Leukotriene B4 receptor type 2 protects against pneumolysin-dependent acute lung injury

Although pneumococcal infection is a serious problem worldwide and has a high mortality rate, the molecular mechanisms underlying the lethality caused by pneumococcus remain elusive. Here, we show that BLT2, a G protein-coupled receptor for leukotriene B4 and 12(S)-hydroxyheptadecatrienoic acid (12-HHT), protects mice from lung injury caused by a pneumococcal toxin, pneumolysin (PLY). Intratracheal injection of PLY caused lethal acute lung injury (ALI) in BLT2-deficient mice, with evident vascular leakage and bronchoconstriction. Large amounts of cysteinyl leukotrienes (cysLTs), classically known as a slow reactive substance of anaphylaxis, were detected in PLY-treated lungs. PLY-dependent vascular leakage, bronchoconstriction, and death were markedly ameliorated by treatment with a CysLT1 receptor antagonist. Upon stimulation by PLY, mast cells produced cysLTs that activated CysLT1 expressed in vascular endothelial cells and bronchial smooth muscle cells, leading to lethal vascular leakage and bronchoconstriction. Treatment of mice with aspirin or loxoprofen inhibited the production of 12-HHT and increased the sensitivity toward PLY, which was also ameliorated by the CysLT1 antagonist. Thus, the present study identifies the molecular mechanism underlying PLY-dependent ALI and suggests the possible use of CysLT1 antagonists as a therapeutic tool to protect against ALI caused by pneumococcal infection.

Scientific RepoRts | 6:34560 | DOI: 10.1038/srep34560 (AA) by 5-lipoxygenase (5-LO) and LTA 4 hydrolase 11 ; however, we later found that another AA metabolite 12(S)-hydroxyheptadeca-5Z,8E,10E-trienoic acid , which is produced downstream of cyclooxygenases (COXs), is the high-affinity endogenous ligand of BLT2 12 . We previously showed that BLT2 is expressed in intestinal epithelial cells and skin keratinocytes, and that 12-HHT/BLT2 signaling maintains intestinal barrier function and accelerates skin wound healing 13,14 . Recently, we found that BLT2 expressed in keratinocytes enhances the expression of several tight junctional proteins, including claudin 4, and protects against invasion by foreign antigens 15 . Taken together, these data suggest a crucial role for BLT2 in epithelial barrier homeostasis.
Based on the finding that BLT2 is expressed in pulmonary epithelial cells and vascular endothelial cells in the mouse lung, we hypothesized that BLT2 may play a protective role in lung epithelia. Thus, we treated mice with PLY to induce ALI and explored the physiological role of BLT2 in the lung.

Results
BLT2-knockout mice are highly susceptible to acute lung damage upon intratracheal administration of PLY. To investigate the physiological function of BLT2 in the lung, BLT2-knockout (BLT2-KO) and the littermate control wild-type mice (BLT2-WT) on BALB/c (Fig. 1A) and C57BL/6 ( Fig. 1B) backgrounds were intratracheally injected with 50 ng of recombinant PLY. Surprisingly, about 80% of BLT2-KO mice on both backgrounds died within 10 to 30 min after PLY administration, in clear contrast to BLT2-WT mice (Fig. 1A,B). As BLT1 16 and BLT2 are highly homologous, we repeated the experiment using BLT1-knockout (BLT1-KO) and the littermate control wild-type (BLT1-WT) mice and found that BLT1-KO mice were not susceptible to the effects of intratracheal PLY administration (Fig. 1C). Furthermore, intravenous injection of PLY did not cause death at all (Fig. 1D), showing that systemically administrated PLY is not lethal by itself; rather, it causes topical injury and inflammation upon local administration in the lungs.
Thus, we next examined PLY-induced acute lung damage by measuring the concentrations of total protein and albumin, along with LDH activity, in bronchoalveolar lavage (BAL) fluid. All of these parameters were increased after PLY administration, even in BLT2-WT mice; however, values were significantly higher in BLT2-KO mice, which showed more severe lung damage ( Fig. 1E-G). Taken together, these results suggest that BLT2-KO mice are more susceptible to lung damage caused by intratracheal PLY administration, regardless of their genetic background.

BLT2 is expressed in alveolar epithelial type II and vascular endothelial cells in the mouse lung.
