New Regulatory Roles of Galectin-3 in High-Affinity IgE Receptor Signaling

Abstract

Aggregation of the high-affinity receptor for IgE (FcεRI) in mast cells initiates activation events that lead to degranulation and release of inflammatory mediators. To better understand the signaling pathways and genes involved in mast cell activation, we developed a high-throughput mast cell degranulation assay suitable for RNA interference experiments using lentivirus-based short hairpin RNA (shRNA) delivery. We tested 432 shRNAs specific for 144 selected genes for effects on FcεRI-mediated mast cell degranulation and identified 15 potential regulators. In further studies, we focused on galectin-3 (Gal3), identified in this study as a negative regulator of mast cell degranulation. FcεRI-activated cells with Gal3 knockdown exhibited upregulated tyrosine phosphorylation of spleen tyrosine kinase and several other signal transduction molecules and enhanced calcium response. We show that Gal3 promotes internalization of IgE-FcεRI complexes; this may be related to our finding that Gal3 is a positive regulator of FcεRI ubiquitination. Furthermore, we found that Gal3 facilitates mast cell adhesion and motility on fibronectin but negatively regulates antigen-induced chemotaxis. The combined data indicate that Gal3 is involved in both positive and negative regulation of FcεRI-mediated signaling events in mast cells.

INTRODUCTION

Mast cells are important immune cells involved in multiple biological processes (1, 2). Under pathological conditions, they are responsible for IgE-mediated hyperreactivity and participate in severe diseases, such as allergy and asthma (3). Antigen (Ag)-mediated mast cell activation leads to the release of secretory granules containing a variety of preformed mediators (e.g., histamine and proteases), de novo synthesis of cytokines and chemokines, and enhanced production of arachidonic acid metabolites (4, 5). The principal surface receptor involved in mast cell activation is the high-affinity receptor for IgE (FcεRI), which belongs to the family of multichain immune recognition receptors. FcεRI is a tetrameric complex formed by an IgE-binding α subunit, a signal-amplifying β subunit, and a homodimer of disulfide-linked γ subunits. Each FcεRI β and γ subunit contains one immunoreceptor tyrosine-based activation motif (ITAM), which, after tyrosine phosphorylation, serves as a docking site for other signaling molecules, such as the SRC family kinase LYN or spleen tyrosine kinase (SYK). These two enzymes, together with other kinases, then phosphorylate various adaptor proteins, including linker of activated T cells 1 (LAT1) and LAT2 (also known as non-T cell activation linker [NTAL]). These adaptors are involved in activation of phospholipase Cγ (PLCγ) and subsequent signal transduction events, leading to calcium response and degranulation (6). FcεRI signaling is a complex process that depends on the magnitude of receptor aggregation and a balance between positive and negative signals that determine the extent of the response (7, 8). Although signaling pathways leading to mast cell activation have been extensively studied in recent years, they are far from being completely understood.

In recent years, RNA interference (RNAi) technology has become an indispensable tool in the elucidation of protein functions. RNAi-based high-throughput screening techniques have contributed significantly to identification of signal transduction pathway components in multiple systems (9–12). In this study, we took advantage of a lentiviral delivery method to transduce otherwise minimally transfectable mast cells and to induce knockdown (KD) of selected genes. We developed a short hairpin RNA (shRNA)-based high-throughput screening system to identify new regulators of FcεRI signaling and tested 432 shRNAs specific for 144 selected genes for their effects on FcεRI-mediated mast cell degranulation. Using this method, we identified 11 negative and 4 positive potential regulators of mast cell degranulation. Detailed analysis of one such regulator, galectin-3 (Gal3), revealed previously unrecognized functions of Gal3 in FcεRI signaling.

MATERIALS AND METHODS

Antibodies and reagents.

