Leukotriene E₄–induced pulmonary inflammation is mediated by the P2Y₁₂ receptor

To determine if P2Y₁₂ receptors could mediate responses to LTE₄, we stably transfected Chinese hamster ovary (CHO) cells (which do not natively express classical CysLTRs [Maekawa et al., 2001] or P2Y₁₂ receptors) with constructs encoding the human P2Y₁₂ receptor protein in forward and reverse orientation. The transfectants were stimulated with exogenous LTC₄, LTD₄, or LTE₄. Fura-2 AM–loaded CHO cells expressing P2Y₁₂ receptors failed to flux calcium in response to LTE₄ but did exhibit a response to ADP, reflecting endogenous P2Y₁ receptors (Fig. 1 A). However, CHO cells transfected with the P2Y₁₂ receptor construct in forward (but not reverse) orientation responded to LTE₄ (and to LTD₄) with ERK activation (Fig. 1 B), which occurred at doses as low as 1 nM and reached a plateau at 100 nM (Fig. S1). The cys-LT–induced ERK activation was partially blocked by pretreatment of the cells with 2-methylthioadenosine monophosphate (2-MesAMP), a selective antagonist of the P2Y₁₂ receptor (Fig. 1 C), and was also sensitive to pertussis toxin (PTX; Fig. 1 D) but was resistant to MK571 (Fig. 1 D), an inhibitor which blocks CysLT₁R and some P2Y receptors (Mamedova et al., 2005). Although ERK activation in the P2Y₁₂ receptor transfectants was at least as robust in response to LTE₄ as to LTD₄ (Fig. 1 D), CysLT₁R and CysLT₂R transfectants reacted to LTD₄ in marked preference to LTE₄, even at high (500 nM) ligand concentrations. These responses were resistant to PTX and 2-MesAMP and, in the case of the CysLT₁R transfectants, susceptible to MK571.

We had previously reported that LAD2 cells (a well-differentiated human MC sarcoma line; Kirshenbaum et al., 2003) exhibited activation responses to LTE₄ > LTD₄. To determine whether P2Y₁₂ receptors were responsible for these responses, Fura-2 AM–loaded LAD2 cells were stimulated with LTD₄ or LTE₄ in the presence or absence of 2-MesAMP or MK571. As reported previously (Paruchuri et al., 2008), MK571 blocked calcium responses of LAD2 cells to both ligands (unpublished data). In contrast, 2-MesAMP treatment failed to attenuate LTD₄- or LTE₄-mediated calcium flux in LAD2 cells (Fig. 2 A). We determined the effect of MK571 and 2-MesAMP on MIP-1β generation by LAD2 cells in response to stimulation for 6 h with LTD₄ and LTE₄, using IgE plus anti-IgE as a positive control. LTE₄ at doses of 100 and 500 nM induced the generation of large amounts of MIP-1β, exceeding the amounts generated in response to LTD₄ and to IgE plus anti-IgE (Fig. 2 B). Pretreatment of the LAD2 cells with 2-MesAMP blocked the LTE₄-mediated increment in MIP-1β production by >50% (Fig. 2 B). 2-MesAMP also reduced the response to the higher concentrations of LTD₄. MK571 suppressed the response to both ligands and was additive with 2-MesAMP for the suppression of the response to LTD₄. Neither antagonist altered the production of MIP-1β in response to IgE plus anti-IgE. To exclude potential off-target effects of the inhibitors, samples of LAD2 cells were transfected with lentiviruses encoding a P2Y₁₂ sequence-specific shRNA, a CysLT₁R-specific shRNA, or an empty vector control before stimulation. Knockdown of P2Y₁₂ receptors decreased the receptor messenger RNA (mRNA) expression by ∼90% (Fig. 2 D) and did not alter expression of CysLT₁R or CysLT₂R proteins (not depicted). P2Y₁₂ receptor knockdown nearly abrogated MIP-1β production, completely eliminated PGD₂ production in response to LTE₄, and minimally affected the response to LTD₄. The knockdown of CysLT₁R completely blocked MIP-1β and PGD₂ production in response to LTD₄ and slightly (but not significantly) decreased the responses to LTE₄ (Fig. 2 C, bottom). Neither knockdown altered MIP-1β generation or PGD₂ production in response to IgE–anti-IgE (Fig. 2 C). LTE₄-mediated MIP-1β generation (Fig. S2 A) and ERK activation (Fig. S2 B) were not altered by treatment of LAD2 cells with apyrase to degrade extracellular ADP.