BLT2 is mainly expressed in tissues exposed to the external environment, such as skin and intestine 17 . We previously reported that BLT2 protects mouse intestinal epithelial cells against colitis 14 , and that BLT2 expression in keratinocytes is important for accelerating skin wound healing 13 . Because no study has examined expression of BLT2 in the lung, we next investigated whether BLT2 is expressed in mouse lung by detecting BLT2 mRNA by quantitative PCR (Fig. 2A). To identify BLT2-expressing cells in mouse lung, we performed immunohistochemical analysis of serially sectioned lung tissues obtained from BLT2-WT and BLT2-KO mice. BLT2 signals were detected in proSP-C-positive alveolar epithelial type II (AT II) cells, but not in T1α -positive alveolar epithelial type I (AT I) cells (Fig. 2B). BLT2 was also expressed in CD31-positive lung vascular endothelial cells (Fig. 2C). No BLT2 signals were detected in AT II cells and vascular endothelial cells in BLT2-KO mice, confirming the specificity of the anti-mouse BLT2 antibody.

BLT2 deficiency augments PLY-induced vascular hyperpermeability and bronchoconstriction.
We next investigated the mechanisms underlying the PLY-induced death of BLT2-KO mice with respect to the function of AT II cells and vascular endothelial cells. One of the important functions of AT II cells is to produce and secrete pulmonary surfactant; therefore, we examined whether BLT2 deficiency affects surfactant production. To do this, we measured the expression of mRNAs for surfactant proteins in isolated AT II cells, and the concentration of phosphatidylcholine (PC), the major surfactant lipid, in BAL fluid from BLT2-WT and BLT2-KO mice 18 . The levels of surfactant protein A, B and C mRNA (Sftpa, Sftpb, and Sftpc, respectively) in isolated AT II cells were not affected by BLT2 deficiency (Fig. 3A). The number of AT II cells was also similar in BLT2-WT and BLT2-KO mice (data not shown). The concentration of PC in BAL fluid from BLT2-WT and BLT2-KO mice was also similar (Fig. 3B). These results indicate that BLT2 deficiency has no effect on AT II cell function, at least in terms of surfactant production.
Because a previous study reported PLY-induced lung vascular leakage during the early phase of infection 19 , we next examined endothelial function in BLT2-WT and BLT2-KO mice. Intratracheal administration of PLY caused leakage of Evans blue from vasculature in both BLT2-WT and BLT2-KO mice, but it was more obvious in BLT2-KO mice (Fig. 3C). Quantification of the leaked dye clearly showed that PLY-dependent vascular leakage was significantly higher in BLT2-KO mice than in BLT2-WT mice (Fig. 3D). Examination of hematoxylin and eosin (H&E)-stained lung sections showed that, in addition to vascular leakage, BLT2-KO mice developed severe bronchoconstriction upon PLY administration (Fig. 3E). However, PLY did not induce bronchoconstriction in BLT2-WT mice (Fig. 3E). Thus, we next assessed pulmonary function using the flexiVent. We found that both airway resistance and elastance in BLT2-KO mice were higher after PLY administration; these effects were time-dependent and were not observed in BLT2-WT mice (Fig. 3F,G). Airway resistance is a quantitative measure of airflow limitation, and elastance reflects the rigidity of the lung tissue. The observed increase in both parameters in BLT2-KO mice might account for the bronchoconstriction and the vascular hyperpermeability induced by PLY. Taken together, these data suggest that BLT2 deficiency causes severe vascular leakage and bronchoconstriction, but has no effect on the production of pulmonary surfactant.
PLY triggers the production of cysteinyl leukotrienes. The vascular hyperpermeability and bronchoconstriction observed in BLT2-KO mice after PLY administration resemble a lethal asthma attack in a human Scientific RepoRts | 6:34560 | DOI: 10.1038/srep34560 patient; thus, we hypothesized that these characteristics may be linked to bronchoconstrictors such as cysteinyl leukotrienes (cysLTs; leukotriene C 4 (LTC 4 ), leukotriene D 4 (LTD 4 ) and leukotriene E 4 (LTE 4 )), collectively known as slow reactive substance of anaphylaxis (SRS-A) 20 . Thus, we next measured the eicosanoid content of the BAL fluid and lung tissue from PLY-treated BLT2-WT and BLT2-KO mice using LC-MS/MS (Fig. 4A,B and Table 1) 21 . Surprisingly, PLY administration led to the production of large amounts of cysLTs in the BAL fluid and lung tissue, which were almost undetectable in vehicle-treated mice (Fig. 4A,B). PLY also increased the levels of other eicosanoids (LTB 4 , 12-HETE) in BAL fluid and lung tissue (Table 1), although there were present at much lower levels than cysLTs. Basal levels of prostaglandin (PG) and 12-HHT were detectable, but were not affected by PLY (Table 1). PLY treatment increased the concentration of cysLTs in diluted BAL fluid to about 4 nM, suggesting that high concentrations of cysLTs are produced in the lung after PLY administration, followed by subsequent activation of three receptors for cysLTs: CysLT1, CysLT2, and GPR99 22 . However, BLT2 deficiency did not affect PLY-induced cysLTs production.
