CysLT₂ receptors interact with CysLT₁ receptors and down-modulate cysteinyl leukotriene–dependent mitogenic responses of mast cells

The cysteinyl leukotrienes (cys-LTs) LTC₄, LTD₄, and LTE₄ are inflammatory mediators derived from arachidonic acid and generated in vivo by mast cells (MCs), eosinophils, basophils, and alveolar macrophages. Cys-LTs are powerful contractile agonists for vascular and bronchial smooth muscle,¹  reflected by bronchoconstriction in human subjects,²  and by rapid vascular leak in passive cutaneous anaphylaxis and experimental peritonitis in mice.³  The known functions of cys-LTs are mediated by 2 G protein–coupled receptors (GPCRs), termed CysLT₁ and CysLT₂.^4,5  A recently deorphanized GPCR (GPR17) is dual selective for cys-LTs and uracil nucleotides and dominantly expressed in the brain.⁶  Although the CysLT₁ and CysLT₂ receptors induce similar signaling events (calcium flux and inositol phosphate accumulation) when heterologously expressed in isolation by Xenopus oocytes,^4,5  they are not functionally redundant. Each mediates distinct responses in vivo,⁷  and each has a unique pattern of cellular and tissue distribution.^4,5  Several leukocyte subsets in peripheral blood and inflammatory lesions express both CysLT₁ and CysLT₂ receptors.⁸  It is thus possible that these receptors interact to regulate cys-LT–mediated responses in inflammation, although the functional significance of coexpression of these receptors by cells of the immune system remains unknown

MCs are tissue-dwelling effector cells that are central to asthma and allergic diseases in humans.⁹  MCs also initiate vascular leak and inflammation in mouse models of allergy, innate immunity, and autoimmune diseases (reviewed in Rottem and Mekori¹⁰ ). MC hyperplasia is a common feature of allergic mucosal inflammation. MCs express both CysLT₁ and CysLT₂ receptors, and cys-LTs potently induce calcium flux¹¹  and phosphorylation of mitogen-activated protein kinase-kinase-1 (MEK-1) and its downstream effector, extracellular signal-regulated kinase (ERK),^12,13  in human MCs (hMCs) cultured from cord blood. These signaling events result in MC proliferation,¹⁴  reflecting transactivation of c-Kit, a receptor tyrosine kinase (RTK) that binds stem cell factor (SCF) and is essential for MC development in vivo.¹⁵  Mice lacking leukotriene C₄ synthase (LTC₄S), the terminal enzyme needed for cys-LT synthesis, show markedly blunted mucosal MC hyperplasia in the lungs when sensitized and challenged with allergen, revealing a physiologic function of cys-LTs for MC development in vivo.¹⁶  Most cys-LT–mediated responses in vitro can be blocked by treatment of the MCs with a CysLT₁ receptor–selective antagonist, MK571. The lack of selective CysLT₂ receptor antagonists has precluded a comprehensive analysis of the contributions from this receptor to the integrated functions of cys-LTs on MCs and other primary cell types to date.

In the current study, we report that LTD₄-mediated proliferation is abrogated in hMCs subjected to short-hairpin RNA (shRNA)–mediated knockdown of CysLT₁ receptor expression, and is also absent in mouse bone marrow–derived MCs (mBMMCs) from Cysltr1^−/− mice. In striking contrast, LTD₄-induced proliferation of MCs is enhanced by knockdown or knockout of CysLT₂ receptors. Although the absence of CysLT₂ does not change the total cellular level of CysLT₁ protein, it increases CysLT₁ expression at the cell surface and enhances LTD₄-induced ERK phosphorylation. CysLT₁ and CysLT₂ receptors colocalize at the plasma membrane, nucleus, and nuclear envelope of MCs, and form heteromeric complexes as determined by Ab-based fluorescent lifetime imaging microscopy (FLIM) to measure fluorescent resonance energy transfer (FRET)¹⁷  in intact LAD2 cells (a human MC line).¹⁸  CysLT₂ receptors limit the strength of signaling through CysLT₁ receptors, at least partly by controlling membrane expression, and thus can function as negative regulators of cys-LT–induced MC responses.

