biosynthesis, with leukocytes, platelets, and tissue resident cells having important roles.^12,15 

ATL shares the potent anti-inflammatory actions of lipoxins but is more resistant to metabolic inactivation.¹⁶  The anti-inflammatory, proresolving properties of both LXA₄ and 15(R)-epi-LXA₄ are mediated through the formyl peptide receptor type 2 (termed Fpr2/3 in the mouse), also called the lipoxin A₄ receptor (ALX),¹⁷  which is expressed in a variety of cell types including endothelial cells and myeloid lineages.^17,18  In addition to native lipoxins and ATL, ALX recognizes a diverse array of bioactive mediators including annexin A1, resolvin D1, and serum amyloid A protein.^17,19  Recent work has elucidated the molecular mechanisms behind the dual actions of ALX (pro- vs anti-inflammatory responses), which is determined by unique receptor dimerization patterns induced by specific ligands.²⁰ 

Here, we demonstrate the role of ATL in regulating leukocyte-platelet aggregation, and consequently lung injury, in both direct and indirect experimental models of ALI: intratracheal instillation of lipopolysaccharide (LPS) and a 2-event model of transfusion-related acute lung injury (TRALI).²¹  We present evidence that platelets have a central role in producing pulmonary inflammation and injury based on critical interactions with neutrophils that are regulated by ATL.

The following reagents were used: ASA, LPS from Escherichia coli O55:B5 (Sigma-Aldrich), Boc2 (Genway), and 15-epi-LXA₄ (EMD Biosciences).

Experimental procedures were performed in 8- to 12-week-old male BALB/c wild-type (WT) mice (Charles River) housed under pathogen-free conditions in the Animal Barrier Facility at University of California, San Francisco. Mice were pretreated (intraperitoneally) with 0.1 mg/g of ASA or dimethylsulfoxide (DMSO) vehicle, 24 and 2 hours before ALI induction. Mice were challenged with intratracheal phosphate-buffered saline (PBS; controls) or LPS (5 μg/g), and analysis was performed 4 to 48 hours later. In selected experiments, mice received 10 μg/kg intraperitoneally of Boc2 30 minutes before ASA injections and 24 hours after LPS instillation or 100 to 5000 ng/mouse of 15-epi-LXA₄ 30 minutes before and 24 hours after LPS challenge. There is no appropriate peptide control for Boc2. Fpr2/3^−/− mice were obtained from Rod Flower (St. Barts and the London School of Medicine and Dentistry).¹⁹ 

Bronchoalveolar lavage (BAL) was obtained at 48 hours by placing a catheter into the trachea, through which 1 mL of cold PBS was flushed back three times. Protein concentration in the cell-free BAL was determined using a BCA protein assay kit (Thermo Scientific). BAL leukocytes were counted using a Coulter counter (Beckman Coulter), and differential was determined by cytospin preparation (Cytospin 3; Thermo Electron Corp.) and Diff-Quick staining. BAL hemoglobin was measured using LowHb microcuvettes (HemoCue).

A 2-event TRALI model was used, as previously described.²¹  Briefly, after being primed with LPS (0.1 mg/kg, intraperitoneally) for 24 hours, mice were challenged with MHC I monoclonal antibodies (mAb) (H2K^d; immunoglobulin G_2a, κ; 1.0 mg/kg) injected into the jugular vein and euthanized after 2 hours. To quantify pulmonary edema formation, bloodless, extravascular lung water was measured.^21,22  We also measured lung vascular permeability to protein by instilling mice with intravenous ¹²⁵I-labeled albumin. The radioactivity in the blood and the bloodless lung was measured with a gamma counter (Packard 5000 Series), and the ratio was used to calculate the lung extravascular plasma equivalents (EVPEs).^21,22  ASA (0.1 mg/g), or DMSO control, was delivered intraperitoneally 30 minutes prior to LPS priming and again 2 hours prior to H2K^d mAb challenge. Boc2 (10 μg/kg), or vehicle (PBS), was administrated intraperitoneally 30 minutes prior to the ASA injections. 15-Epi-LXA₄, or vehicle (ethanol), was administered via tail vein 30 minutes prior to H2K^d mAb injections.