We next identified the cells responsible for PLY-dependent cysLTs production. LTC 4 is produced from AA through 5-LO and LTC 4 synthase, and is subsequently converted into LTD 4 and then LTE 4 extracellularly 23 . The most important and well-known source of LTC 4 is mast cells; therefore, we treated bone marrow-derived mast cells (BMMCs) with PLY and measured the cysLTs concentrations in the culture supernatant. Extremely high amounts of LTC 4 (~3 nmol/1 × 10 6 BMMCs) were detected after PLY treatment (Fig. 4C), indicating that PLY stimulates both the production and release of cysLTs from mast cells, leading to vascular hyperpermeability and bronchoconstriction in the mouse lung. Previous reports demonstrate that BLT2 is expressed by mast cells 12,24 .
To compare the functions of mast cells in BLT2-WT and BLT2-KO mouse lungs, we measured the expression of mRNA for mast cell markers, including Hdc (histidine decarboxylase), Mcpt4 (chymase), Mcpt6 (tryptase), Cpa3 (carboxypeptidase), and Kit (stem cell factor receptor) 25 . None of these mast cell markers was affected by BLT2 deficiency (Fig. 4D), suggesting that mast cell maturation is intact in BLT2-KO mice. These data suggest that PLY stimulates the production of large amounts of cysLTs by mast cells in both BLT2-WT and BLT2-KO mice.

12-HHT/BLT2 signaling suppresses CysLT1 expression in vascular endothelial cells. Because
PLY-induced cysLTs production was comparable between BLT2-WT and BLT2-KO mice (Fig. 4A,B), we next examined the expression level of cysLTs receptor CysLT1. CD31-positive mouse lung endothelial cells (MLEC) were isolated and subjected to quantitative PCR. Expression of BLT2 mRNA was observed both in MLEC and CD31-negative cells from BLT2-WT mouse lung (Fig. 5A, Supplementary Fig. S1), consistent with the immunohistochemical analysis (Fig. 2). CysLT1 mRNA levels in MLEC were significantly higher in BLT2-KO than in BLT2-WT (Fig. 5B), though those in whole lung (data not shown) and in CD31-negative lung cells (Fig. 5B) from BLT2-WT and BLT2-KO mice were similar. Because measurement of Ca 2+ mobilization or transepithelial electrical resistance (TER) with isolated primary MLEC was unsuccessful, we employed HUVEC, which endogenously express both BLT2 (Fig. 5C) and CysLT1 (Fig. 5D) for further analysis. The blockade of 12-HHT/BLT2 signaling by the BLT2 antagonist LY255283 induced CysLT1 mRNA expression (Fig. 5E) and decreased TER after LTD 4 stimulation (Fig. 5F) in HUVEC. These data suggest that BLT2 deficiency enhances CysLT1 expression in vascular endothelial cells, resulting in the enhanced susceptibility to cysLTs produced by PLY administration in BLT2-KO mice.
A CysLT1 antagonist improves PLY-induced ALI in BLT2 deficient mice. CysLT1 is one of the receptors for cysLTs and is deeply involved in asthmatic responses such as bronchoconstriction and vascular leakage; thus, several CysLT1 antagonists are clinically used to treat bronchial asthma and allergic rhinitis [26][27][28] . To examine the association between cysLTs and PLY-dependent lethality, mice were treated with a CysLT1-selective antagonist, montelukast, by gavage at 24 and 4 h before intratracheal administration of PLY. The survival of montelukast-pretreated BLT2-KO mice was markedly higher than that of vehicle control mice (Fig. 6A). Of note, PLY-induced vascular leakage was completely abrogated in both BLT2-WT and BLT2-KO mice by pretreatment with montelukast (Fig. 6B,C). The PLY-dependent increases in airway resistance and elastance in BLT2-KO mice were also entirely ameliorated by pretreatment with montelukast (Fig. 6D,E). These results clearly indicate that CysLT1 is involved in PLY-induced ALI in BLT2-KO mice. Finally we asked whether PLY-dependent mortality is affected by treatment of mice with NSAIDs which inhibit the production of the BLT2 ligand 12-HHT. As expected, pretreatment of WT mice with aspirin and loxoprofen increased mortality by PLY administration (Fig. 7A,B, respectively). Moreover, montelukast treatment also improved the survival ratio of aspirin-pretreated mice after PLY administration (Fig. 7C). These results emphasize that 12-HHT/BLT2 signaling plays important role in protecting against PLY-related ALI, possibly by decreasing CysLT1 expression.