LTD₄ and MK571 were purchased from Cayman Chemical (Ann Arbor, MI). Polyclonal antihuman Abs against the LTB₄ receptor BLT₁, CysLT₁ (against C-terminal amino acids 318–337 [YVPRKKASLPEKGEEICKV]), and CysLT₂ (against N-terminal amino acids 1–18 [MERKFMSLQPSISVSEME], and against C-terminal amino acids 330–346 [TKCVFPVSVWLRKETRV]), and a rabbit IgG control were all purchased from Cayman Chemical. A polyclonal antipeptide Ab (RB34, against the conserved sequence DEKNNTKCFEPPQNN of EC3 of both mouse and human CysLT₁ receptors) and the corresponding blocking peptide were custom generated by Orbigen (San Diego, CA). A monoclonal anti-CysLT₂ peptide Ab (clone 1B3, IgG2a against the N-terminal amino acid sequence MERKFMSLQPSISVSEME of human CysLT₂ receptor) and blocking peptide were provided by Nita Frank (Millennium Pharmaceuticals, Cambridge, MA). FITC-conjugated donkey antimouse and antirabbit Abs, antihuman polyclonal anti-α₄ integrin, and monoclonal anti-β₁ and anti-CXCR2 Abs were purchased from Pharmingen (San Diego, CA). Alexa Fluor 488–conjugated goat anti–rabbit IgG and Cy3-conjugated donkey anti–mouse IgG were purchased from Invitrogen (Frederick, MD). Alexa Fluor 488, Alexa Fluor 546 (Molecular Probes, Eugene, OR), and Cy3 (Amersham, Arlington Heights, IL) were used for direct labeling according to the manufacturers' protocols. Recombinant human and mouse cytokines were all purchased from R & D Systems (Minneapolis, MN).

The use of cord blood for this study is covered under a protocol for the use of discarded human materials approved by the Partners Healthcare Human Research Committee. hMCs were derived from cord blood mononuclear cells as described.¹⁹  mBMMCs were derived from bone marrow¹⁴  and studied at 4 to 6 weeks when virtually all stained with toluidine blue. LAD2 cells (provided by Dr Arnold Kirshenbaum, NIH, NIAID, Bethesda, MD) were maintained in SpemPro34 medium containing SCF (100 ng/mL).¹⁸  Mitogenic assays were performed on triplicate cell samples in the presence of SCF (10 ng/mL) with or without LTD₄ (500 nM). This dose was chosen based on its ability to saturate both receptors,^4,5  and provide maximal CysLT₁-dependent proliferation and ERK activation.¹⁴ 

Cysltr1^−/− mice (C57BL/6-Cysltr1^tm1Ykn) and Cysltr2^−/− mice (C57BL/6-Cysltr2^tm1Ykn) were derived as described.^7,20  All animal studies were approved by the Animal Care and Use Committee of the Dana-Farber Cancer Institute.

Cytokine-starved hMCs were stimulated for 5 minutes with various doses of LTD₄, SCF (100 ng/mL), or medium, lysed, and processed for Western blotting as described previously.¹⁴  Blots were probed with anti–Active ERK, total ERK (Cell Signaling Technology, Beverly, MA), and phosphorylated mouse c-Kit (Zymed, South San Francisco, CA). The signals were detected by chemiluminescence and the densities of the bands were quantitated by densitometry using a Chemi Imager 4400 (Alpha Innotech, San Leandro, CA). Phospho-ERK signals were corrected as a ratio of the corresponding total ERK. The corrected phospho-ERK signals for the unstimulated cells did not differ significantly between genotypes, and thus these values were expressed as 100% of control for that genotype in each experiment. The dosing range of LTD₄ (10-1000 nM) was chosen so as to be at or higher than the threshold for saturation of both receptors.

Abs were coupled to columns of a ProFound immunoprecipitation kit (Pierce, Rockford, IL) according to the manufacturer's protocol. Lysates were generated from primary hMCs and LAD2 cells, and 2.5 × 10⁷ cell equivalents were used for each sample. Immunoblotting of the eluted proteins was performed using the indicated Abs.

