The ATP-gated P2X1 ion channel contributes to the severity of antibody-mediated Transfusion-Related Acute Lung Injury in mice

The biological responses that control the development of Transfusion-Related Acute Lung Injury (TRALI), a serious post-transfusion respiratory syndrome, still need to be clarified. Since extracellular nucleotides and their P2 receptors participate in inflammatory processes as well as in cellular responses to stress, we investigated the role of the ATP-gated P2X1 cation channel in antibody-mediated TRALI. The effects of NF449, a selective P2X1 receptor (P2RX1) antagonist, were analyzed in a mouse two-hit model of TRALI. Mice were primed with lipopolysaccharide (LPS) and 24 h later challenged by administrating an anti-MHC I antibody. The selective P2RX1 antagonist NF449 was administrated before the administration of LPS and/or the anti-MHC I antibody. When given before antibody administration, NF449 improved survival while maximal protection was achieved when NF449 was also administrated before the sensitization step. Under this later condition, protein contents in bronchoalveolar lavages were dramatically reduced. Cell depletion experiments indicated that monocytes/macrophages, but not neutrophils, contribute to this effect. In addition, the reduced lung periarteriolar interstitial edemas in NF449-treated mice suggested that P2RX1 from arteriolar smooth muscle cells could represent a target of NF449. Accordingly, inhibition of TRPC6, another cation channel expressed by smooth muscle cells, also reduced TRALI-associated pulmonary interstitial and alveolar edemas. These data strongly suggest that cation channels like P2RX1 or TRPC6 participate to TRALI pathological responses.

contributions of the different blood populations. Cell depletion and/or transfer experiments have indicated that neutrophils are either essential [7][8][9] or dispensable 10 for lung edema formation. Other cells participate to TRALI responses like monocytes and/or macrophages 10,11 , while they control neutrophil recruitment in the lungs through MIP2 secretion 11 . TRALI develops similarly in wild type and rag2-deficient mice, indicating that lymphocytes do not have a major impact 9,10 . In contrast, in another mouse model, suppressor T cells or Tregs have been reported to inhibit TRALI through IL-10-dependent pathway(s) 7,12 . In addition, platelets play a critical role 9 or are dispensable 13 for the early TRALI-associated responses that lead to lung edema formation.
During acute lung injury, a plethora of stimuli can induce the release of the damage associated molecular pattern adenosine 5′-triphosphate (ATP) from various cell types such as endothelial and immune cells, platelets and/ or stressed erythrocytes. Two classes of membrane receptors mediate the effects of ATP, the ligand-gated P2X cation channels and the G protein-coupled P2Y receptors 14,15 . They are expressed in various combinations depending on the cell type, notably by cells participating in vascular homeostasis [15][16][17][18] . ATP receptors positively regulate the recruitment and activation of inflammatory cells 19 or control vascular parameters such as endothelial barrier integrity and hemodynamics 17 . Among these receptors, the P2X1 receptor (P2RX 1 ) is expressed on numerous cell types involved in vascular homeostasis and/or immunity, namely arterial smooth muscle cells (SMCs), neutrophils, macrophages and platelets and therefore might control inflammatory processes involved in the pathogenesis of TRALI. We investigated whether this receptor influences the pathogenesis of TRALI and if so, which P2RX 1 + cells could be involved and could potentially represent a biological target relevant to TRALI-associated pathologic responses.
Mice. Experiments were performed using 8 to 10 week-old male BALB/cByJ mice (JAX collection) purchased from Charles River Laboratories and used within one week after arrival. The mice were housed in individual ventilated cages of the animal facilities of the EFS-Grand Est (agreement No. E67-482-10). Ethical approval for the animal experiments was obtained from the French Ministry of Research, in accordance with the guidelines of the Regional Committee for Ethics in Animal Experimentation of Strasbourg (CREMEAS, CEEA-35).