To determine whether P2Y₁₂ receptors directly mediated LTE₄ binding, membranes were prepared from LAD2 cells with and without P2Y₁₂ knockdown. We first performed competitive radioligand binding assays using [³H]ADP (the known natural ligand of P2Y₁₂ receptors) and unlabeled LTs as competitors. Unlabeled LTE₄ competed with labeled ADP, blocking 39 ± 9 and 50 ± 9% of specific ADP binding at doses of 0.1 and 1 nM, respectively, and reaching a plateau (60 ± 7%) at 1 µM (mean ± SEM for four separate experiments, as shown for one experiment [Fig. S3 A]). LTE₄ was more efficacious than LTD₄ (Fig. S3 B). Knockdown of P2Y₁₂ receptors reduced binding of [³H]ADP by 40–60% and completely eliminated competition by LTE₄ (Fig. S3, A and B, right). LAD2 cell membranes weakly bound [³H]LTE₄ (converted from commercially prepared [³H]LTD₄; Fig. S3 C), but specific binding of [³H]LTE₄ was not altered by the knockdown of P2Y₁₂ receptor (Fig. S3 C). To determine whether LTE₄ could block the ADP binding of P2Y₁₂ receptors expressed in isolation, the human forward and reverse P2Y₁₂ constructs were transiently expressed in COS-7 cells. The membranes from the transfectants expressing the forward construct bound [³³P]2-MesADP, a selective P2Y₁₂ receptor agonist. In contrast to ADP binding to LAD2 membranes, this binding was not competed by unlabeled LTE₄ or LTD₄ (Fig. S3 D). Additionally, these membranes failed to bind [³H]LTE₄ (unpublished data).

To determine whether LTE₄ could induce or potentiate histological signatures of pulmonary inflammation in mice, we determined the effect of 2.2 nmol of intranasal LTE₄ administered on three successive days on the lung histology of naive BALB/c mice. We compared the effect to that of LTD₄. The lungs of naive BALB/c mice showed no evidence of cellular influx or goblet cell metaplasia after three doses of either cys-LT alone (unpublished data). Thus, we sought to determine whether either cys-LT amplified pulmonary inflammation induced by the inhalation of low-dose allergen in sensitized mice. 2 wk after sensitization with chicken egg OVA by i.p. injection, BALB/c mice received inhalation challenges on three consecutive days with low-dose OVA (0.1%) for 30 min. 30 min before each challenge, the mice received 2.2 nmol of intranasal LTD₄, LTE₄, or a buffer control. A cohort of mice treated with 1% OVA were maintained as a positive control. The mice were then euthanized, their bronchoalveolar lavage (BAL) fluid was collected, and their lungs were examined histologically for evidence of inflammation and goblet cell metaplasia. Compared with saline-treated sensitized mice, mice challenged with low-dose OVA demonstrated low-grade BAL fluid eosinophilia (Fig. 3 A). These mice showed very mild pulmonary inflammation, as indicated by the accumulation of lymphocytes, plasma cells, and eosinophils around the bronchovascular bundles (Fig. 3, B and C). The administration of LTD₄ did not increase BAL fluid eosinophilia (Fig. 3 A) and slightly potentiated both bronchovascular inflammation (Fig. 3, B and C) and goblet cell metaplasia (Fig. 3, D and E). In contrast, LTE₄ significantly enhanced BAL fluid eosinophilia (Fig. 3 A), inflammation (Fig. 3, B and C), and goblet cell metaplasia (Fig. 3, D and E). The extent of the cellularity and goblet cell responses of the LTE₄-treated animals approached the levels of these parameters in the mice treated with 1% OVA.

To determine whether P2Y₁₂ receptors accounted for the LTE₄-mediated augmentation of pulmonary inflammation, sensitized mice were treated with clopidogrel, an antithrombotic agent that is converted in vivo to an active metabolite that covalently binds to and irreversibly inactivates P2Y₁₂ receptors (Savi et al., 2006). Drug treatment began 2 d before the first administration of LTs to allow conversion of the prodrug. A separate cohort of sensitized mice was maintained without clopidogrel treatment as a control group. Both groups were challenged with low-dose OVA with or without additional LTE₄. Treatment with clopidogrel eliminated LTE₄-induced potentiation of both inflammation and goblet cell metaplasia (Fig. 4, A and B). To determine the effect of P2Y₁₂ receptor blockade on the induced expression of mRNAs encoding proteins involved in goblet cell metaplasia, real-time PCR was used to analyze the lungs of the mice for the expression of IL-13 and the goblet cell–associated glycoprotein MUC5AC. LTE₄ modestly increased the expression of both transcripts, whereas clopidogrel treatment substantially suppressed the expression of both (Fig. 4 C). There was no detectable induction of IL-4 or IL-5 transcripts (unpublished data).