CysLT₁ and CysLT₂ receptor proteins were detected in permeabilized cells using anti–C-terminal Abs as described elsewhere.²¹  Specificity was confirmed by demonstrating diminished levels of staining in LAD2 cells and hMCs treated with receptor-selective shRNAs. Anti-CysLT₁ (RB34, 7 μg/mL) and anti-CysLT₂ Abs (1B3, 10 μg/mL) were used to detect the receptors on nonpermeabilized hMCs, LAD2 cells, and (in the case of RB34) mBMMCs. Specificities of the new Abs were confirmed using blocking peptides and, in the case of RB34, fluorescence-activated cell sorting (FACS) analysis with CysLT₁^−/− mBMMCs. Nonspecific rabbit and mouse IgG (BD Biosciences, San Jose, CA) were used as controls. Flow cytometry analyses were performed on a FACSort calibur flow cytometer, and data were analyzed using Cell Quest Pro software (BD Immunocytometry Systems, San Jose, CA) as described.¹¹  Data were calculated as net MFI (fluorescence intensity of the anti-CysLT₁ staining minus the staining of the IgG control). There were no significant differences in the staining of different mBMMC populations with the IgG control (MFI: 5.4 ± 1.3, 4.6 ± 1.1, and 6.1 ± 0.9 for the WT, CysLT₁^−/−, and CysLT₂^−/− mBMMCs, respectively).

shRNA constructs were purchased from Open Biosystems (Huntsville, AL). The constructs were designed to include a hairpin of 21 base pair of a sense strand and an antisense strand, and a 6 base-pair loop. The sequences were (reading from 5′ to 3′) as follows: CysLT₁: CCGGGCGTGACTTATGTACCCAGAACTCGAGTTCTGGG TACATAAGTCACGCTTTTT; CysLT₂: CCGGCCTGCAGGATTATGTCTTATTCTCGAGAATAAGACATAAT-CCTGCAGGTTTTT Each hairpin sequence was cloned in frame with a lentiviral vector (pLKo1; Open Biosystems). Clones were verified by automated sequencing. Infectious viral particles were prepared by cotransfection of 293FT cells with the pLKo1 and a lenitival packaging mix (Virapower; Invitrogen) according to the manufacturer's protocol. After titering the viral stocks in Hela cells, the lentivirus particles were concentrated by high-speed centrifugation. LAD2 and hMC suspensions were prepared at a density of 2.5 × 10⁵ cells/mL in StemPro 34 medium in a 75-cm flask. Viral stocks were added directly to the medium at a quantity sufficient to achieve a multiplicity of infection (MOI) of 10, based on preliminary experiments in which FACS analysis was used to assess the effect of the transfection on the expression of the target receptors. The cells were incubated overnight with the particles at 37°C. The following day, the medium was replaced and the cells were cultured for an additional 48 hours before performing the functional studies.

LAD2 cells were fixed in suspension with 2% paraformaldehyde in phosphate-buffered saline (PBS) on ice for 10 minutes. The cells were washed once with wash buffer (PBS containing 0.1% sodium azide) and resuspended in blocking buffer (PBS containing 5% horse serum). The cells were immobilized on round 12-mm coverslips by cytocentrifugation and postfixed with methanol at − 20°C for 10 minutes. The slides were washed twice with wash buffer and then blocked with blocking buffer, shaking at room temperature for 30 minutes. The cells were stained with primary Abs that were labeled directly with Alexa Fluor 488 or Alexa Fluor 546 (Molecular Probes) according to the manufacturer's specifications. Alternately, cells were stained with unlabeled primary Abs, followed by donkey anti–rabbit-FITC at 1:100, and donkey anti–mouse-Cy3 at 1:1000. Five percent donkey serum (Jackson Immunoresearch, Bar Harbor, ME) was added to the blocking buffer when indirect staining was used. The cells were counterstained with DRAQ5 nuclear stain (1:1000; Biostatus, Leicestershire, United Kingdom). Anti-CysLT1 C-terminal Ab (1:100), RB34 (1:100), and monoclonal or polyclonal anti-CysLT₂ N-terminal Abs (1:50) were used at the indicated dilutions, and species- and isotype-matched control IgGs were used at the same concentrations. Subcellular localization of the receptors was assessed using a Nikon TE2000-U inverted microscope with a Nikon C1 plus laser scanning confocal system (Melville, NY). Images were acquired through a 30-μm pinhole with a Nikon 60× Oil CFI Plan Apochromat objective and analyzed with Nikon EZ-C for Nikon C1 Confocal Software, Gold Version 3.40 build 691. The resolution was set for the highest possible given the objective and lasers according to the Nikon software (in most cases, 71 nm/pixel ± 1 nm). A 1.68 usec/pixel dwell time was used to capture a single z section. The 2-dimensional images were captured as 12-bit color images, which were then saved as 8-bit images by adjusting the histogram to remove any superfluous gray values using Photoshop CS1 (Adobe Systems, San Jose, CA).