Two-event TRALI model. A previously described two-hit model was used 9 with a few modifications. Male BALB/cByJ mice (8 to 10 weeks old) were primed with an intraperitoneal injection of LPS (0.1 mg/kg) 24 h before being anesthetized (xylazine (20 mg/kg body weight) and ketamine (100 mg/kg body weight)) and challenged with an anti-H-2K d mAb (clone 34-1-2S, 0.5 mg/kg), its isotype control or saline, injected retro-orbitally. The mice were maintained at 37 °C and sacrificed 10 min or 2 h after injection of the anti-MHC I mAb. All experiments were performed in a blinded manner.
Drug administration and specific cell depletion. NF449 (10 mg/kg in saline) was delivered i.v. 5 min prior to LPS sensitization and again 5 min before anti-MHC I mAb injection. In some experiments, the mice received NF449 either before LPS or before the anti-MHC I mAb. SAR7334 (10 mg/kg in saline) was injected retro-orbitally into isoflurane (Vetflurane) anesthetized animals, 30 min before TRALI experiments. Anti-Ly6G antibody (clone 1A8, 5 mg/kg in saline) was delivered i.v. 24 h prior to LPS. Anti-CCR2 antibody (clone MC-21, 0.4 mg/kg in saline) was injected i.p. and clodronate liposomes (2 mL/kg) were injected i.v., 18 h after LPS administration, i.e., 6 h prior to injection of the anti-MHC I mAb.
Leukocyte counts were determined in whole blood collected into EDTA (6 mmol/L) by severing the tail of anesthetized mice, using a Scil Vet ABC automatic cell counter (Scil Animal Care Company) set to murine parameters. The percentages of peripheral neutrophils and inflammatory monocytes were determined as described in Flow cytometry paragraph, the absolute counts of the cells were deduced. Partial oxygen pressure (pO 2 ). The partial oxygen pressure (pO 2 ) was measured in blood drawn into a syringe containing heparin anticoagulant (50 U/mL) from the abdominal aortas of anesthetized mice, 10 min after anti-MHC I injection, using a Critical Care Analyser (OMNIS S; Roche Diagnostics).
Protein content of bronchoalveolar lavages (BALs). At 10 min or 2 h after induction of TRALI, a ligation was performed to isolate the left lung (used later for immunohistochemistry), the tracheas of sacrificed mice were exposed and cannulated and 1 mL of sterile PBS solution was injected into the right lung. The bronchoalveolar lavages (BALs) were centrifuged at 13,000 × g for 5 min at 4 °C and the supernatants were stored at −80 °C. The protein content of BALs was quantified using a bicinchoninic acid assay (Thermo Fisher Scientific) according to the manufacturer's instructions with bovine serum albumin (BSA) as the standard.
Immunohistochemistry. At the end of the experiments, the left lungs of the mice were removed and fixed in 4% paraformaldehyde overnight. The fixed samples were washed in PBS and maintained in 10% and then 20% sucrose solutions, each time for 24 h, before being snap-frozen in optimal cutting temperature (OCT) embedding medium. Sections (8 μm) were stained with hematoxylin and eosin before histological examination. The areas www.nature.com/scientificreports www.nature.com/scientificreports/ occupied by edema and by the arterioles were quantified using Image J software on images acquired with a Leica DMLB microscope and the ratio was calculated. The investigator was blinded to the various experimental groups.
Flow cytometry. Briefly, after TRALI experiments, lungs were perfused with PBS via the right ventricle and an incisure was made in the vena cava as an outlet. Lungs were removed en bloc and minced with a scalpel. Minced lungs were placed in PBS with 500 U/mL type IV collagenase and 0.02 mg/mL DNase I and were incubated at 37 °C for 30 min. Digested lungs were passed through a 40 μm filter to obtain single cell suspension. After incubation, cells were counted (ADAM Automated Cell Counter, Digital Bio), resuspended in PBS-1% BSA and Fcγ receptors were blocked with FcR blocking reagent (Miltenyi Biotec). Cells were then stained with directly conjugated anti-CD45, -CD11b, -CD11c, -CD24, -Gr1, -CD170, -F4/80 and -MHC class II mAbs to determine the percentages of neutrophils and inflammatory monocytes, respectively. Cell depletions were controlled by flow cytometric analysis of blood cells stained with directly conjugated anti-CD45, -CD115 and -CD11b mAbs, or with anti-CD45 and anti-Ly6C mAbs. Flow cytometric data were acquired on a Fortessa-X20 flow cytometer (BD Biosciences) and the cell populations were analyzed using BD FACSDiva software. All subsets were gated using fluorescence minus one (FMO) analysis.