The following antibodies and their conjugates were used: mouse IgE monoclonal antibody (MAb) specific for 2,4,6-trinitrophenol (TNP), clone IGEL b4 1 (13), SYK-specific MAb (14), rabbit anti-IgE (15), FcεRI β subunit-specific MAb (JRK) (16), mouse IgE MAb specific for dinitrophenol (DNP) clone SPE-7 (Sigma-Aldrich), rat anti-KIT–allophycocyanin conjugate (17-1171) and hamster anti-FcεRI-α–fluorescein isothiocyanate (FITC) conjugate (eBioscience; 11-5898), rabbit anti-pSYK (2710) and mouse anti-phosphorylated c-Jun N-terminal kinase (anti-pJNK) (Cell Signaling; 9255S), rabbit anti-GRB2 (sc-255), actin (sc-8432), pAKT (sc-7985), extracellular signal-regulated kinase (ERK) (sc-93), pERK (sc-7976), CBL (sc-170), pCBL (sc-26140), pPLCγ1 (sc-12943), JNK1 (sc-571), Gal3 (sc-20157), galectin-1 (Gal1) (sc-28248), PLCγ1 (sc-81), goat anti-AKT1 (sc-1618), rat MAb specific for lysosomal-associated protein 1 (LAMP1) (sc-19992), horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG, goat anti-rabbit IgG, and donkey anti-goat IgG (Santa Cruz Biotechnology), phosphotyrosine-specific MAb PY-20–HRP conjugate (610012), rabbit antiphosphotyrosine (anti-pY) (610010), and V450-conjugated rat anti-mouse LAMP1 (560648) (BD Biosciences), mouse MAb specific for ubiquitinated proteins (FK2 clone; Affinity Research Products; PW8810), anti-β1-integrin antibodies (HM β1-1 and 9EG7; BD Pharmingen), secondary antibodies anti-rabbit, anti-mouse, and anti-rat IgG conjugated to Alexa Fluor 488 (AF488) or AF568 (Invitrogen), AF488-conjugated anti-hamster IgG (Life Technologies), Fcγ-specific anti-rat IgG (Jackson ImmunoResearch Laboratories), and Fura-2 AM- and AF488-conjugated phalloidin (Life Technologies). TNP-bovine serum albumin (BSA) conjugate (15 to 25 mol TNP/mol BSA) was produced as described previously (17). Mouse recombinant Gal3 was obtained from R&D Systems. DNP-human serum albumin (HSA) conjugate (30 to 40 mol DNP/mol HSA) and all other reagents were obtained from Sigma-Aldrich if not otherwise specified.

Mice, cells, and lentiviral transduction.

Mouse bone marrow mast cells (BMMCs) were derived from femurs and tibias of 8- to 10-week-old BALB/c mice bred, maintained, and used in accordance with the Institute of Molecular Genetics guidelines (permit number 12135/2010-17210) and national guidelines (2048/2004-1020). The cells were cultured in RPMI 1640 medium supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin, 71 μM 2-mercaptoethanol, minimum essential medium (MEM) nonessential amino acids, 0.7 mM sodium pyruvate, 2.5 mM l-glutamine, 12 mM d-glucose, recombinant mouse stem cell factor (SCF) (15 ng/ml; PeproTech EC), mouse recombinant interleukin 3 (IL-3) (20 ng/ml; PeproTech EC), and 10% fetal calf serum (FCS). A stable cell line derived from BMMCs (BMMCL [18]) was donated by M. Hibbs (Ludwig Institute for Cancer Research, Melbourne, Australia) (19). The cells exhibited strong Ca²⁺ (15) and degranulation (18) response after activation by Ag. The cells were cultured in medium as described above but in the absence of SCF. Lentiviruses for infections were prepared by mixing 1.5-ml aliquots of Opti-MEM medium (Invitrogen) with 21 μl of ViraPower Lentiviral Packaging Mix (Invitrogen), 14 μg of lentiviral construct, and 105 μl of polyethylenimine (1 mg/ml; 25 kDa; linear form; Polysciences). The mixture was incubated for 20 min at room temperature before it was added to HEK-293FT packaging cells in medium for cultivation of BMMCs in a 150-cm² tissue culture flask. Forty-eight hours later, virus-containing medium was filtered through 45-μm-pore-size nitrocellulose filters (Merck Millipore) and used directly for BMMC infection in the presence of 1 μg/ml protamine. The next day, the cells were subjected to the new virus load. Two days later, the cells were transferred to fresh medium containing 3 μg/ml puromycin (Apollo Scientific) for selection of positive transductants. A set of murine shRNAs targeting Lgals3 cloned into the pLKO.1 or pLKO_TRC005 vector (TRCN0000301479 [shRNA_1], TRCN0000054863 [shRNA_2], TRCN0000054867 [shRNA_3], TRCN0000301547 [shRNA_4], and TRCN0000301480 [shRNA_5]) was purchased from the RNAi Consortium (Broad Institute, Cambridge, MA). The cells were transduced with individual shRNAs or with a pool of shRNAs prepared by mixing shRNA_1, shRNA_2, shRNA_3, and shRNA_5, in equimolar ratios to obtain 14 μg of constructs transfected into packaging cells. The pool of shRNAs gave results similar to those with individual shRNAs but allowed us to scale up the experiments. Therefore, most of the pilot experiments were performed with individual shRNAs and confirmed with cells transduced with the shRNA pool.

Degranulation and Ca²⁺ response.