The cells were fixed and stained for the FLIM analysis as described for confocal imaging. Cells were double stained with 1:100 dilutions of polyclonal Abs against CysLT₁ (RB34), BLT₁, or α₄ integrin, in various combinations with monoclonal (1B3) or polyclonal Abs to CysLT₂, monoclonals against CXCR2 or β₁ integrin, or with isotype controls. The cells were counterstained with secondary antirabbit–Alexa Fluor 488 (donor) and antimouse-Cy3 (acceptor) conjugates. In some experiments, anti-CysLT₁ Ab RB34 and the polyclonal anti–N-terminal CysLT₂ peptide Ab were labeled directly with Alexa Fluor 488 and Alexa Fluor 546 (or Cy3) to eliminate any potential FRET signal caused by interactions between secondary Abs and reorientation of fluorophores. Primary Abs were omitted in some samples as specificity controls. Cy3 and Alexa Fluor 546 were chosen as acceptor fluorophores because their excitation spectra overlap with the emission spectrum of Alexa Fluor 488. IgG alone was directly labeled as a control. Some cell samples were stimulated with 100 nM LTD₄ for various intervals before fixation and staining. To confirm double labeling and colocalization of receptors, confocal images were taken using 488-nm and 543-nm single-photon excitation to excite Alexa Fluor 488 and Cy3 (or Alexa Fluor 546), respectively (Zeiss LSM510/NLO; Carl Zeiss, Heidelberg, Germany). Emissions were separated in multitrack mode. FLIM was performed using a microscope (Zeiss LSM510/NLO) and a femtosecond-pulsed Ti:Sapphire laser (Chameleon, Coherent, Santa Clara, CA), with a multichannel plate photon counting detector (R3809; Hamamatsu, Okayama City, Japan). A time-correlated single-photon counting board and software (SPC830; Becker and Hickl, Berlin, Germany) was used for acquisition, and images were analyzed using SPCImage software (version 2.6.1.2711; Becker and Hickl) with monoexponential and biexponential lifetime curve fits.²²  A minimum of 6 fields was analyzed for each condition in each experiment, with each field containing a minimum of 4 cells and a maximum of 30 cells. Lifetimes were analyzed as whole-cell measurements. FRET is indicated by shortening of the donor fluorophore's lifetime and will occur if the 2 fluorophores are within 10 nm of each other. The non-FRETing fluorescence lifetime of the donor fluorophore (Alexa Fluor 488) was measured in the absence of the acceptor fluorophores (negative control) in each experiment. Results using Cy3 as an acceptor were similar to those using Alexa Fluor 546, and direct and indirect labeling methods revealed similar degrees of FRET.

Data are expressed as mean plus or minus SEM from at least 3 experiments except where otherwise indicated. Differences between treatment groups were determined with the Student t test. For the FRET studies, differences in lifetimes were assessed by ANOVA, with a Bonferroni analysis to correct for multiple comparisons.