Because pharmacologic antagonists can act in an off-target manner, we sought to determine the receptors required to mediate the effect of LTE₄ in the lungs of allergen-sensitized and -challenged mice using a molecular approach. First, we studied the ability of LTE₄ to amplify pulmonary inflammation in OVA-sensitized and -challenged C57BL/6 mice lacking P2Y₁₂ receptors (p2ry12^−/− mice; Andre et al., 2003) along with age- and sex-matched C57BL/6 controls. Although cellular infiltration into the lung and BAL fluid was less pronounced in the C57BL/6 mice than in the BALB/c mice, LTE₄ potentiated BAL fluid eosinophilia (Fig. 4 D), goblet cell metaplasia (Fig. 4 E), and bronchovascular infiltration (Fig. 4 F) in the WT controls, all of which were severely blunted in the p2ry12^−/− mice (Fig. 4, D–F). To determine whether classical GPCRs for cys-LTs were also required for the LTE₄ effects, BALB/c mice lacking both CysLT₁R and CysLT₂R (Cysltr1/Cysltr2^−/− mice; Maekawa et al., 2008) were also studied. Because of limited numbers of available mice, all animals were sensitized and challenged with low-dose OVA, and half were treated with exogenous LTE₄. Strikingly, the potentiation of OVA-induced BAL fluid eosinophilia (Fig. 4 G), goblet cell metaplasia (Fig. 4 H), and inflammation (Fig. 4 I) by LTE₄ were completely intact in the Cysltr1/Cysltr2^−/− mice, indicating that LTE₄ was working independently of the known GPCRs for cys-LTs. Thus, the intrapulmonary actions of LTE₄ in vivo require P2Y₁₂ receptors but not classical cys-LT–reactive GPCRs.

To determine whether pharmacologic blockade of P2Y₁₂ receptors altered pulmonary inflammatory responses in a more physiological model of pulmonary inflammation, C57BL/6 mice were administered an extract of house dust mite Dermatophagoides farinae (Der f) intranasally twice weekly for 3 wk, with or without clopidogrel treatment. Two different doses of Der f were used to elicit moderate (3 µg) and severe (10 µg) inflammation, respectively. At both antigen doses, the mice treated with clopidogrel showed ∼75% attenuation of BAL fluid eosinophilia (unpublished data), as well as significant reductions in pulmonary inflammation, and goblet cell metaplasia compared with the cohort that did not receive clopidogrel (Fig. 5, A and B).

Because P2Y₁₂ receptors are essential for normal platelet activation in vivo (Andre et al., 2003), we sought to determine whether platelets were required for the response of sensitized challenged mice to exogenous LTE₄. Platelets were depleted in sensitized BALB/c mice by the i.v. injection of a rat mAb against mouse CD42b (GPIbα; Nieswandt et al., 2000) or an isotype-matched control IgG 48 h before to the first administration of LTD₄ or LTE₄. Treatment with the anti-CD42b antibody depleted platelets almost completely (99% depletion [unpublished data]) at the time of the first dose of LT. Platelet depletion resulted in a complete loss of the LTE₄-mediated potentiation of airway eosinophilia, inflammation, and goblet cell metaplasia (Fig. 6, A and B). The effect of platelet depletion was identical to the effect of treatment of the mice with clopidogrel, and the two treatments were not additive (Fig. 6 B). Platelet depletion, like clopidogrel treatment, sharply reduced the LTE₄-mediated expression of mRNA encoding IL-13 and MUC5AC (unpublished data). To determine whether LTE₄ alone induced platelet activation (leading to ADP release with potential resultant autocrine stimulation of P2Y₁₂ receptors), human blood platelets were stimulated with 1 µM LTE₄ or with 100 µM ADP, and degranulation was assessed by cytofluorographic detection of P-selectin (CD62P). As expected, ADP elicited CD62P expression, but no CD62P expression was detected in response to stimulation with LTE₄ (Fig. 6 C).