BMMCs were sensitized with TNP-specific IgE (1 μg/ml) in SCF- and IL-3-free culture medium for 16 h. The cells were then washed in buffered saline solution (BSS) (135 mM NaCl, 5 mM KCl, 1.8 mM CaCl₂, 5.6 mM glucose, 20 mM HEPES, pH 7.4) supplemented with 0.1% BSA (BSSA) and activated with Ag (TNP-BSA). The extent of degranulation was evaluated by determining the concentration of β-glucuronidase, as previously described (20). An Infinite 200M (Tecan) plate reader at 355-nm excitation and 460-nm emission wavelengths was used. In some experiments, cells were pretreated with lactose for 30 min prior to activation; lactose was also present during the activation. Alternatively, cells were incubated with recombinant Gal3 for 30 min prior to activation. For analysis of calcium response, cells were loaded with Fura-2 AM as described previously (20). Kinetic measurements of intracellular Ca²⁺ were determined by spectrofluorometry using an Infinite 200M plate reader with excitation wavelengths at 340 and 380 nm and with constant emission at 510 nm.

shRNA screening and high-throughput degranulation assay.

Genes for RNAi screening were selected on the basis of their expression in mast cells, as determined by microarray gene expression profiling using an Affymetrix Gene Titan HT MG-430 PM 24-array plate, as previously described (21), and/or a BioGPS gene portal (http://biogps.org/), together with estimated functions of their products in early signaling events/calcium signaling, cytoskeleton dynamics, cell adhesion/migration, and/or other plasma membrane functions (see Table S1 in the supplemental material). The lentivirus-based shRNA library containing 432 shRNA sequences specific for 144 genes was obtained from the RNAi Consortium (see Table S1 in the supplemental material). For shRNA screening in BMMCL, the following protocol was established. Cells were cultured in Iscove's modified Dulbecco's medium supplemented with mouse recombinant IL-3 (30 ng/ml) and 10% FCS. Each plate with samples contained in-plate negative-control shRNAs targeting irrelevant sequences, green fluorescent protein (GFP) (TRCN0000072181 and TRCN0000072186) or red fluorescent protein (RFP) (TRCN0000072212 and TRCN0000072209), and positive-control shRNAs targeting STIM1 (shRNA_5; TRCN0000175139) or SYK (shRNA_3; TRCN0000234763). Rows containing in-plate negative and positive controls were shuffled in different plates to minimize potential edge effects. On day 1, cells were seeded in 96-well flat-bottom plates (5 × 10⁵ per well) and transduced with lentiviruses at a multiplicity of infection (MOI) of 15 in medium supplemented with Polybrene at the optimal concentration of 8 μg/ml. The plates were centrifuged at 830 × g for 1 h and then incubated for 16 h at 37°C. On day 2, the medium containing viruses was changed for fresh medium, and on day 4, fresh medium supplemented with puromycin (5 μg/ml) was added. The medium containing puromycin was replenished every 3 days. The β-glucuronidase assay was performed more than 14 days after the start of puromycin selection. Before the assay, puromycin was removed for 24 h, and the cells were sensitized for 4 h with DNP-specific IgE (1 μg/ml) in medium depleted of IL-3. The cells (approximately 10⁵) were then washed in BSSA and split into two parts for analysis of nonactivated and Ag-activated samples (25 μl of cell suspension per well). Activation was triggered by addition of 25 μl of Ag (DNP-HSA; 10 ng/ml final concentration). This relatively low concentration of Ag allowed us to score both positive and negative regulatory hits. The cells were activated for 30 min at 37°C, and β-glucuronidase released from the cells was determined in 20 μl of cell supernatants by mixing with 28 μl of 4-methylumberylliferyl-β-d-glucuronide substrate (Invitrogen), 50 μM final concentration. The cell pellets were then lysed by the addition of 20 μl Triton X-100 at a final concentration of 1% in BSSA to each well and incubated at 37°C for 30 min. The total β-glucuronidase content was analyzed in 20 μl of cell lysate as described above, with the exception that a SpectraMax Gemini reader (Molecular Devices) was used. Degranulation was enumerated as the percentage of the β-glucuronidase activity in the supernatant from the total β-glucuronidase activity in the supernatant and lysate. Transduced cells were assayed to obtain a minimum of 3 replicates meeting quality control criteria for data analysis. Data were collected from two independent lentiviral transductions (runs). The extent of gene expression knockdown was analyzed in cells transduced in the second screen run using reverse transcription-quantitative PCR (RT-qPCR).

shRNA screen data analysis.