Statistical analyses.
Statistical analyses were performed with GraphPad software (Prism 5.02). Survival studies were analyzed using the log-rank test (Figs 1D, 2A,B and 3B) or the χ2 test (Fig. 2B). Other data are reported as the mean ± SEM and were analyzed using 1-or 2-way ANOVA followed by a Dunn's test or a Bonferroni post hoc analysis as appropriate. A P-value of <0.05 was considered to be statistically significant.

Results
The P2RX 1 receptor antagonist NF449 decreases the severity of experimental TRALI. To explore the role of P2RX 1 , we used an experimental TRALI model consisting of administration of the anti-MHC I mAb 34-1-2S (anti-H-2K d , -H-2D d , 0.5 mg/kg) in LPS-sensitized H-2 d BALB/c mice, taking advantage of the availability of the selective P2RX 1 antagonist NF449 21,22 . NF449 displays good selectivity for the P2RX 1 with respect to other P2 receptor subtypes and potently antagonizes (IC 50 = 5.8 ± 2.2 µM) the P2RX 1 -mediated calcium influx induced by α,β-methyleneadenosine 5′-triphosphate (α,βMeATP) in murine platelets [22][23][24] . BALB/c mice were treated with NF449 (10 mg/kg, i.v.) or vehicle (saline) prior to LPS injection (0.1 mg/kg, i.p.) and anti-MHC I antibody challenge (0.5 mg/kg, i.v.). Ten minutes later, the features of experimental TRALI, such as lung edema and the drop in blood oxygenation, can be measured whereas survival is evaluated within a 2 h period. Lung edema is evaluated by quantifying the proteins in bronchoalveolar lavages and respiratory distress by the measure of the partial oxygen pressure (pO 2 ) in aortic blood as previously described 13 . After 10 min, the protein content of BALs was significantly lower in NF449-treated than in saline-treated animals (Fig. 1A, left  panel). Accordingly, histological analyses of lung tissues revealed larger areas of interstitial periarteriolar edema in saline-treated mice as compared to those receiving NF449 (Fig. 1B,C), while no edema was observed in the proximity of the post-alveolar microcirculation ( Supplementary Fig. 1). Nevertheless, pO 2 dropped similarly in both groups (Fig. 1A, right panel).
Within the two hours following antibody injection, 56% of the control mice (treated with saline) died, while this percentage fell to 16% in mice receiving NF449 (Fig. 1D, left panel). The protein content of BALs was significantly increased in saline-treated mice challenged with the anti-MHC I antibody as compared to non-challenged mice. On the other hand, treatment with NF449 reduced the protein concentration in BALs by more than two-fold (Fig. 1D, right panel). Altogether, these results indicated that in mice, NF449 reduces the disruption of the endothelial barrier provoked by antibody-mediated TRALI, which allows better survival. www.nature.com/scientificreports www.nature.com/scientificreports/ NF449 needs to be administered before each step of experimental TRALI to provide maximal protection. We then investigated whether NF449 needed to be administered at each step of the two-hit experimental TRALI model, i.e., before LPS sensitization and/or before anti-MHC I mAb challenge, to reduce its pathology. The survival of the animals that received saline or NF449 prior to LPS sensitization only was similar ( Fig. 2A). Therefore, administration of NF449 exclusively before LPS sensitization did not improve survival.
In contrast, when NF449 was delivered only before the anti-MHC I antibody challenge, 33% of the mice died as compared to 60% of saline-receiving animals (Fig. 2B). Interestingly, whereas the whole survival curves of saline-and NF449-treated mice revealed no significant differences, survival at the 120 min endpoint was significantly improved in those receiving NF449 (n = 32, *p = 0.0384, χ2 test). Thus, treatment with NF449, before antibody administration only, improved survival, while maximal reduction of the severity of experimental TRALI required administration of NF449 also before the sensitization step.