This study establishes that P2Y₁₂ receptors are essential for the actions of LTE₄, the only long-lived stable abundant member of the cys-LTs, in inflamed lung. Our findings help to explain long-recognized but unexplained properties of LTE₄ in airway biology. The involvement of cys-LTs in the pathobiology of asthma is established by the fact that 5-LO inhibitors (Israel et al., 1996; Liu et al., 1996) and CysLT₁R antagonists (Knorr et al., 1998) have clinical efficacy. The cloning and functional characterization of the CysLT₁R (Lynch et al., 1999) and CysLT₂R (Heise et al., 2000) explained the pharmacology of LTC₄ and LTD₄ predicted from studies of contractile tissues (Lee et al., 1984). The finding that neither GPCR showed significant binding or reactivity to LTE₄ was surprising given the plethora of data in human and animal studies that indicate the unique characteristics of this stable ligand relative to its short-lived precursors (Christie et al., 1992a, 1993; Laitinen et al., 1993; Gauvreau et al., 2001). Because we had previously demonstrated that LTE₄ could activate LAD2 cells by a mechanism independent of CysLT₁R and CysLT₂R (Paruchuri et al., 2008), and because LTE₄ was previously identified by an in silico model as a potential surrogate ligand for the P2Y₁₂ receptor (Nonaka et al., 2005), we undertook this study to determine whether the P2Y₁₂ receptor was a bona fide LTE₄-reactive receptor and to determine its contribution to LTE₄-induced proinflammatory events.

We first established that recombinant human P2Y₁₂ receptor protein conveyed activation responses by CHO cells to LTE₄. Unlike most members of the P2Y receptor class, native P2Y₁₂ receptors do not couple to Gαq proteins or activate calcium flux; instead, they induce signaling through PTX-sensitive Gαi2 proteins when stimulated with ADP (Foster et al. 2001; Lova et al., 2002; Woulfe et al., 2002). It was, thus, not surprising that CHO cells expressing P2Y₁₂ receptors failed to flux calcium in response to cys-LTs (Fig. 1 A) or that blockade of P2Y₁₂ receptors on LAD2 cells with 2-MesAMP failed to alter cys-LT–mediated calcium flux (Fig. 2 A), which was totally abrogated by MK571. However, heterologously expressed P2Y₁₂ receptors responded to LTE₄ with PTX-sensitive ERK activation (Fig. 1). The dose range for this response (Fig. S1) is similar to the LTE₄ dose range required to compete with radiolabeled ADP for binding to LAD2 cell membranes (Fig. S3) and is consistent with that reported by Nonaka et al. (2005) using a P2Y₁₂-Gα16 fusion protein to demonstrate calcium flux, also in CHO cells. The finding that cys-LT–induced ERK activation in the transfectants was resistant to MK571 (Fig. 1 D) implies that P2Y₁₂ receptors could contribute to an element of cys-LT–driven pathobiology that is relatively selective for LTE₄ and may resist conventional CysLT₁R antagonists. This element may be especially relevant to pathological situations where LTE₄ is abundant as a result of its relative stability.

In our previous study, LTE₄ had exhibited unanticipated potency for inducing ERK activation and the generation of MIP-1β and COX-2–dependent PGD₂ by LAD2 cells (Paruchuri et al., 2008). Because both ERK and calcium-dependent transcriptional events are essential for MC activation, we sought to determine the potential contribution of P2Y₁₂ receptors to the activation responses of LAD2 cells to LTE₄ and to contrast these responses to those elicited by LTD₄, the most potent CysLT₁R ligand. Although ineffective for blocking cys-LT–induced calcium flux, 2-MesAMP effectively blocked the LTE₄-mediated increment in MIP-1β production (Fig. 2 B) and also reduced the response to the higher concentrations of LTD₄, indicating that LTD₄ at high concentrations can also activate native P2Y₁₂ receptors, as supported by its actions as an agonist for ERK activation in the P2Y₁₂ receptor transfectants (Fig. 1). MK571 suppressed MIP-1β generation in response to both ligands (likely reflecting the requirement for calcium flux for chemokine generation) and was additive with 2-MesAMP for the suppression of the response to LTD₄. The shRNA-mediated knockdowns of CysLT₁R and P2Y₁₂ receptors revealed strong dependence of LTD₄-mediated activation on CysLT₁R, whereas LTE₄-mediated responses were clearly P2Y₁₂ receptor dependent (Fig. 2 C). Although there is segregation of the receptor requirements for these two related ligands on the same cell, the ability of LTE₄ to “cross over” and induce some calcium signaling through the CysLT₁R may permit some complementarities between the CysLT₁R (via Gαq proteins and calcium-induced pathways) and P2Y₁₂ receptors (via Gαi proteins), particularly for chemokine generation. This is analogous to the cooperation by Gαq-linked P2Y₁ receptors with P2Y₁₂ receptors in regulating ADP responses of platelets (Lova et al., 2002; Woulfe et al., 2002). The P2Y₁₂-dependent activation of LAD2 cells does not likely reflect autocrine effects of released ADP because it was resistant to treatment of the cells with the ectonucleotidase apyrase (Fig. S2).