Data collected from degranulation assays were filtered using the following quality control criteria for each well: (i) β-glucuronidase activity detected in the well was above the measurement noise level (>10,000 relative fluorescence units [RFU]) and (ii) the β-glucuronidase released from nonactivated cells was less than 30% of the mean β-glucuronidase release from activated in-plate negative controls. Subsequently, the data from each well were Z-normalized to the mean degranulation of activated in-plate negative controls and are shown as a degranulation ratio. As a quality control for each plate, the following criteria were used: (i) the mean degranulation of activated in-plate negative controls reached at least 18% and (ii) the degranulation of activated in-plate SYK positive controls showed at least 1.5 times the standard deviation (SD) (on a Z-normalized scale) difference from the mean degranulation of activated in-plate negative controls. The Z-score for each shRNA was determined using the mean degranulation and SD of activated in-plate negative controls, and average Z-scores were calculated using all the replicates in each run. The statistical significance of differences in shRNA Z-scores versus the Z-score of negative controls was computed using a Mann-Whitney test. To limit potential false-negative results from the assay that showed significant variability (Z′ = 0.37), we evaluated the effects of shRNAs in two independent runs. Based on this evaluation, we selected the shRNAs with average Z-scores greater than ±1 and reproducible effects with P values of <0.05 across two runs; we also accepted P values of <0.075 across two runs with at least one P value of <0.05 to define shRNA hits. shRNAs with effects with opposite directionalities were excluded from the list of potential mast cell regulators.

RNA isolation and RT-qPCR.

RNA was isolated using an RNeasy minikit (Qiagen) when single samples were isolated. For RNA isolation from 96-well plates, a Macherey-Nagel kit for 96-well extraction was used according to the manufacturer's instructions. RNA was reverse transcribed using either an iScript cDNA synthesis kit (Bio-Rad) or Moloney murine leukemia virus (MMLV) reverse transcriptase (Invitrogen) according to the manufacturer's protocol. RT-qPCRs were performed in 384-well plates using a PCR mastermix supplemented with 0.2 M trehalose, 1 M 1,2-propanediol, and SYBR green I as described previously (22). RT-qPCRs were performed in a LightCycler 480 (Roche Diagnostics). All assays were performed at least in duplicate, and reaction mixtures in 10-μl volumes were processed under the following cycling conditions: initial 3-min denaturation at 95°C, followed by 50 cycles at 95°C for 10 s, 60°C for 20 s, and 72°C for 20 s. A melting-curve analysis was carried out from 72°C to 97°C with 0.2°C increments; threshold cycle (C_T) values for each sample were determined by automated threshold analysis. Data from shRNA-transduced cells from the RNAi screen were normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA levels of negative controls present in each 96-well plate with the transduced cells. Primer pairs specific for positive controls and hits in the screen are listed in Table S2 in the supplemental material. For analysis of cytokine and chemokine mRNAs, cells were not activated or activated with Ag (TNP-BSA; 100 ng/ml; 1 h). Genes for GAPDH, actin, and ubiquitin were used as reference genes, and the expression levels of all mRNAs were normalized to the geometric mean of the expression of the reference genes. The relative increase in the expression level of a cytokine was normalized to the level of expression by nonactivated pLKO.1 control cells in each experiment. Primer pairs used for analysis of cytokine and chemokine mRNAs have been listed elsewhere (20).

Cloning and mast cell transfection.

For experiments with transient Gal3 overexpression, a plasmid encoding human Gal3 (hGal3) (23), kindly provided by F.-T. Liu (University of California, Davis, CA), was used for amplification of hGal3 sequence using the forward primer 5′-aaagaattcgccaccATGGCAGACAATTTTTCGCT -3′ (the EcoRI restriction site is underlined, and the coding sequence is in uppercase) and the reverse primer 5′-tttggatccttTATCATGGTATATGAAGCAC-3′ (the BamHI restriction site is underlined, and the coding sequence is in uppercase). The amplified DNA was cloned, using EcoRI and BamHI restriction sites, into a pEGFP-N1 expression vector (Clontech Laboratories Inc.; 6085-1). Mouse Gal3 (mGal3) was obtained by PCR amplification from cDNA produced as described above and cloned into the pCDH-CMV-MCS-EF1-Puro (Systembio; pCDH) lentiviral expression vector using NheI and NotI restriction sites. For PCR, the following primers were used: 5′-aaaaagctagcgccaccATGGCAGACAGCTTTTCGCT-3′ (forward; the NheI restriction site is underlined) and 5′-tttttgcggccgcTTAGATCATGGCGTGGTTAG-3′ (reverse; the NotI restriction site is underlined). All the constructs were verified by DNA sequencing. A plasmid with hGal3 was introduced into BMMCs by Amaxa nucleofection according to the manufacturer′s instructions. A plasmid with mGal3 was used for production of lentiviruses as described above and introduced into the cells by lentiviral transduction followed by puromycin selection.

PGD₂ production.

IgE-sensitized BMMCs were seeded at 2 × 10⁵ cells per well in 100 μl of BSSA in 96-well flat-bottom culture plates. The cells were activated with Ag (TNP-BSA; 100 ng/ml; 25 min). Cell-free supernatants were collected and assessed for prostaglandin D₂ (PGD₂) using a competitive enzyme immunoassay according to the manufacturer's instructions (Cayman Chemicals). To be read within the range of the respective standard curves, supernatants of nonactivated and activated cells were diluted 1/10 and 1/40, respectively.

Immunoblotting and immunoprecipitation.