We had previously reported that human MCs express both CysLT₁ and CysLT₂ receptor proteins.^11,21  A recent study of human bronchial fibroblasts²³  questioned the specificity of the anti–C-terminal Ab against human CysLT₁ that we had used for our previous studies. We therefore sought to validate this Ab reagent, and to generate additional reagents for the identification of both receptors. FACS using anti–C-terminal peptide Abs on permeabilized LAD2 cells and hMCs detected both CysLT₁ and CysLT₂ proteins (Figure 1A). The signals for each receptor were decreased in the cells treated with the corresponding shRNA (Figure 2A). The receptors were also detected by FACS on nonpermeabilized LAD2 cells (Figure 1B) and hMCs (not shown) using RB34, a polyclonal anti-CysLT₁ extracellular domain 3 Ab, and 1B3, a monoclonal anti–N-terminal CysLT₂ Ab. The signals elicited by each Ab were abrogated by prior treatment of the cells with corresponding blocking peptides, without “cross-blocking” of the nontarget receptors (Figure 1B). The RB34 Ab also detected a strong signal on the surface of WT mBMMCs, but not on mBMMCs from Cysltr1^−/− mice (Figure 1C). Confocal imaging of LAD2 cells using the polyclonal anti–C-terminal Abs (Figure 1D) and hMCs (not shown) confirmed that the 2 receptors colocalized at the plasma membrane, the nuclear envelope, and discreet nuclear inclusions (Figure 1D). Similar results were obtained regardless of the combinations of Abs used. Strong membrane and nuclear localization of CysLT₁ was also detected in WT and Cyslt2r^−/− mBMMCs (not shown).

To determine the potential functional relevance of CysLT₁-CysLT₂ interactions for MC responses to cys-LTs, we treated hMCs and LAD2 cells with shRNA constructs designed to suppress the expression of each receptor packaged into recombinant lentivirus.²⁴  FACS of permeabilized cells revealed that the knockdowns suppressed the expression of the target receptor proteins. In LAD2 cells (but not in hMCs), CysLT₁ knockdown increased CysLT₂ expression levels by approximately 2-fold relative to mock-transfected controls (n = 3, as shown for one experiment, Figure 2A). FACS of nonpermeabilized cells revealed that knockdown of CysLT₂ modestly but consistently increased the level of surface CysLT₁ (∼ 2-fold increase), but CysLT₁ knockdown did not change surface levels of CysLT₂ (n = 3, as shown for one experiment, Figure 2B). A similar trend was observed in mBMMCs from Cyslt2r^−/− mice (Figure 2C). Both basal and LTD₄-stimulated proliferation of the LAD2 cells was suppressed even by the control transfection, and thus these cells were not used for the mitogenic analysis. However, hMCs transfected with empty vector showed the same 2-fold LTD₄-mediated increase in the rate of thymidine incorporation that was observed in the nontransfected cells (not shown). CysLT₁ receptor knockdown abrogated this increase, while CysLT₂ receptor knockdown significantly amplified the increase (Figure 2D). In mBMMCs, the LTD₄-mediated increment in proliferation was absent in Cysltr1^−/− cells, but amplified in Cysltr2^−/− cells (Figure 2E). The LTD₄-induced increment in proliferation observed in the Cysltr2^−/− mBMMCs was abrogated by treatment of the cells with MK571. Basal SCF-dependent proliferation of hMCs was unaltered by the knockdowns, and did not differ among the mBMMCs from the 3 genotypes (mean ± SEM: 14 968 ± 3913, 14 223 ± 6817, and 13 270 ± 6550 cpm for wild-type, Cysltr1^−/−, and Cysltr2^−/− cells, respectively; n = 4).

Stimulation of wild-type mBMMCs with LTD₄ alone for 5 minutes caused ERK phosphorylation in a dose-dependent manner, which required the presence of CysLT₁ receptors. The absence of CysLT₂ receptors enhanced LTD₄-mediated ERK phosphorylation at every concentration (Figure 3A,B). Both WT and Cysltr2^−/− mBMMCs showed robust and equal c-Kit phosphorylation in response to LTD₄ at 100 and 1000 nM; this was not observed for Cysltr1^−/− mBMMCs. The cells of all 3 genotypes responded to SCF (Figure 3A).