Surprisingly, although P2Y₁₂ receptors were essential for competition between LTE₄ with ADP for binding to membranes of LAD2 cells (Fig. S3, A and B), they were not involved in the direct low-affinity binding of [3H]LTE₄ to these same membranes (Fig. S3 C). Additionally, LTE₄ could not compete for binding to P2Y₁₂ receptors expressed in isolation on COS-7 cells (Fig. S3 D). The fact that P2Y₁₂ receptors do not directly bind LTE₄ despite their essential nature implies that they are components of a complex with another LTE₄-reactive GPCR, perhaps with the putative CysLT_ER reported in the mouse skin (Maekawa et al., 2008). The facts that P2Y₁₂ receptors are required for LTE₄ to activate transfected CHO cells (Fig. 1) and LAD2 cells (Fig. 2), and for competition between LTE₄ and ADP for binding to LAD2 cells (Fig. S3 B), and that knockdown of P2Y₁₂ receptors did not reduce direct binding of radiolabeled LTE₄ (Fig. S3 C) are all consistent with this thesis. Precedents for such complexes on MCs include CysLT₁R and CysLT₂R heterodimers (Jiang et al., 2007) and a functional requirement of CysLT₁R for the uridine diphosphate-reactive P2Y6 receptor (Jiang et al., 2009). Our data indicate that presence of P2Y₁₂ is required for signaling and activation by LTE₄ in a cell-specific context.

The fact that LTE₄, but not LTD₄, induces bronchial eosinophilia when administered by inhalation to the airways of individuals with asthma (Gauvreau et al., 2001) argued for the presence of an LTE₄-reactive receptor in inflamed lung. In our model, which was designed to study potentiation of submaximal bronchial inflammation in sensitized mice, we found that LTE₄ exceeded the potency of LTD₄ for potentiating BAL fluid eosinophilia (Fig. 3 A), cellular infiltration of the bronchovascular bundles (Fig. 3, B and C), and goblet cell metaplasia (Fig. 3, D and E). Thus, LTE₄ exceeds the efficacy of LTD4 for potentiating bronchial inflammation in both mouse and man, a pattern not explicable by the known properties of CysLT₁R or CysLT₂R. Several pieces of evidence link these LTE₄-mediated responses in mice to the P2Y₁₂ receptor. First, treatment of the mice with a highly potent selective antagonist of the P2Y₁₂ receptor, clopidogrel, completely eliminated the histological response to LTE₄ (Fig. 4, A and B), which is associated with blockade of LTE₄-potentiated expression of IL-13, the Th2 cytokine most closely linked to the development of goblet cell metaplasia (Zhu et al., 1999), and of MUC5AC, a major mucus glycoprotein which is controlled by IL-13 (Fig. 4 C). Second, LTE₄ fully amplified pulmonary inflammation in Cysltr1/Cysltr2^−/− mice (Fig. 4, G–I). Lastly, the ability of LTE₄ to potentiate mucosal inflammation and goblet cell metaplasia was nearly completely abrogated in the absence of P2Y₁₂ receptors (Fig. 4, D–F). Thus, LTE₄ requires P2Y₁₂ receptors, but not the classical CysLTRs, to amplify the histological signatures of allergen-induced pulmonary inflammation. Because LTD₄ is converted in vivo to LTE₄, it was somewhat surprising that the ligands do not behave interchangeably in potentiating bronchial eosinophilia in mouse or man. In addition to P2Y₁₂-dependent effects of LTE₄, LTD₄ may initiate inhibitory signals through CysLT₂R (Jiang et al., 2007) or other yet-to-be-identified GPCRs that recognize LTD₄ but not LTE₄. The effects of clopidogrel in the model of airway disease induced by the natural allergen Der f without the use of exogenous LTE₄ (Fig. 5) support the importance of the P2Y₁₂ pathway in the integrated biology of pulmonary inflammation (although this does not discriminate between effects mediated by LTE₄ from ADP). The blockade of the response by clopidogrel distinguishes the P2Y₁₂ receptor-dependent response of the lung to LTE₄ from the clopidogrel-resistant LTE₄ response in the skin (Maekawa et al., 2008). Thus, it is likely that different receptors or receptor complexes mediate response to the stable ligand LTE₄ in distinct anatomical distributions.