Discussion
In this study, we explored the effects of BLT2-deficiency on the acute lung injury using BLT2-KO mice. BLT2 was originally cloned as a highly homologous receptor to BLT1, the receptor for a potent chemoattractant LTB 4 16 , with 45% on amino acid level, and its gene locus is overlapped on the promoter region of the BLT1 gene 11 . As extremely high amount of LTB 4 is required to activate BLT2, we explored the endogenous ligand and identified the 12-HHT as a high-affinity endogenous ligand for BLT2, which had long been assumed as a functionless by-product of thromboxane A 2 synthesis 12 . 12-HHT is massively produced via thromboxane A 2 synthase (TxA 2 S) in association with blood coagulation, but is also constantly produced in a TxA 2 S-independent fashion under the presence of COXs 21 . Previously we reported that 12-HHT/BLT2 axis plays biological roles in maintaining intestinal barrier function 14 and in accelerating skin wound healing 13 , and that BLT2 enhances the expression of tight junctional proteins 15 . Based on these evidences, we investigated whether BLT2 plays biophylactic ability in the lung, which is also exposed to outer environment, using a pneumococcal toxin PLY. Of particular interest is that, although BLT2 deficiency caused PLY-dependent sudden death, WT mice rarely died; this suggests that inhibiting BLT2 signaling critically exacerbates PLY-induced ALI. Although there is no report on BLT2 deficiency in human, several reports suggest the relationship between pneumococcal infection and NSAIDs in human. Recent reports show that NSAIDs given during the early stages of lower respiratory tract infections may aggravate serious pneumococcal pneumonia in human patients 29,30 . In addition, a murine model revealed the detrimental role of aspirin in experimental pneumococcal infection 31 . Because the BLT2 ligand 12-HHT is produced from AA via COXs, its production is completely inhibited by NSAIDs 13,21 . Consistent with these previous reports, pretreatment with aspirin or loxoprofen certainly aggravated the survival ratio of PLY-treated mice. These data explain the possible molecular mechanisms underlying the finding that patients taking NSAIDs exhibit more severe symptoms of pneumococcal pneumonia 29,30 . Furthermore, clinicians occasionally encounter ALI flare-ups during a course of antibiotics, possibly due to the release of large amounts of PLY from dead bacteria. Here, we also show that the CysLT1 antagonist blocks PLY-dependent ALI both in BLT2-KO mice and in NSAIDs-pretreated mice, suggesting that the CysLT1 antagonist is a candidate drug for treating severe ALI in patients with pneumococcal pneumonia.
Although the comparable levels of PLY-dependent cysLTs production were observed in BLT2-WT and BLT2-KO mice, the expression of CysLT1 mRNA in MLEC was enhanced in BLT2-KO mice. Consistently, pharmacological inhibition of BLT2 signaling by LY255283 increased the mRNA expression of CysLT1 and subsequently enhanced the LTD 4 -dependent barrier dysfunction in HUVEC. Recent reports show that CysLT1 signaling is negatively regulated by several GPCRs, including CysLT2 and GPR17 [32][33][34] . These receptors interact with CysLT1, thereby inhibiting LTD 4 -dependent signaling. Our findings suggested that BLT2 negatively regulates CysLT1 function in lung, by suppressing CysLT1 expression; hence, BLT2 deficiency augments cysLTs-CysLT1-dependent bronchoconstriction and vascular leakage, which in turn results in acute respiratory failure and sudden death in BLT2-KO mice after PLY administration.
Although the CysLT1 antagonist completely inhibited PLY-dependent vascular leakage and bronchoconstriction, it did not completely protect all BLT2-KO mice from death, indicating that alternative mechanisms of PLY-induced lethality may be in play. One possibility is that epithelial fragility may occur in BLT2-KO mouse lung; indeed, a recent report showed that BLT2 augments tight junctions 15 . BLT2 expressed in AT II cells played no role in surfactant production; however, the role of BLT2 in other AT II cell functions (e.g., as progenitors of AT I cells) is unknown. Another possibility would be the involvement of other lipid mediators. Recently, a growing number of reports describe a relationship between specialized pro-resolving lipid mediators (SPMs) and acute lung inflammation 35 . Although we failed to detect several SPMs (including lipoxin A 4 , protectin D1, and resolvins) in BAL fluid in the presence/absence of PLY (data not shown), other SPMs such as maresin 1 36 might be involved in lethal ALI.