For histological analysis, 7 μm of optimal cutting temperature compound-embedded frozen lung sections (without lung perfusion) was fixed in cold acetone and blocked with PBS containing 8% rabbit serum and 4% bovine serum albumin. Sections were incubated with CD41 antibody (clone MWReg30; BD Biosciences), followed by peroxidase-conjugated secondary Abs, color development, and hematoxylin counterstaining.²¹ 

Western blots on digested lungs and BAL samples were done using standard techniques and immunoblotted with CD41 and α-tubulin Abs.

15-epi-LXA₄ and thromboxane B₂ concentrations in plasma and BAL were determined using enzyme-linked immunosorbent assay kits from Neogen and GE Healthcare, respectively. We followed the manufacturer’s instructions for lipid extraction.

Cell-free BAL and myeloperoxidase (MPO) standards (Sigma-Aldrich) were placed onto a 96-well plate, and 200 μL of o-dianisidine solution and 10 μL of 0.1% H₂O₂ were added. The absorbance was read after 5 minutes at 405 nm and expressed as units per milliliter of BAL supernatant using a standard curve.²¹ 

Whole blood was collected into acid citrate dextrose (Sigma-Aldrich). Samples in the presence of FcγRII/III blocking Ab (2.4G2) were diluted with Tyrode’s buffer supplemented with 10 U/mL of heparin (APP), 7 U/mL Apyrase (Sigma-Aldrich), and 0.5 μM prostaglandin I₂ (Sigma-Aldrich). eFluor450-CD11b (eBioscience), phycoerythrin-Ly6G (BD Bioscience), allophycocyanin-CD41 (eBioscience), and isotope controls Abs were used to detect leukocyte and platelets antigens. Samples were examined with a LSRII/Fortessa flow cytometer (BD Bioscience). Neutrophils and monocytes were gated by their forward- and side-scatter characteristics and by their Ly-6G⁺/CD11b⁺ (neutrophil) or Ly-6G⁻/CD11b⁺/ Ly-6C⁺ (monocyte) expression pattern. Platelet-neutrophil or platelet-monocyte aggregates were detected by CD41 Ab staining. In selected experiments, whole lungs were digested,²¹  followed by the same general protocol described above. All data were analyzed using FlowJo software (TreeStar).

Whole blood was collected from C57BL/6 WT and Fpr2/3^−/− mice into acid citrate dextrose and used to isolate platelets.²³  Bone marrow neutrophils were isolated, as previously described.²⁴  Neutrophils and platelets (1:100 ratio) from WT and Fpr2/3^−/− mice were incubated for 1h at 37°C with LPS (5 μg/mL) or with 15-epi-LXA₄ (300 nM), in presence of Tyrode’s buffer. After this incubation, Tyrode’s buffer with supplements (see above) was added. Flow cytometry was performed as described above.

We used 2-photon intravital microscopy²⁵  to record in real-time neutrophil and platelet interactions in LPS-induced ALI. To facilitate tracking of both platelets and neutrophils, PF4-cre × Rosa26-LSL-tdTomato mice (both from The Jackson Laboratories) were crossed with LysM-eGFP mice (obtained from E. Robey, University of California, Berkeley, CA). To permit identification of the lung vasculature, Cascade-blue dextran (50 μL of 25 mg/mL; Life Technologies) was injected into the jugular vein. Mice were treated with LPS ± ASA and compared with LPS untreated mice (PBS). We captured a 0.4-mm² x-y surface area at 40 μm z-depth, capturing a complete image every 1 minute for 60 minutes. Images were analyzed using Imaris 7.6.1 software (Bitplane) to quantify mobile neutrophils (green) that colabeled for platelets (red). The number of events was averaged per minute, and the mean of the highest 20 values obtained was recorded as the number of neutrophil-platelet aggregates per 0.1 mm² of lung tissue at 40 μm z-depth.

All in vivo and in vitro experiments were repeated a minimum of 3 independent times. Results are reported as both individual data points and mean ± standard deviation (SD). To determine significance, 2-tailed Student t test, analysis of variance, and log-rank (Mantel-Cox) tests were used as appropriate (GraphPad PRISM version 5.0). P ≤ .05 was deemed to be significant.

All experiments were approved by the Institutional Animal Care and Use Committee at the University of California, San Francisco.