Ab-based multiphoton FLIM^17,22  was used to assay for physical interactions between CysLT₁ and CysLT₂ receptors in intact MCs based on FRET. We separately tested for FRET using different combinations of each polyclonal Ab against the CysLT₁ receptor with either the polyclonal or monoclonal Abs against the CysLT₂ receptor to ensure consistency of the results, and used both direct and indirect labeling of these Abs with the fluorophores to control for potential artefactual interactions between secondary Abs. FLIM of the cells stained with anti–CysLT₁–Alexa Fluor 488 alone (the negative control for FRET) gave an average lifetime of 1.9 ns (Figure 4C, as shown for one field, Figure 4A). LAD2 cells stained with anti-CysLT₁ Ab RB34 in combination with either polyclonal or monoclonal anti-CysLT₂ N-terminal Abs showed significant FRET (n = 4, as shown for one experiment using polyclonal anti-CysLT₂, Figure 4A). Similar degrees of FRET were observed using either direct or indirect labeling methods, and using either Cy3 or Alexa Fluor 488 as acceptor fluorophores (Figure 4C). Areas of FRET were detected in the nucleus as well as in proximity to the plasma membrane. In contrast, no FRET was detected between BLT₁ and CysLT₂ or CXCR2 and CysLT₁ receptors used as specificity controls, whereas the α₄ and β₁ integrins (used as a positive control) showed strong interaction (n = 3, as shown for one experiment, Figure 4B). Immunoprecipitates generated from hMCs (not shown) and LAD2 cells with either anti-CysLT₁ or anti-CysLT₂ polyclonal Abs demonstrated coprecipitation of the 2 receptors by Western blot (Figure 4D). Abs against the BLT₁ receptor for LTB₄ or IgG alone failed to coprecipitate cys-LT receptors (Figure 4D). LAD2 cells (Figure 4D) and hMCs (not shown) expressed BLT₁ protein.

The abundance of extracellular cys-LTs in allergic mucosal inflammation reflects production of these mediators by resident MCs and infiltrating eosinophils, basophils, and macrophages. Unlike these other cells, MCs can proliferate in situ, and the lack of antigen-induced MC hyperplasia in the lungs of LTC₄S null mice may in part reflect a role for cys-LTs in the control of this process. Like many hematopoietic cells found in inflammatory lesions, MCs express both CysLT₁ and CysLT₂ receptors.^21,25  When heterologously expressed by transfection into cell lines, GPCRs can heterodimerize with other GPCRs that recognize the same or similar ligands^26,27 ; these heterodimers can profoundly alter signaling responses. This led us to suspect that cys-LT–dependent mitogenic responses of MCs might be regulated by interactions between CysLT₁ and CysLT₂ receptors.

To avoid the potential confounding influences of artifacts induced by overexpressing GPCRs in transfected cells, we used primary hMCs, mBMMCs, and LAD2 cells (a human MC line)¹⁸  for studies of receptor interactions. Both receptors were detected by FACS on permeabilized LAD2 cells and hMCs using respective human-specific Abs against the C-termini of each receptor (Figure 1A), as well as without permeabilization using a new polyclonal anti-CysLT₁ extracellular domain 3 (EC3) Ab (RB34), and a monoclonal anti–N-terminal CysLT₂ Ab (1B3). The binding of each new Ab was abrogated by treatment of the cells with corresponding blocking peptides, without “cross-blocking” of the nontarget receptors (Figure 1B). The RB34 Ab (against a peptide conserved between the EC3 of the mouse and human receptors) also detected a strong signal on the surface of WT mBMMCs, but not on mBMMCs from CysLT₁^−/− mice (Figure 1C), further confirming the specificity of this reagent.