The protective effect of NF449 does not rely on neutrophils and inflammatory monocytes. Subsequently, we evaluated the contribution of P2RX 1 + immunological cells to TRALI, namely neutrophils, monocytes and macrophages. Neutrophils can be selectively depleted for at least three days by treatment with the anti-Ly6G mAb 1A8 25 . This antibody was administered 24 h before LPS sensitization and flow cytometric analyses performed 48 h later, before anti-MHC I injection, confirmed their depletion (Supplementary Fig. 2A). The pathology of TRALI at the 10 min time point was similar in 1A8-treated and untreated mice, as assessed by the measure of the protein concentration in BALs and the pO 2 (Fig. 3A). Immunohistology confirmed that neutrophils could not be detected in the lungs of 1A8-treated mice after TRALI, indicating that even residual circulating neutrophils had not been recruited ( Supplementary Fig. 2B).
We next tested whether NF449 still has a protective effect on survival in neutropenic animals over a 2 h time course. In the absence of NF449 treatment, neutrophil depletion did not modify the severity of antibody-mediated TRALI in terms of survival at the 2 h endpoint (Fig. 3B, left panel). In contrast, 80% of mice treated with NF449 survived regardless of the neutrophil count (Fig. 3B, left panel). Accordingly, the protein content of BALs was similarly elevated after TRALI in neutropenic and non-neutropenic mice that did not receive NF449, but significantly decreased in NF449-treated animals (Fig. 3B, right panel). These results indicated that neutrophils do not play a major role in experimental TRALI and hence that other cell types must be responsible for the contribution of P2RX 1 to TRALI.
We then used clodronate liposomes, which eliminate all types of monocytes and also macrophages. To allow inflammation to proceed, the liposomes were delivered 18 h after LPS, i.e., 6 h before anti-MHC I mAb injection. Flow cytometric analyses confirmed that this procedure depleted tissue macrophages and peripheral monocytes ( Supplementary Fig. 3A,B). Under these conditions, antibody-mediated TRALI was totally abolished. Indeed, 10 min after anti-MHC I administration, the protein content of BALs was dramatically reduced in clodronate-treated mice as compared to control animals (Fig. 4A, left panel) and no interstitial edema was observed on histological preparations (Fig. 4A, right panel).
We then attempted to evaluate the specific contribution of monocytes. Monocytes can be divided into two main subtypes, the patrolling and the inflammatory monocytes. The former weakly express Ly6C and are CCR2 − , whereas the latter are Ly6C + and CCR2 + and represent the major subclass of blood monocytes (80% of monocytes in BALB/c mice). This pattern allows the depletion of inflammatory monocytes, whereas there is no available treatment to specifically deplete patrolling monocytes in BALB/c mice. Administration of the anti-CCR2 mAb MC-21, 18 h after LPS treatment, i.e., after the maximum concentration of inflammatory cytokines in the blood had been reached 26 , resulted in the depletion of inflammatory monocytes ( Supplementary Fig. 2C), in accordance with previous studies 20 . This cell specific depletion did not modify the development of TRALI triggered 6 h later, www.nature.com/scientificreports www.nature.com/scientificreports/ as assessed by measurement of the protein content of BALs (Fig. 4B, left panel) and the aortic pO 2 (Fig. 4B, right  panel). We could therefore conclude that inflammatory monocytes are dispensable for the initiation of experimental antibody-mediated TRALI.
Altogether, these findings strongly indicate that activation of patrolling monocytes and/or macrophages triggers the pathology of TRALI and that their P2RX 1 could contribute to TRALI.

The protective effect of NF449 is consistent with a contribution from P2RX 1 expressed on lung periarteriolar smooth muscle cells (SMCs). The development of pulmonary periarteriolar edemas dur-
ing TRALI early biological responses might be driven in part by the pulmonary arterial pressure. The blood pressure in the lung pre-capillary arterioles is controlled by SMCs. Contraction of these cells is positively regulated by P2RX 1 present on their surface 27 and their deficiency prevents the development of pulmonary arterial hypertension 27 . Thus, the reduced edema areas in NF449-treated mice could be due to the inhibition of periarteriolar SMC contraction. Interestingly, inhibition of another cation channel, the TRPC6 calcium channel, which is expressed by pulmonary SMCs, also inhibits pulmonary arterial hypertension and the associated edema during acute lung injury 28 .