To determine the role of platelets in experimental ALI, we used an intratracheal LPS model of lung injury. Immunohistochemistry for the platelet specific marker CD41, 48 hours after intratracheal LPS instillation (5 µg/g), revealed intense sequestration of platelets within LPS-injured lungs compared with PBS-instilled control animals (Figure 1A-B). We cannot rule out that areas of particularly intense CD41 staining at vessel junctions could indicate the presence of megakaryocytes. CD41 protein expression in total lung digest was analyzed by western blot and confirmed the sequestration of platelets within the lungs of LPS-challenged animals (Figure 1D). Thromboxane A₂ is a marker of platelet activation and is rapidly hydrolyzed to its inactive stable metabolite, thromboxane B₂ (TXB₂). Plasma TXB₂ levels were increased after LPS challenge (Figure 1E). Pretreatment with ASA 24 and 2 hours before LPS challenge prevented platelet accumulation in the lungs (Figure 1C-D) and decreased plasma TXB₂ levels (Figure 1E) compared with controls (DMSO).

The total number of white blood cells (WBCs) and neutrophils in BAL was measured 48 hours after intratracheal PBS (control) or LPS instillation. MPO activity, an index of neutrophil activation, was also measured in cell-free BAL supernatant. The number of WBCs and neutrophils within BAL increased with LPS challenge, as did MPO activity (Figure 2A-B). BAL total protein was measured as an indicator of lung permeability and increased after LPS instillation (Figure 2C). Mice treated with ASA had significantly decreased BAL WBCs, neutrophils, MPO activity, and total protein compared with vehicle controls (Figure 2A-C).

CD41 immunohistochemistry of lungs after LPS challenge revealed intense sequestration of platelets within the pulmonary capillaries but also suggested the presence of intra-alveolar platelets (Figure 1B, inset). To further test for intra-alveolar platelet accumulation, we probed for CD41 protein via western blot analysis on cells obtained from BAL. CD41 protein was increased in BAL from LPS-challenged mice compared with PBS controls (Figure 2D). We also used flow cytometry and a CD41 mAb to determine the presence of platelets in the BAL and did not detect intra-alveolar platelets in control animals, but a sharp increase was detected after LPS challenge (Figure 2E). We tested for pulmonary hemorrhage as a potential cause of intra-alveolar accumulation of platelets by measuring BAL hemoglobin and found a minimal increase in hemoglobin in the LPS-challenged animals (supplemental Figure 1, available on the Blood Web site). We detected significantly increased TXB₂ concentrations in the cell-free BAL supernatant of LPS-treated mice compared with controls (Figure 2F). These results demonstrate that platelets accumulate and are activated in the alveolar spaces after LPS challenge. ASA treatment yielded a significant reduction in both alveolar LPS-induced CD41 expression (Figure 2D-E) and platelet activation (Figure 2F) and did not affect BAL hemoglobin measurements.

We investigated the LPS-induced formation of LPAs in whole blood and BAL using flow cytometry. The gating strategies for whole blood and BAL are illustrated in supplemental Figure 2A-B. Briefly, antibodies against the leukocyte-specific surface markers CD11b, Ly6G, or Ly6C, in addition to the platelet-specific marker CD41, were used to identify MPAs, defined as CD11b⁺/Ly6C⁺/Ly6G⁻/CD41⁺, and NPAs, defined as CD11b⁺/Ly6G⁺/CD41⁺. Whole blood obtained by cardiac puncture revealed that LPS induced both NPA (Figure 2G) and MPA (Figure 2H) formation as early as 4 and 24 hours, respectively, after LPS instillation. Blood NPA and MPA levels progressively increased over time, reaching a maximum at 48 hours after LPS instillation (NPA = 62 ± 5.6%, MPA = 48.6 ± 3.4%). The percentage of NPA in the BAL of LPS-challenged mice also increased (Figure 2I), but we did not detect a similar increase in BAL MPAs (data not shown). Treatment with ASA significantly decreased blood NPA and MPA formation, as well as BAL NPA formation, at 48 hours after LPS (Figure 2G-I).