Whole-cell extracts from nonactivated or Ag (TNP-BSA)-activated cells were prepared by solubilizing pelleted cells in sodium dodecyl sulfate (SDS) sample buffer (24), followed by sonication and denaturation of samples at 97°C for 5 min. Proteins were size fractionated by SDS-polyacrylamide gel electrophoresis (PAGE), electrophoretically transferred onto nitrocellulose, and analyzed by immunoblotting with protein- or phosphoprotein-specific antibodies. For FcεRI immunoprecipitation, cells were pelleted and solubilized in ice-cold lysis buffer containing 25 mM Tris-HCl, pH 8.0, 140 mM NaCl, 1 mM Na₃VO₄, 2 mM EDTA, 100-fold-diluted protease inhibitor cocktail (Sigma-Aldrich; P8340), 1 mM phenylmethylsulfonyl fluoride, and 0.2% Triton X-100. After 30 min incubation on ice, the lysates were spun down (16,000 × g; 5 min; 4°C), and IgE complexes in postnuclear supernatants were immunoprecipitated with rabbit anti-IgE antibody prebound to UltraLink-immobilized protein A (Thermo Scientific). Immunoprecipitates were size fractionated as described above and immunoblotted with phosphotyrosine-specific PY-20–HRP conjugate or with antibody specific for ubiquitinated proteins. Bound primary antibodies were detected with HRP-conjugated secondary antibodies. HRP signal was detected with chemiluminescence reagent (25) and quantified with a luminescent image analyzer (LAS 3000; Fuji Photo Film). Aida software (Raytest GmbH) was used for signal quantification.

Flow cytometry analysis.

To analyze surface levels of FcεRI and KIT receptor, labeling was performed as described previously (20). For analysis of internalized IgE, IgE-sensitized BMMCs were activated or not with different concentrations of Ag (TNP-BSA) for various times at 37°C. Activation was stopped by centrifugation at 4°C (300 × g; 4 min), and samples were split into two; one half was subjected to acid stripping by a 10-min incubation in 0.5 M NaCl and 0.2 M acetic acid (pH 2.7) on ice to remove surface-accessible IgE, as previously described (26). All the samples were then fixed for 20 min in 2% paraformaldehyde, washed in PBS, and then permeabilized in 0.05% Triton X-100 for 10 min. IgE was stained using AF-conjugated anti-mouse Ig (cross-reacting with IgE) in PBS-1% BSA. After labeling, the cells were washed 3 times with PBS and analyzed with an LSRII flow cytometer (BD Biosciences). Median fluorescence intensities were determined and further processed using FlowJo software (TreeStar, Ashland, OR). To quantify surface expression of β1-integrin and its activated epitope 9EG7, BMMCs were activated or not with Ag (TNP-BSA; 100 ng/ml; 3 min). The cells were then exposed to the anti-β1-integrin antibody HM β1-1 or 9EG7 for 30 min on ice, followed by a 30-min incubation with the corresponding AF488-conjugated anti-hamster or anti-rat (Fcγ-specific) secondary antibody, respectively. After 30 min of incubation on ice, the cells were washed in ice-cold PBS and analyzed as described above. For analysis of surface LAMP1, BMMCs were activated or not with various concentrations of Ag (TNP-BSA; 10 min; 37°C). Activation was stopped by centrifugation at 4°C. The cells were then resuspended in 50 μl of PBS-1% BSA containing 200-fold-diluted V450-conjugated rat anti-mouse LAMP1 and stained on ice for 30 min. After washing with PBS, the cells were analyzed as described above.

Confocal microscopy.

Eight-well multitest microscopy slides (MP Biomedicals) were coated with CellTak (8 μl in 1 ml of PBS; BD Biosciences). The cells were left to attach to coated slides for 15 min in BSSA and then activated with Ag (TNP-BSA; 100 ng/ml; 37°C). After activation, the cells were fixed with 3% paraformaldehyde in PBS for 30 min and permeabilized with 0.1% Triton X-100 for 30 min. After washing with PBS, free binding sites were blocked with 1% BSA and subsequently labeled with specific antibodies, followed by AF-conjugated secondary antibodies. IgE was detected with AF-conjugated donkey anti-mouse Ig-specific antibody. After labeling, the cells were washed and mounted in 50% (wt/vol) glycerol in PBS, pH 8.5, supplemented with 0.1% p-phenylenediamine and 1 μg/ml Hoechst 33258 to label nuclei. For F-actin staining, cells were permeabilized with l-α-lysophosphatidylcholine (80 μg/ml) and simultaneously labeled with 10 μg/ml of AF488-conjugated phalloidin for 1 h in PBS-1% BSA. Samples were examined with a confocal laser scanning microscope (Leica TCS SP5) equipped with a 63× (numerical aperture, 1.4) oil immersion objective. In each experiment, images were acquired at identical microscope settings. Image analyses were performed using a pipeline generated in CellProfiler cell image analysis software (Broad Institute, Cambridge, MA) (27). For examination of protein distribution, 20 to 40 cells were identified using nuclei, and signals from fluorescently labeled proteins were segmented into 4 concentric rings centered on the nucleus. The fluorescent intensity of each cell was determined, and the fraction of the fluorescence in 3 inner rings (cytoplasmic) was calculated.