We next used confocal imaging to determine the subcellular localization of the CysLT₁ and CysLT₂ receptors in LAD2 cells. We stained the cells using various combinations of each anti-CysLT₁ Ab along with either 1B3 or a polyclonal anti–N-terminal CysLT₂ Ab directed to the same peptide. The cells were stained with primary Abs directly labeled with fluorophores (Alexa Fluor 488 for CysLT₁, Alexa Fluor 546 or Cy3 for CysLT₂) in some experiments, and in others were indirectly labeled with fluorophore-conjugated secondary Abs. Regardless of the combination of Abs or labeling method used, CysLT₁ colocalized with CysLT₂ on the plasma membrane, at the nuclear envelope, and within discreet nuclear inclusions (Figure 1D). Nonspecific IgG did not stain the cells. CysLT₁ constitutively localizes to the outer nuclear envelope of human colonic carcinoma cells,²⁸  and translocates to this location in LTD₄-stimulated primary epithelial cells, where it regulates mitogenic signaling and intranuclear activity of the ERK pathway. The obligate biosynthetic enzyme LTC₄S also localizes to the nuclear envelope of MCs and other hematopoietic cells,²⁹  and the presence of 2 cys-LT receptors at the synthetic organelle for LTC₄ may permit intracrine functions as proposed in our previous study.¹⁴  Moreover, the constitutive colocalization of these receptors to both the nucleus and cell surface of MCs highlights the potential for receptor-receptor interactions to regulate cys-LT–dependent signaling.

Although blockade by selective antagonists such as MK571 infers cellular responses mediated by CysLT₁, MK571 can also have off-target effects, such as interference with sphingosine-1-phosphate transport³⁰  and blockade of purinergic receptors.³¹  Moreover, selective antagonists of CysLT₂ receptors are not currently available. Thus we used a shRNA “knockdown” approach for each receptor to determine the potential functional relevance of CysLT₁-CysLT₂ interactions for MC responses to extracellular cys-LTs. FACS of permeabilized cells using anti–C-terminal Abs revealed that the knockdowns suppressed the expression of the target receptor proteins (Figure 2A). While the knockdowns did not change the total cellular expression of the nontarget receptors, CysLT₂ knockdown unexpectedly caused a 2-fold increase in surface expression of CysLT₁ by LAD2 cells (as shown in Figure 2B). Likewise, the levels of CysLT₁ protein surface expression (but not total cellular expression) were significantly higher on mBMMCs from Cyslt2r^−/− mice than from wild-type controls (Figure 2C). In both human and mouse MCs, the loss of CysLT₁ receptors abrogated the increment in proliferation induced by LTD₄ (Figure 2D,E), confirming the requirement for this receptor for the mitogenic response.¹⁴  In striking contrast, the loss of CysLT₂ receptors significantly enhanced LTD₄-dependent proliferation of both cell populations (Figure 2D,E). Thus CysLT₂ receptors down-modulate CysLT₁ receptor–dependent proliferation in the MCs of both mice and humans, possibly by limiting the number of CysLT₁ receptors expressed at the cell surface and thus access of the latter to extracellular ligands. Alternatively, CysLT₂ receptors could serve as a physiologic “decoy” and limit access of CysLT₁ to its ligands. The availability of Abs that recognize extracellular epitopes should prompt a reanalysis of levels of membrane presentation of the cys-LT receptors, as opposed to their total levels, as a mechanism for control of cys-LT responsiveness.

Exogenous cys-LTs drive proliferation of cord blood–derived hMCs by inducing CysLT₁-mediated transactivation of the c-Kit tyrosine kinase and consequent ERK phosphorylation,¹⁴  the latter initiating the cascade of signaling and transcriptional events necessary for incremental proliferation. We sought to determine whether CysLT₂ receptors participated in this process. Stimulation of wild-type mBMMCs with LTD₄ alone for 5 minutes caused both ERK and c-Kit phosphorylation. Both of these responses were absent in Cysltr1^−/− mBMMCs, which nevertheless showed normal responses to SCF, the natural c-Kit ligand. The absence of CysLT₂ receptors enhanced LTD₄-mediated ERK phosphorylation at every concentration tested, even lower than the dose required for c-Kit transactivation (Figure 3A,B), and shifted the dose of LTD₄ needed to induce ERK signaling by at least one log relative to the wild-type cells (n = 5, Figure 3B). CysLT₂ did not alter the LTD₄-mediated phosphorylation of c-Kit. The phospho-ERK signals in each genotype returned to baseline by 30 minutes (not shown). Thus CysLT₂ receptor–dependent control of CysLT₁ surface expression limits the strength of mitogenic ERK signaling in response to extracellular cys-LTs, and consequently alters the downstream signals responsible for cell-cycle progression. We cannot exclude the possibility that CysLT₂ receptor engagement induces activation of an ERK-specific phosphatase.