We used 2-photon intravital lung microscopy²⁵  to determine the spatial and temporal formation of NPA within the pulmonary microcirculation at baseline and after LPS-induced ALI. To track platelets, Rosa26-LSL-tdTomato mice were crossed with PF4-cre mice,²⁶  and in the peripheral blood of these mice, >99% of the tdTomato⁺ events were also CD41⁺ (supplemental Figure 3). These mice were crossed with LysM-eGFP mice to facilitate neutrophil identification. Cascade-blue dextran was injected intravenously to delineate the pulmonary microvasculature. In the triple transgenic mice without LPS exposure, we observed the dynamic formation of NPA within the lung microcirculation, with intravascular neutrophils interacting with platelets (example shown in Figure 3A and supplemental Movie 1). We also observed NPAs in PBS-challenged mice (Figure 3B,E; supplemental Movie 2), and in the first hours after LPS exposure, there was a sharp increase in NPA formation (Figure 3C,F; supplemental Movies 3 and 4), which was ameliorated with ASA pretreatment, with NPA formation decreasing to basal levels (Figure 3D,G; supplemental Movie 5). At later time points after LPS challenge (24 hours), we detected immobile NPAs that were outside of the intravascular dextran signal (Figure 3H-I; supplemental Movie 9), consistent with intra-alveolar NPAs. Areas of neutrophil-platelet colocalization were analyzed by surface rendering (supplemental Movies 6-8) and quantified in Figure 3J.

Plasma and BAL 15-epi-LXA₄ concentrations were measured by enzyme-linked immunosorbent assay after PBS and LPS challenge ± aspirin treatment. LPS did not affect plasma or BAL 15-epi-LXA₄ levels, but ASA treatment significantly increased both plasma and BAL 15-epi-LXA₄ concentrations in LPS-challenged mice at 48 hours compared with PBS or vehicle controls (Figure 4A-B).

To determine whether elevated 15-epi-LXA₄ contributed to the decreased lung injury observed after aspirin treatment, we tested the ALX (Fpr2/3) antagonist, Boc2.²⁷  The administration of Boc2 30 minutes before aspirin treatment, and repeated 24 hours after LPS instillation, reversed the protective effects of aspirin on lung injury (Figure 4C-D). Additionally, Boc2 partially reversed the decrease in whole blood NPAs after ASA treatment (Figure 4E). Boc2 treatment alone (in the absence of ASA) had no effect on LPS-induced lung inflammation or injury (data not shown). Also, Boc2 did not change the plasma 15-epi-LXA₄ concentrations of ASA treated mice (data not shown).

Next, we confirmed the results obtained with pharmacologic blockade of Fpr2/3 using the recently characterized Fpr2/3^−/− mice.¹⁹  ASA treatment of Fpr2/3^−/− animals increased plasma concentrations of 15-epi-LXA₄ in the LPS model (Figure 4F). Compared with WT controls, Fpr2/3^−/− mice had no significant differences in BAL WBCs, neutrophils, or total protein after LPS challenge (data not shown), and we observed no significant decreases in BAL WBCs, neutrophils, or total protein after ASA treatment (Figure 4G-H). The ASA-induced decrease in blood and BAL NPA was also blocked in Fpr2/3^−/− mice (Figure 4I-J).

To evaluate whether pharmacologic doses of 15-epi-LXA₄ can modulate LPS-induced lung injury, mice were treated with increasing doses of this lipid mediator. Intravenous administration of between 100 and 1500 ng of 15-epi-LXA₄ did not produce a significant change in BAL WBC counts or total protein after LPS challenge (Figure 5A-B). However, doses of 5 µg resulted in a significant decrease in BAL WBCs (Figure 5A) and attenuated lung permeability (Figure 5B) and whole blood NPA formation (Figure 5C) in a similar manner to that observed with aspirin treatment.

To determine whether lipoxin receptor signaling on neutrophils and/or platelets mediated the protective effects observed after aspirin and 15-epi-LXA₄ treatment, we developed a coculture system in which neutrophils and platelets, freshly isolated from WT or Fpr2/3^−/− mice, were incubated with LPS (5 μg/mL) to induce NPA formation. We observed a basal formation of NPA after a 1-hour incubation of WT neutrophils and platelets (1:100), which increased after LPS treatment (Figure 5D). Pretreatment with 15-epi-LXA₄ (300 nM) significantly decreased LPS-induced NPA formation (Figure 5D). We next tested combinations of WT and Fpr2/3^−/− neutrophils and platelets and observed that the presence of a functional lipoxin receptor was required on both neutrophils and platelets for 15-epi-LXA₄ to decrease NPA formation (Figure 5D).