Cell adhesion, motility, and chemotaxis.

IgE-sensitized BMMCs were loaded with calcein-AM and transferred into wells of a 96-well plate (Thermo Scientific) coated with fibronectin at a concentration of 10 μg/ml (28). After activation with Ag (TNP-BSA; 100 ng/ml; 30 min), unbound cells were washed out using a microplate washer (HydroSpeed; Tecan), and bound cells were determined using an Infinite 200M fluorometer with excitation and emission filters at 485 nm and 538 nm, respectively. Analysis of cell motility on fibronectin after activation with PGE₂ was performed as described previously (29). Ag (TNP-BSA; 250 ng/ml)- or PGE₂ (100 nM)-mediated chemotactic responses were evaluated using 24-well Transwell chambers (Corning) with 8-μm polycarbonate filters, as described previously (28). Cells migrating into lower compartments within the 8-h incubation period were counted using an Accuri C6 flow cytometer (Becton Dickinson). Cell migration toward PGE₂ (100 ng/ml) was performed in the same way, except that the cells were not sensitized with IgE. In some experiments, the supernatants from IgE-sensitized and Ag-activated (TNP-BSA; 100 ng/ml; 6 h) BMMCs at a concentration of 2 × 10⁶/ml were used as a source of chemoattractants for unsensitized cells.

Data analysis.

All experiments were repeated independently as indicated in the figure legends. Averages or representative results from all repeats are shown. Unless otherwise indicated, the significance of intergroup differences was evaluated by Student's t test.

RESULTS

Search for new regulators of FcεRI signaling using an RNAi-based high-throughput degranulation assay.

To identify new regulators of FcεRI signaling, we developed an shRNA high-throughput mast cell degranulation assay using BMMCL. The cells were transduced with shRNA lentiviruses, followed by puromycin selection (Fig. 1A). Our optimized protocol included infection of the cells at an MOI of 15 in the presence of Polybrene at 8 μg/ml. Two weeks after selection, the cells were sensitized with Ag-specific IgE and activated with Ag, and the amount of β-glucuronidase released from preformed secretory granules was assessed by enzymatic assay (Fig. 1B). shRNAs targeting GFP and RFP were used as negative controls, and shRNAs targeting STIM1 and SYK were used as positive controls. STIM1 shRNAs led to 50 to 80% knockdown efficiency and up to a 5-fold decrease in degranulation, and SYK shRNAs led to more than 70% knockdown of gene expression and up to a 5-fold decrease in degranulation (Fig. 1C and D). Notably, the STIM1 shRNAs that had smaller effects on STIM1 expression levels also had poorer effects on mast cell degranulation. SYK shRNAs were used to calculate the Z′ factor as 0.37, which is considered acceptable for screening for RNAi cell-based assays (30).

FIG 1.

High-throughput degranulation assay protocol and shRNA screening in mast cells. (A) Time axis of lentiviral transduction of BMMCL, selection in puromycin (Puro) and degranulation assay in one transduction run. (B) Assay design and protocol for high-throughput degranulation assay. Appropriate numbers of cells (∼10⁵) were transferred from the culture plate (transduced cells) to a new plate and sensitized with IgE (IgE-sensitized cells). The cells were split into two parts for nonactivated samples (Ag⁻) or Ag-activated samples (shaded; Ag⁺). The positions of 4 negative controls (blue), positive-control shRNAs targeting STIM1 (red), and SYK (green) are indicated. After 30 min of activation, β-glucuronidase (β-glucur.) was measured in cell-free supernatants and in lysates from cell pellets. (C) Degranulation of control shRNAs. Representative results from pilot experiments are shown. The cells were transduced with negative-control shRNAs targeting irrelevant sequences (2 GFP and 2 RFP; blue) or positive-control shRNAs targeting STIM1 (red) or SYK (green). The Z′ factor for the screening assay is also shown. (D) RT-qPCR analysis of gene KD of STIM1 and SYK from cells in panel C. (E) Schematic workflow of data collection, criteria for data quality control (QC), and hit calling. (F) Overview of all individual shRNAs in the second run of the screen ranked according to percent inhibition or activation in the degranulation assay. The red bars indicate the positions of all shRNAs targeting Lgals3.