Pairs of GPCRs cotransfected into cell lines form combinations of homodimers and heterodimers that have very different signaling properties, with heterodimers regulating surface expression of some GPCRs.²⁶  Few studies have shown that these interactions and functions occur in primary cells.³²  The colocalization of CysLT₁ and CysLT₂ (Figure 1D) and the fact that CysLT₂ deficiency caused increased membrane expression of CysLT₁ (Figure 2) led us to suspect direct interactions between these receptors on MCs. We used Ab-based multiphoton FLIM^17,33  to confirm these interactions based on FRET, a quantitative measure of protein-protein interactions. FRET was consistently detected in resting LAD2 cells when they were stained with combinations of directly or indirectly labeled Abs against peptide sequences in extracellular domains (N-terminus of CysLT₂, EC3 of CysLT₁) (n = 4 experiments, as shown for one, Figure 4A). The 2 receptors could also be coimmunoprecipitated from lysates of unstimulated LAD2 cells using Abs directed against either one (Figure 4D). Collectively, these studies suggest that a fraction of these receptors constitutively exists as heterodimers on MCs. Coprecipitation was demonstrated with various combinations of Abs and did not occur with anti-BLT₁, nonspecific IgG (Figure 4D), or beads alone (not shown). We also detected significant FRET when the FLIM assay was performed using Cy3 rather than Alexa Fluor 488 as the acceptor fluorophore (Figure 4C). With all combinations of Abs, FRET was observed at the cell surface and in the nucleus, suggesting that the areas of colocalization (Figure 1D) included heterodimeric pairs. The complete lack of FRET between BLT₁ and CysLT₂ or CXCR2 and CysLT₁ receptors supports the specificity of this interaction, and combined with the FRET observed with directly labeled Abs rules out a spurious association caused by reorientation of fluorophores on the secondary Abs.

This study establishes that CysLT₂ receptors interact with CysLT₁ receptors and inhibit functional responses of the latter on MCs. We speculate that the presence of CysLT₂ receptors on MCs and other cells expressing both receptor subtypes (macrophages, eosinophils, basophils) may attenuate the strength of cys-LT–mediated signaling by limiting the formation of CysLT₁ receptor homodimers and/or controlling their surface expression, although additional mechanisms (heterologous desensitization, competition for ligand) may contribute as well. The possibility that CysLT₁/CysLT₂“counterbalance” regulates cys-LT–dependent functions in immune and inflammatory responses is supported by the observation that the presence of a coding sequence variant of the CysLT₂ receptor that exhibits reduced ligand affinity is associated with substantially increased rates of atopy and asthma in human subjects.³⁴  The fact that deletion of either receptor totally abrogates cys-LT–mediated cutaneous edema⁴  suggests that CysLT₁/CysLT₂ heterodimers may regulate vascular responses to cys-LTs. Moreover, there may be circumstances where effector responses that are CysLT₂-receptor dependent are dampened by CysLT₁ receptors, as suggested by the fact that Cysltr1^−/− mice show increases in bleomycin-induced pulmonary fibrosis relative to wild-type controls,³⁵  whereas Cysltr2^−/− mice are protected from this injury.⁷  Cys-LT–regulated cellular and tissue responses via heterodimers likely vary depending on the cell type in which the receptors are expressed, the relative abundance of each receptor, and the corresponding cell type–specific signaling pathways involved for each.

This work was supported by National Institutes of Health Grants AI-48802, AI-52353, AI-31599, HL-36110, and EB-00768, and by grants from the Charles Dana Foundation and the Vinik Family Fund for Research in Allergic Diseases.

National Institutes of Health

Contribution: Y.J. performed all of the functional analyses; L.A.B. carried out the FLIM experiments; B.J.B. supervised the FLIM and interpreted the FRET results; Y.K. developed the critical gene-targeted mice; J.A.B. was responsible for the overall conception of the study and composition of the paper.

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

Correspondence: Joshua A. Boyce, One Jimmy Fund Way, Smith Bldg Rm 626, Boston, MA 02115; e-mail: jboyce@rics.bwh.harvard.edu.