We previously demonstrated that ASA treatment decreases platelet sequestration, lung vascular permeability, and edema formation, and increases survival in an experimental TRALI model.^2,24  Here, we investigated whether ATL was required for these effects. In a 2-event model of TRALI,²¹  we treated animals with aspirin ± Boc2, or vehicle (DMSO), 30 minutes prior to the intraperitoneal LPS priming, and again 2 hours prior to challenge with H2K^d mAb. Both aspirin and ASA + Boc2 treatment significantly increased 15-epi-LXA₄ concentrations in this model (Figure 6A). Lung edema formation, lung vascular permeability, hemoconcentration, and mortality were all reduced with ASA treatment compared with vehicle-treated controls (Figure 6B-D,F). Extravascular lung water and EVPE values in LPS-primed controls are 0 and <10 µL, respectively.²¹  Hematocrit values in LPS-primed controls are ∼50%.²¹  TRALI induced the formation of lung NPA, which was ameliorated by aspirin treatment (Figure 6E). Pretreatment with Boc2 significantly reversed the protective effects of aspirin on lung injury, NPA formation, and mortality (Figure 6B-F). The administration of Boc2 alone did not affect TRALI severity (data not shown).

We subsequently tested the impact of treatment with 15-epi-LXA₄ or vehicle (intravenous) administered 30 minutes prior to H2K^d mAb challenge on lung injury and NPA formation. 15-Epi-LXA₄ treatment significantly decreased lung injury and formation of lung NPA and eliminated mortality compared with vehicle-treated animals (Figure 7A-E).

The main conclusions from this study are that (1) pulmonary platelet sequestration and activation, including intra-alveolar accumulation, are major features of ALI; (2) NPAs are dynamically formed in the lung microcirculation and sharply increase in the blood and alveolar compartments after injury; (3) ATL is required for aspirin-mediated protection in ALI, and ATL signals via Fpr2/3 on neutrophils and platelets to regulate the formation of NPA; and (4) in 2 experimental models of ALI, ATL treatment alone significantly reduces lung injury.

Platelets are capable of rapidly responding to inflammatory stimuli, releasing preformed mediators and even synthesizing immune mediators from mRNA transferred from megakaryocytes.²⁸  Platelets and platelet microparticles²⁹  can bind to leukocytes and the endothelium, influencing the function of these cells during inflammation. Indeed, using lung intravital imaging, we now show that the lung is a site of de novo NPA formation under basal conditions. The unique structural features of the lung microcirculation that influence neutrophil trafficking³⁰  and the proposed role of the lung in thrombopoiesis³¹  may enhance these interactions. With inflammation, we found an early and rapid increase in dynamic NPA formation in the lung microcirculation, and many of these aggregates were ultimately detected in the alveolar spaces. Previous reports of NPA formation in inflammation have shown that neutrophils bound to the endothelium secondarily capture platelets,^32,33  but our direct imaging of the lung microcirculation reveals that these aggregates are mobile and are eventually capable of migrating into the alveolar spaces. It has recently been described that the PF4-cre mouse is potentially leaky in nonmegakaryocyte lineages³⁴ ; however, in the peripheral blood, we found that >99% of the PF4-cre–driven tdTomato⁺ events were platelet specific (CD41⁺).

There is a current lack of understanding of the regulation of LPA formation. Here, we propose that anti-inflammatory lipoxins and ATL specifically regulate this process. Lipoxins and ATL exert potent anti-inflammatory and pro-resolution bio-actions. In particular, they block chemotaxis and adherence to microvasculature, inhibit nuclear factor-κB activation in leukocytes, and block the release of proinflammatory cytokines such as interleukin-6, interleukin-8, and tumor necrosis factor-α.^16,35-38  A potential limitation of our study is that microgram quantities of ATL were required to limit lung injury in our experiments, whereas others have reported that nanogram amounts of proresolving lipid mediators are sufficient.¹⁴  However, we treated with ATL at just 2 time points in our 48-hour LPS model, and more frequent dosing of ATL, which has a half-life of a few minutes, may have allowed for lower overall amounts. Also, we have not ruled out the contribution of other ASA-triggered lipid mediators, such as resolvins, but results from our pharmacologic experiments support a major role of ATL in mediating the protective effects of ASA.