For screening, we selected 144 genes that are expressed in mast cells and could be involved in FcεRI signaling. The list included genes with potential roles in early signaling events/calcium signaling, cytoskeleton dynamics, cell adhesion/migration, and/or other plasma membrane functions. As shown in Table S1 in the supplemental material, we tested 432 lentivirus-encoded shRNAs, with 3 different shRNAs targeting each gene. As the Z′ factor of the assay did not pass the 0.5 threshold to be considered an excellent screening assay, we collected data from two independent transduction runs. In each run, we collected at least 3 degranulation assays that passed quality control criteria (see Materials and Methods) to increase confidence in the identification of potential mast cell regulators. To limit potential false-negative results from the assay that showed significant variability (Z′ factor = 0.37), we used a P value of <0.05 (Mann-Whitney test) across two runs, and we also accepted a P value of <0.075 across two runs with at least one P value of <0.05 for shRNAs with an average Z-score greater than ±1. To limit potential false-positive results, the effect of each shRNA had to pass statistical significance using the above P values in the two independent runs. Using these criteria, we identified potential regulators with consistent effects on shRNA levels on mast cell degranulation: 10 hits using a P value of <0.05 across two runs and 5 hits using a P value of <0.075 across two runs with at least one P value of <0.05. Data collection and the subsequent processes of quality control and hit calling are summarized in Fig. 1E. Average Z-scores and average degranulation ratios calculated for all shRNAs in both screen runs, together with an analysis of gene expression of the identified hits from the second run, are shown in Table S1 in the supplemental material. Among 15 potential regulators, 11 were negative regulators and 4 were positive regulators of cell degranulation (Table 1). The effects of all shRNAs on degranulation in the second screen run are shown in Fig. 1F. Among the negative regulatory genes, the KD of the Lgals3 gene, encoding Gal3, increased cell degranulation in all 3 tested shRNAs, displaying 1.67 ± 0.03-fold change in mast cell degranulation (P < 0.05). This was a surprising finding, because previous studies using BMMCs from mice with Gal3 knockout (KO) suggested that Gal3 is a positive regulator of mast cell degranulation (31). We therefore focused on deciphering the mechanism of the negative regulatory role of Gal3 in mast cell activation.

TABLE 1.

Potential regulators of mast cell degranulation identified in shRNA screening in mast cells

Symbol	NCBI gene ID	GenBank accession no.	Official full name	Fold change in degranulation (mean ± SD)a
Negative regulators
Adora2b	11541	NM_007413.2	Adenosine A2b receptor	1.75 ± 0.76
Basp1	70350	NM_027395.1	Brain acid-soluble protein 1	1.4 ± 0.36
Cyfip2	76884	NM_133769.1	Cytoplasmic FMR1-interacting protein 2	1.26 ± 0.19
Diap1	13367	NM_007858.2	Diaphanous homolog 1	1.65 ± 0.42
Lgals3	16854	NM_010705.3	Lectin, galactose binding, soluble 3	1.67 ± 0.03
Lpxn	107321	NM_134152.1	Leupaxin	1.38 ± 0.35
Mctp1	78771	XM_127419.3	Multiple C2 domains, transmembrane 1	1.6 ± 0.36
Plscr1	22038	NM_011636.2	Phospholipid scramblase 1	1.65 ± 0.32
S100a10	20194	NM_009112.2	S100 calcium binding protein A10	1.52 ± 0.27
Spn	20737	NM_009259.3	Sialophorin	1.33 ± 0.07
Tjp1	21872	NM_009386.1	Tight-junction protein 1	1.27 ± 0.06
Positive regulators
Add2	11519	NM_013458.2	Adducing 2 (beta)	0.67 ± 0.1
Coro7	78885	NM_030205.2	Coronin 7	0.46 ± 0.18
Dock8	76088	NM_028785.3	Dedicator of cytokinesis 8	0.52 ± 0.27
Ier2	15936	NM_010499.4	Immediate-early response 2	0.42 ± 0.17