Collectively, the data reported herein reveal a novel role for BLT2 in protecting lung tissues in addition to its previously reported role in the intestine and skin. Thus, BLT2 plays a crucial role in tissues exposed to the external environment, raising the possibility that a CysLT1 antagonist may protect patients with pneumococcal pneumonia from ALI. 100 mg/kg) and xylazine (SIGMA; 10 mg/kg), and oro-tracheally intubated with 20-gauge intravenous catheters (TERUMO). PLY (50 ng) or an equivalent volume of saline (50 μ L; as vehicle) were injected through the catheter and allowed to absorb into the alveoli by spontaneous respiration. Mice were then monitored for 72 h to evaluate survival. For the NSAIDs-pretreatment experiments, mice were started to receive aspirin (SIGMA; 0.18 mg/mL in drinking water) or loxoprofen (Tokyo Chemical Industry; 30 μ g/mL in drinking water) 2 days or 7 days before PLY injection as reported 13,39 . For experiments with the CysLT1 antagonist, mice were pretreated with montelukast (CAYMAN) as previously reported, with a minor modification 40 . In short, montelukast (5 mg/kg) was suspended in 1% methylcellulose and administered by gavage twice, at 1 d and 4 h prior to PLY treatment.
Analysis of BAL fluid. BAL was performed 10 min after PLY administration. One milliliter of PBS containing 2 mM EDTA was infused into the lungs through the intubated catheter, and 700 μ L of the solution was collected as BAL fluid. The total protein concentration, albumin concentration, and LDH activity in the BAL fluid were measured using Protein Assay Bicinchoninate kit (nacalai), Mouse Albumin ELISA Quantitation Set (Bethyl Laboratories), and LDH Cytotoxicity Detection Kit (Takara), respectively, according to the manufacturers' protocols. The concentrations of eicosanoids were measured by LC-MS/MS as described previously 21 . The PC concentration was measured in an enzyme-based fluorescent assay as previously reported 18 .
Vascular permeability assay. The permeability of the lung vasculature was measured according to the accumulation of intravenously injected Evans blue in the tissues as described previously 41,42 , with a minor modification. Briefly, mice were intravenously injected with 100 μ L of 0.5% Evans blue (nacalai) at 30 min prior to PLY administration. The lungs were then perfused with saline and harvested 10 min after PLY treatment, followed by snap freezing in liquid nitrogen. Lungs were then homogenized in 2 mL of PBS. Evans blue was extracted from the lung homogenate by adding twice the volume of formamide and incubating at 60 °C for 18 h. Samples were centrifuged at 12,000 × g for 30 min. The concentration of Evans blue in the supernatants was then measured at the absorbance of 620 and 740 nm using a dual wavelength spectrophotometric method. The concentration (as represented by the corrected A620 value that omits contaminating heme pigments) was calculated using the following formula: Corrected A620 = A620 − (1.426 × A740 + 0.03).
Quantitative RT-PCR. Mouse AT II cells were isolated as previously reported 43  H&E staining. BLT2-KO mouse lungs were harvested immediately after death. WT mice were sacrificed at the time of BLT2-KO mouse death caused by PLY. Lungs were collected, fixed in 10% formalin, paraffin-embedded, sectioned, and stained with H&E.
Mast cell stimulation. BMMCs were prepared in IL-3-containing medium, as described previously 25  TER measurement. HUVECs were purchased from Lonza and cultured in EGM ™ BulletKit ™ Medium (Lonza). TER was measured as previously reported 15,45 . Briefly, cells were seeded at 1 × 10 5 cells/well on fibronectin (25 μ g/mL)-coated membrane inserts (Millicell ® Cell Culture Inserts; Millipore) in culture medium, supplemented with 1 μ M LY255283 or 0.01% DMSO as vehicle control, and cultured for 3 days to achieve confluency. TER was measured before and 5 min after 1 μ M LTD 4 stimulation, using Millicell-ERS-2 volt-ohmmeter (Millipore). The value was calculated with the following formula: TER (Ωcm 2 ) = (R sample − R blank) × effective membrane area, and standardized by the basal TER value.
Statistics. Data are expressed as the mean ± SEM. Survival data were analyzed using the log-rank test. Two data sets were compared using a two-tailed Student's t test. ANOVA was used for multiple comparisons. P values < 0.05 were considered statistically significant. All data were calculated using Prism version 5.0 (GraphPad Software).