We therefore tested the effects on TRALI of SAR7334, a specific inhibitor of TRPC6. SAR7334 was administered i.v. 30 min before the anti-MHC I mAb and TRALI responses were analyzed 10 min later. Remarkably, the quantity of proteins in BALs was reduced in animals receiving the inhibitor (Fig. 5). Histological  www.nature.com/scientificreports www.nature.com/scientificreports/ analyses of the lungs revealed that the extent of periarteriolar edema was diminished after SAR7334 treatment ( Supplementary Fig. 4). Thus, inhibition of the TRPC6 calcium channel expressed by SMCs reduced the formation of TRALI-associated lung edema. Combined administration of NF449 and SAR7334 did not however further inhibit TRALI-associated lung edema (Fig. 5). These findings strongly suggest that the P2RX 1 present on lung periarteriolar SMCs could also contribute to the protective effect of NF449.

Discussion
Our present study shows that NF449 reduces the severity of TRALI in mice, indicating a major contribution of the P2X 1 receptor. Maximal reduction of the experimental TRALI features required treatment with NF449 before the sensitization step with LPS and before administration of the anti-MHC I mAb. Of note, NF449 treatment before injection of the anti-MHC I mAb only, also resulted in a significant improvement in the survival of the animals. P2RX 1 appeared to control different mechanisms at each step of the two-hit model. LPS responsive cells, namely monocytes or macrophages, which express P2RX 1 29 , were most likely the primary targets of NF449 in the first hit, although the role of P2RX 1 on these cells during LPS sensitization remains to be clarified. In the second step, the protective effect of NF449 could be related to all cell types expressing the P2RX 1 and participating in vascular homeostasis. Endothelial cells (ECs) and platelets were not considered, since ECs do not express P2RX 1 30 , while platelets are dispensable for the initiation of TRALI 13 . We first explored the role of neutrophils since P2RX 1 has been shown to control neutrophil responses by amplifying chemotactic signals 31,32 . Here, their immunodepletion did not modify the features of experimental TRALI at the 10 min or 2 h time points.
This absence of a major role of neutrophils may appear to be in conflict with the accepted view of their importance in TRALI. However, in the literature, part of the evidence supporting the contribution of neutrophils to TRALI responses is indirect. Firstly, in man, post mortem histological analyses have shown that TRALI is accompanied by the recruitment of neutrophils to the lungs, but these observations cannot indicate whether this neutrophil recruitment is the result or the cause of TRALI initiation 33 . Anti-HLA class I antibodies could nevertheless induce TRALI in a neutropenic receiver (less than 0.1 10 9 cells/L) 34 . Secondly, in animal models, neutrophils are ideal candidate cells based on their recruitment to the lungs and their ability to produce damaging reactive oxygen species in vitro, while conditions inhibiting this production protect the animals from developing TRALI 11,[35][36][37][38][39] .
Direct evidence for the requirement for neutrophils in antibody-mediated TRALI exists, but can be debated. It is based on in vivo studies using transfer of isolated bone marrow neutrophils 8 , which may not be representative of circulating neutrophils since it is technically difficult to isolate these cells in a resting state, or on cell depletion in animals followed by intratracheal instead of intraperitoneal administration of LPS 9 , which produces fundamentally different inflammatory conditions. Indeed, intratracheal administration of LPS leads to neutrophil recruitment to the alveoli, whereas intraperitoneal injection of low doses of LPS, as in our work, does not. A second team reported that neutrophil depletion blocked lung edema formation 7 , while a third failed to observe any protective effect 10 . It is difficult to compare our observations with those of these two groups, since the experimental conditions were very different: (i) LPS sensitization in our study versus no sensitization in the two others (two-hit versus one-hit model), (ii) different doses of anti-MHC I antibody (0.6 mg/kg in our study versus 5 to 50 mg/kg), (iii) variable breeding conditions (breeder or animal facility environment) and (iv) different genetic backgrounds (BALB/cByJ, BALB/cAnNCrl, BALB/cAnNTac). Notably, with respect to the last two points, BALB/ cByJ and BALB/cAnNCrl mice have been reported to display a different susceptibility to Theiler's murine encephalomyelitis virus-induced demyelinating disease 40 , and the commensal microbiota may affect TRALI responses 41 . Nonetheless, we do not exclude that neutrophils could participate in later TRALI responses such as extravasation, which requires P2RX 1 activity 32 .