Platelets accumulate in the lungs of individuals with asthma (Jeffery et al., 1989) and also are recruited to the lungs of OVA-sensitized and -challenged mice by an IgE-dependent mechanism (Pitchford et al., 2008). Activated platelets generate mediators (serotonin and thromboxane) that can potentiate airway inflammation. The essential nature of P2Y₁₂ receptors for normal platelet function (Andre et al., 2003) led us to examine the effect of platelet depletion on the response of sensitized challenged mice to exogenous LTE₄. Indeed, the response to LTE₄ was totally abrogated by platelet depletion (Fig. 6, A and B). Whether the requirement for platelets is direct or indirect is less clear. The fact that LTE₄ did not induce platelet expression of CD62P (an activation marker used as a surrogate for ADP release; Abrams and Shattil, 1991; Fig. 6 C) argues that LTE₄ does not trigger an ADP-dependent autocrine loop initiated by LTE₄-mediated activation of another receptor. Because platelet activation involves complementary signaling in both Gi and Gq protein–mediated pathways (Jin et al., 2002; Lova et al., 2002; Woulfe et al., 2002) and P2Y₁₂ provides only the Gi-linked component, LTE₄ (and P2Y₁₂) likely synergizes in vivo with a second agonist to facilitate platelet functions in the lung. It is noteworthy that LTE₄-mediated potentiation of the contractile responses of both guinea pig and human airway smooth muscle is COX dependent (Christie et al., 1992a) and was attributed to secondary generation of thromboxane, a major platelet-derived eicosanoid (Jacques et al., 1991). In retrospect, this finding may have reflected P2Y₁₂ receptor-dependent signaling on platelets or other cell types. The fact that LTE₄ potentiated inflammation only in sensitized challenged mice may reflect the previously described platelet-mediated pathway for leukocyte recruitment that depends on sensitization and IgE (Pitchford et al., 2008). The complete lack of LTE₄ reactivity in the face of platelet depletion argues against a role for MCs and other P2Y₁₂-bearing cell types, at least in this model.

Our findings suggest a potential therapeutic application for P2Y₁₂ receptor antagonists that may be especially relevant to AERD, which is associated with both high levels of LTE₄ in the urine (Christie et al., 1992b) and selective hyperresponsiveness to LTE₄ (Christie et al., 1993). In AERD, both 5-LO antagonists and CysLT₁R antagonists provide disease control and attenuate the consequences of aspirin challenge (Israel et al., 1993; Dahlén et al., 1998, 2002; White et al., 2006), in which the reaction is characterized by a marked surge in LTE₄ generation. The fact that the effects of LTE₄ persist in the absence of CysLT₁R and CysLT₂R implies that this pathway would be resistant to the available cys-LT receptor antagonists, all of which selectively block CysLT₁R. We speculate that simultaneous interference with the bronchoconstrictive effects of LTD₄ (via CysLT₁R) and with proinflammatory effects of LTE₄ (by P2Y₁₂), or more complete blockade of 5-LO, might improve clinical efficacy. Our study furthermore highlights the biological significance of the structural and functional relationships between the P2Y and cys-LT–reactive classes of GPCRs.

All animal protocols were approved by the Dana Farber Cancer Institute Animal Care and Utilization Committee. BALB/c mice lacking both CysLT₁R and CysLT₂R (Cysltr1/Cysltr2^−/− mice) and their WT littermate controls were derived as described elsewhere (Maekawa et al., 2008). Mice lacking P2Y₁₂ receptors (p2ry12^−/− mice) were derived as described elsewhere (Andre et al., 2003) on a mixed C57BL/6-129 background and backcrossed for 10 generations with C57BL/6 mice. WT BALB/c and C57BL/6 mice were purchased from Taconic.

The LAD2 line (Kirshenbaum et al., 2003) isolated from the bone marrow of a patient with MC leukemia was a gift from A. Kirshenbaum (National Institutes of Health). These cells were cultured in StemPro-34 (Invitrogen) supplemented with 2 mM l-glutamine (Invitrogen), 100 IU/ml Pen-strep (Invitrogen), and 100 ng/ml stem cell factor (SCF; Thermo Fisher Scientific). Cell culture medium was hemidepleted every week with fresh medium and 100 ng/ml SCF. CHO cells were grown in DMEM/F-12 with 10% FBS and 100 IU/ml Pen-strep (Invitrogen). Primary cord blood MCs were derived in vitro in RPMI 1640 (Invitrogen) supplemented with 10% fetal bovine serum (Sigma-Aldrich), 2 mM l-glutamine, 0.1 mM nonessential amino acids, 0.2 µM 2-ME, 100 U/ml penicillin, 100 µg/ml streptomycin, and 2 µg/ml gentamycin supplemented with SCF, IL-6, and IL-10 (R&D Systems) as previously described (Ochi et al., 1999).

0.5–1 × 10⁶ cells/sample were washed and labeled with Fura 2-AM for 30 min at 37°C. Cells were stimulated with the indicated concentrations of LTD₄ or LTE₄, and changes in intracellular calcium concentration were measured using excitation at 340 and 380 nm in a fluorescence spectrophotometer (F-4500; Hitachi; Paruchuri et al., 2008). The relative ratios of fluorescence emitted at 510 nm were recorded and displayed as a reflection of intracellular calcium concentration. In some experiments, cells were preincubated with the 1 µM of the CysLT₁R antagonist MK571 or with 100 µM of the P2Y₁₂ receptor antagonist 2-MesAMP for 5 min before the stimulation.