Others have shown protective effects of lipoxins in acute inflammation. LXA₄ and ATL were protective in an acid-induced ALI model through the modulation of neutrophil apoptosis.³⁹  ATL was also protective in mice challenged with LPS through a heme-oxygenase1–dependent mechanism.⁴⁰  Endogenous LXA₄ modulates ischemia-reperfusion injury in the gut, and Fpr2/3^−/− mice have more inflammation and injury compared with WT animals.⁴¹  In this same study, NPA formation was investigated using chimeric experiments to test whether platelet or neutrophil Fpr2/3 was more important in aggregate development, and neutrophil Fpr2/3 was determined to be critical.⁴¹  However, in our neutrophil-platelet aggregation assay, we found that neither neutrophil nor platelet Fpr2/3 is required for the formation of aggregates in response to LPS, but Fpr2/3 on both cells is required for lipoxin-mediated inhibition of NPA formation.

For the first time, we showed the presence of platelets in the alveolar spaces in ALI. Others have shown in experimental allergic inflammation that platelets migrate out of vessels and localize beneath airways⁴²  or into the synovial spaces in rheumatoid arthritis.⁸  Platelets have also been observed by electron microscopy extravasating out of the lung vasculature in patients with ARDS.⁴³  In our studies, many of the platelets were bound to neutrophils, and the dynamics of NPA migration into the alveolar spaces after LPS challenge seems to occur hours after the very early formation of NPAs in the lung vessels. More work is needed to determine the mechanisms of platelet transmigration in the lung, including the chemotactic stimuli for platelets, and whether transmigration is facilitated by tethering to neutrophils or occurs independently. Notably, our assays could have detected intact platelets or platelet microparticles in the alveolar spaces, and the specific role of platelet microparticles in lung inflammation should be addressed in future studies.

We previously reported that aspirin is protective in a 2-event model of TRALI by decreasing platelet sequestration and activation and neutrophil extracellular traps formation.^2,21  Here, we found that ATL and ALX were required for the protective effects of aspirin in TRALI. Indeed, treatment of mice with ATL alone yielded significant protection from TRALI. These results have important clinical implications for patients with ARDS including transfused, critically ill patients. Platelet depletion is impractical, as is aspirin therapy in many critically ill patients, but treatment with ATL, or perhaps other anti-inflammatory and proresolving mediators,⁴⁴  is an attractive approach. Further, the measurement of LPA in patients with ARDS and other critical illnesses could be a useful biomarker of inflammation and could be measured serially to assess therapeutic responses to treatment with pro-resolving lipid mediators.

In summary, we showed in 2 experimental models of ALI that aspirin-triggered lipoxin is a powerful inhibitor of neutrophil- and platelet-mediated lung inflammation and injury. Ironically, surface contact between neutrophil and platelets yields transcellular production of lipoxins and also leukotrienes,⁴⁵  and our results indicate that signaling through ALX on neutrophils and platelets in turn regulates their tethering during inflammation. Aspirin is likely exerting its therapeutic benefit through a combination of inhibiting the production of proinflammatory eicosanoids and triggering the formation of proresolving lipid mediators.⁴⁶  Even low-dose aspirin administered to healthy volunteers is capable of producing bioactive levels of ATL,⁴⁷  and a clinical trial (ClinicalTrials.gov identifier #NCT01504867) is underway randomizing patients at risk for ARDS to aspirin or placebo. ATL and potentially other pro-resolving lipid mediators are attractive therapeutic options for patients with ARDS: a life-threatening syndrome that lacks any effective pharmacotherapies.

The authors thank K. Corbin and H. Pinkard at the University of California, San Francisco Biological Imaging Development Center for assistance with the intravital lung microscopy experiments and Rod Flower at the William Harvey Research Institute, St. Barts, and the London School of Medicine and Dentistry for providing the Fpr2/3^−/− mice.

This work was supported by the National Institutes of Health National Heart, Lung, and Blood Institute grants R01 HL107386 (M.R.L.) and P01 HL024136 (M.F.K).

Contribution: G.O.-M. designed and performed the experiments, analyzed the results, and prepared the manuscript; B.M. performed experiments and edited the manuscript; A.B. assisted with the intravital lung imaging experiments and edited the manuscript; M.H. assisted with the intravital lung imaging experiments and edited the manuscript; M.F.K. assisted with the intravital lung imaging experiments and edited the manuscript; and M.R.L. designed and assisted with the experiments, analyzed the results, provided funding, and prepared and edited the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Mark R. Looney, 513 Parnassus Ave, HSE 1355A, San Francisco, CA 94143-0130; e-mail: mark.looney@ucsf.edu.