We showed that clodronate liposomes, which eliminate monocytes, macrophages and dendritic cells (DCs), protected mice from TRALI. On the other hand, immunodepletion of CCR2 + inflammatory monocytes did not influence the development of TRALI, indicating that inflammatory monocytes are not key players in this syndrome. How and which of these cells contribute positively to TRALI remains to be elucidated. Distal Kupffer cells or splenic macrophages could participate in TRALI, as observed in other ALI models 42 , or more likely arteriolar macrophages 43 or dendritic cells 44 , given the localization of the interstitial lung edema. Administration of diphtheria toxin to transgenic animals expressing its receptor on CD11c + cells, which include DCs, exacerbates TRALI-associated lung damage 7 . However, this experimental set-up can lead to neutrophilia 45 , which could represent a confounding effect in this neutrophil-dependent TRALI model. Thus, complementary experiments will be required to clarify the role of DCs, at least in our experimental model.
Remarkably, interstitial lung edema was observed in the pre-microcapillary arteriolar spaces but not in the post-alveolar circulation. This localization is consistent with the observation that ALI begins with periarteriolar edema which, after exceeding a threshold, floods into the alveoli 46 . Accordingly, real time analyses of the development of lung edema in pigs 47 or in man have shown that the extravascular lung water volume increases before a defect in blood oxygenation is ascertained 48 . These facts strongly suggest that the arteriolar tissue is the primary site of the edematous process during anti-MHC I-induced TRALI. The reduced lung periarteriolar interstitial edemas in NF449-treated mice suggested that P2RX 1 + vascular SMCs could participate in the formation of interstitial lung edema by promoting arteriolar vasoconstriction and thereby a transient increase in pulmonary blood pressure, which would favor plasma leakage 46 . Stimulation of these cells induces vasoconstriction in a P2RX 1 -dependent manner, resulting in pulmonary arterial hypertension, which is in part responsible for edema formation 27 . We note that in a rat model of TRALI, transfusion of the plasma from aged erythrocyte concentrates was accompanied by an increase in pulmonary arterial blood pressure 6 . Accordingly, in two cases of human TRALI observed after the transfusion of blood products containing anti-HLA class I antibodies, a rise in pulmonary arterial tension was detected before the onset of hypoxemia and acute respiratory distress 49,50 . TRPC6 is another calcium channel expressed by SMCs, which positively controls vasoconstriction, pulmonary arterial hypertension and edema formation 28,51 . Remarkably, we showed that the TRPC6 inhibitor SAR7334 also reduced www.nature.com/scientificreports www.nature.com/scientificreports/ the accumulation of proteins in BALs and the extent of the periarteriolar edema provoked by TRALI, which could represent the first example of a protective effect in vivo of a specific TRPC6 inhibitor 28 . The expression profiles of TRPC6 in the hematopoietic lineage indicate that murine monocytes and macrophages do not significantly express this gene (heamosphere.org and immgen.org), while TRPC6 is expressed by ECs. Therefore, we cannot formally exclude that inhibition of endothelial TRPC6 also contributes to the protective influence of SAR7334.
Currently, in the case of TRALI, the mainstay of treatment remains general supportive care, including fluid resuscitation, oxygenation or mechanical ventilation 52 . Recent studies have indicated that the administration of IL-10 or CRP inhibitors could represent promising therapeutic or prophylactic strategies 53 . Our findings now suggest that targeting cation channels like P2RX 1 or TRPC6, by reducing interstitial lung edema, could constitute an additional pharmacological approach for TRALI management, given that intensive medical care lasts several hours.