The expressions of human P2Y₁₂ receptor mRNA and mouse IL-13 and MUC5AC RNAs were determined with real-time PCR performed on an ABI PRISM 7700 Sequence detection system (Applied Biosystems). RNA was isolated with an RNeasy mini kit (QIAGEN) and was treated with RNase-free DNase (Invitrogen) according to the manufacturer’s protocol. cDNA was synthesized from 1 µg RNA with Superscript II RNase H-RT (Invitrogen). RT was performed using TaqMan RT reagents. Primers and FAM-labeled PCR mix were purchased from SABiosciences.

shRNA constructs targeting human CysLT₁R and P2Y₁₂ receptors were purchased from Thermo Fisher Scientific. The constructs were cloned into a lentiviral vector (pLKo1; Thermo Fisher Scientific) and used to generate infectious particles with a lentiviral packaging mix (Virapower; Invitrogen) according to the manufacturer’s protocol. The transfections were performed as described previously (Jiang et al., 2007). FACS analysis was used to confirm the knockdowns of CysLT₁R, whereas quantitative PCR was used to verify the knockdown of the P2Y₁₂ receptor.

A human P2Y₁₂ receptor cDNA was amplified by 30 cycles of PCR from RT total RNA extracted from primary hMCs. The primer sequences were 5′-CAACAAGAAATGCAAGCCGTCGA-3′ and 5′-TACATTGGAGTCTCTTCATTTGG-3′. The fragment was cloned into a TA vector (Invitrogen). After verifying the nucleotide sequence, the fragment was subcloned into the multiple cloning site of the expression vector pEF1/His B, encoding a C-terminal HIS tag. A plasmid expressing the P2Y₁₂ receptor construct in the forward orientation was transfected into CHO cells using Fugene HD reagent according to manufacturer’s protocol. A construct in the reverse sequence was transfected in parallel as a negative control. Stably expressing clones were selected in the medium containing 1,000 µg/ml G418 (Invitrogen), and expression of the construct was confirmed by FACS analysis of permeabilized cells with a mAb against the HIS tag.

LAD2 cells were stimulated with the indicated concentrations of LTD₄ or LTE₄ (Cayman Chemical) in the presence or absence of MK571 (Cayman Chemical) or 2-MesAMP (Sigma-Aldrich) or were passively sensitized with 2 µg/ml of human myeloma IgE (Millipore) overnight and stimulated with 1 µg/ml of rabbit anti–human anti-IgE (Millipore) as detailed elsewhere (Paruchuri et al., 2008). In some experiments, LAD2 cells were incubated with 10 µM apyrase (Sigma-Aldrich) to degrade extracellular nucleotides during the activation. The concentration of MIP-1β was measured by an ELISA (Thermo Fisher Scientific). PGD₂ was quantitated by a PGD₂-methoxylamine hydrochloride assay (Cayman Chemical).

After stimulation with the respective agonists, 0.5 × 10⁶ LAD2 cells and CHO cells were lysed with lysis buffer (BD) supplemented with protease inhibitor cocktail (Roche) and 1 mM sodium vanadate. Lysates were subjected to 4–12% SDS-PAGE and transferred to PVDF membranes. The membranes were incubated with antibodies against phospho and total ERK (Cell Signaling Technology) in PBS, 5% dry milk, and 0.1% Tween-20 (1:1,000) overnight at 4°C on shaker and then with secondary antibody (peroxidase-conjugated anti–rabbit or anti–mouse). Bands were visualized with enhanced chemiluminescence (Thermo Fisher Scientific).

[³H] LTD₄ ([14,15,19,20-³H(N)]; specific activity 100–240 Ci/mmol; PerkinElmer) was converted to [³H]LTE₄ by the dipeptidases present in the serum. In brief, 100 µl (80 nM) [³H]LTD₄ was incubated with 100 µl of 10% serum for 2 h at room temperature. The converted product was extracted into 400 µl of methanol, evaporated in the presence of nitrogen, and diluted to the required concentration with the binding buffer for the binding assay. The conversion was confirmed by running an aliquot on high-performance reverse-phase liquid chromatography. The fraction eluting with the LTE₄ peak accounted for >99% of the radioactivity used in the assays. The specific activity of the resultant LTE₄ was consistently 80–100% that of the LTD₄ from which it was converted. [³H]ADP and [³H]LTE₄ binding assays were performed using membrane proteins as described earlier (Maekawa et al., 2001). In brief, cells were washed in PBS, resuspended in PBS supplemented with protease inhibitor cocktail, and lysed by sonication for 5 min. The lysed cells were centrifuged at 100,000 g for 45 min and the microsomal pellet was resuspended in 1 ml PBS. Protein concentration was determined using a Protein Assay kit (Bio-Rad Laboratories). 50 µg of membrane protein was incubated for 1 h at room temperature in 250 µl of 10 mM Hepes/KOH, pH 7.4, with various concentrations of radiolabeled ligand and cold competitor. Bound [³H] was separated from free [³H] by filtration through GF/C filters (GE Healthcare) and washed twice with 10 mM Hepes/KOH, pH 7.4, containing 0.01% BSA. The residual membrane-associated [³H] on the filter was determined in 2 ml scintillation fluid by β counter (PerkinElmer). For [³H]ADP, specific binding was determined by subtracting the amount of [³H] bound in the presence of 100 µM 2-MesADP from total binding. For [³H]LTE₄, nonspecific binding was calculated as the residual radioactivity bound in the presence of 10 µM of unlabeled LTE₄.

Male BALB/c mice (6–8 wk old) received i.p. injections of 10 µg of reagent-grade chicken egg OVA precipitated with 2.25 µg aluminum hydroxide on days 0 and 7. On days 14–16, the mice received single intranasal doses of 2.2 nmol LTD₄, LTE₄, or buffer. 30 min after each dose, the mice were exposed to an aerosol of 0.1% OVA for 30 min delivered by an ultrasonic nebulizer. As a positive control group, some mice received 1% OVA without LTD₄ or LTE₄ pretreatment. The mice were euthanized 24 h after their third and final OVA aerosol challenge. In some experiments, the mice received 500 µg/ml clopidogrel in drinking water for 3 d before the first intranasal dose of LTs. The treatment was continued throughout the procedure. For Der f–mediated pulmonary inflammation, 6–8-wk-old C57BL/6 mice were lightly anaesthetized and received either 3 or 10 µg Der f extract (Greer Laboratories) on days 0, 3, 7, 10, 14, and 17. Some mice received clopidogrel throughout the duration of the experiment. Mice were euthanized 24 h after the last intranasal instillation.

Sensitized mice were treated i.v. with 50 µg (∼2 µg/g body weight) of a monoclonal rat IgG directed against mouse CD42b (GPIbα) or an equal amount of isotype control (both from Emfret Analytics; Nieswandt et al., 2000). Each antibody was diluted in 50 µl of sterile saline. The depletion of platelets was confirmed by automated counting.

The left lungs were fixed for at least 8 h in 4% paraformaldehyde and embedded in glycol methacrylate, as described previously (Kim et al., 2006). Then, 2.5-µm-thick glycol methacrylate sections were stained with H&E for general morphological examination. For histological study of the mucus-secreting cells of the epithelium (goblet cells) of the airways, lung sections were stained with PAS. The extent of cellular infiltration in the bronchovascular bundles was evaluated without knowledge of the particular treatment and was categorized arbitrarily into the following grades: 0, no inflammation; 1, mild inflammation; 2, moderate inflammation; and 3, severe inflammation. 15 bronchovascular bundles of each mouse were evaluated. The stained goblet cells were enumerated in at least four independent bronchovascular bundles from the lung sections obtained for each animal in the different experimental groups. The length of basal lamina of corresponding bronchus was measured by ImageJ image analysis software (National Institutes of Health). Only the comparable large-caliber preterminal bronchi (diameter 200–220 µm) were examined because minimal changes occur in terminal bronchioles. The data were expressed as the mean of goblet cell counts stained in each bronchus in each section per millimeter of bronchial basal lamina.

Data are expressed as mean ± SEM from at least three experiments except where otherwise indicated. Significance was determined with the Welch’s test for samples of unequal variance. Analysis of variance was used to test differences between multiple groups.

Fig. S1 shows the dose-dependent effects of LTE₄ and LTD₄ on ERK activation in CHO cells stably transfected with the human P2Y₁₂ construct. Fig. S2 shows the effect of apyrase treatment on LT-mediated activation of LAD2 cells. Fig. S3 shows competitive ligand binding assays.

We thank Arlene Simon for her careful editing of this manuscript and K. Frank Austen and Jilly Evans for helpful comments and suggestions. We thank Gretchen Bain for her assistance with the binding assays.

This work was supported by grants from the National Institutes of Health and by a generous contribution from the Vinik Family.

Pamela Conley is employed by Portola Pharmaceuticals. The other authors have no financial conflicts of interest to report.