^aMost effective shRNAs for each gene across both runs (P < 0.05). The following genes had nonsignificant or inconsistent effects on mast cell degranulation: genes encoding Adcy7, Adora3, Ahnak, Akap2, Alox5ap, Arap3, Arhgdia, Arhgdib, Arhgef1, Avil, Bcap31, Bsg, Cab39, Cd151, Cd200r1, Cd200r3, Cd200r4, Cd244, Cd274, Cd276, Cd300a, Cd300lb, Cd93, Cd96, Cd97, Cd99, Cdc42, Cdh2, Cfl1, Coro1a, Cotl1, Csk, Cyfip1, Cytip, Darc, Dock10, Dock11, Dstn, Dusp19, Efhd1, Efhd2, Enah, Fcer1a, Fcer1g, Fermt3, Fert2, Fgfr1op2, Frmd4a, Git2, Gng5, Gp1ba, Gp49a, Gpr34, Grb2, Ier3, Ifitm1, Ifitm2, Inpp5b, Inpp5d, Itgb7, Kcnn4, Kit, Lamp2, Laptm5, Lasp1, Lgals1, Lims1, Lpar6, Lst1, March8, Ms4a2, Msn, Mtss1, Nckap1, Nckap1l, Nckipsd, Ninj1, P2rx7, P2ry14, Pdcd1lg2, Pdlim5, Pdzk1ip1, Pecam1, Pfn1, Plec, Plek, Prnp, Ptger3, Ptp4a1, Ptp4a2, Ptp4a3, Ptpn1, Ptpn11, Ptpn12, Ptpn18, Ptpn2, Ptpn3, Ptpn9, Rab1, Rab27b, Rac2, Rap1a, Rassf5, Rdx, Rnf128, Rock1, S100a6, Scin, Sema4d, Sh3bgrl3, Shc1, Sla, Spna2, Spnb2, Svil, Tagln2, Tjp2, Tln1, Tmem123, Tmem158, Tmem66, Tspan13, Tspan31, Tspan4, Tspan5, Tspan8, Twf2, Tyrobp, and Vcl. For details, see Table S1 in the supplemental material.

Negative regulatory role of Gal3 in FcεRI-mediated mast cell degranulation, cytokine expression, and PGD₂ production.

To confirm the negative regulatory role of Gal3 in mast cell degranulation, we expanded the number of shRNAs and analyzed their effects on Gal3 expression (Fig. 2A and B) and degranulation (Fig. 2C) in BMMCL. The results were compared to those obtained with cells transduced with an empty vector (pLKO.1). The data obtained showed that inhibition of degranulation was mediated by different shRNAs targeting Gal3. Next, we transduced BMMCs with individual shRNAs or pooled shRNA_1, shRNA_2, shRNA_3, and shRNA_5. Because of the similarity of the data obtained with these two approaches, we pooled the data and labeled the cells Gal3 KD. BMMCs with Gal3 KD exhibited ≥70% decrease of Gal3 expression at the level of both mRNA (Fig. 2D) and protein (Fig. 2E). They also exhibited enhanced Ag-induced degranulation (Fig. 2F). The expression of other galectins was not affected by transduction with Gal3-targeted shRNAs (see Fig. S1A in the supplemental material), with the exception of Gal1 mRNA, which was upregulated up to 2-fold at the mRNA level but not at the protein level (see Fig. S1B in the supplemental material). When the cells were activated with thapsigargin, an inhibitor of sarcoplasmic/endoplasmic reticulum Ca²⁺ ATPase, the differences in degranulation between BMMCs with Gal3 KD and pLKO.1 controls were not significant, suggesting that Gal3 is not involved in the mechanism of granule release and acts upstream of calcium signaling (Fig. 2G). BMMCs with Gal3 KD had normal surface expression of FcεRI and KIT (Fig. 2H), suggesting that lentiviral transduction, puromycin selection, and Gal3 KD had no effect on the expression of these two receptors in nonactivated mast cells.

FIG 2.

Upregulated Ag-induced activation of BMMCs with Gal3 KD. (A) RT-qPCR quantification of Gal3 mRNA expression in BMMCL transduced with shRNAs targeting Gal3 or pLKO.1 control vector. The data are representatives of 3 independent experiments. (B) Quantification of Gal3 protein expression by immunoblotting (IB) in whole-cell lysates prepared from cells transduced as in panel A. The numbers under the blot indicate the relative amounts of Gal3 normalized to the relative amounts of LYN, used as a loading control, and Gal3 in pLKO.1 control cells. The results are representative of 3 independent experiments. (C) BMMCs, transduced as in panel A, were sensitized with IgE and activated for 30 min with various concentrations of Ag, and β-glucuronidase release was determined. The data are representative of the results of 3 independent experiments. (D) RT-qPCR quantification of Gal3 mRNA expression in BMMCs with Gal3 KD and in pLKO.1 control cells. Means and standard errors (SE) are shown (n = 7). (E) Quantification of Gal3 protein expression by IB in whole-cell lysates prepared from BMMCs transduced as in panel D; nontransduced cells are labeled as wild type (WT). The numbers under the blots are as in panel B; however, actin was used as a loading control. The data are representative of the results of 3 independent experiments. (F) BMMCs transduced as in panel D were sensitized with IgE and activated with various concentrations of Ag for 30 min, and β-glucuronidase was determined. Means ± SE are shown (n = 7). (G) BMMCs transduced as in panel D were activated for 30 min with 1 μM thapsigargin, and β-glucuronidase release was determined. Means and SE are shown (n = 7). (H) Representative flow cytometry plots showing expression of surface FcεRI and KIT in BMMCs transduced as in panel D. Nontransfected cells are labeled as WT. The control fluorescence profile of BMMCs is denoted nonlabeled. (I) BMMCs transduced as in panel D were sensitized with IgE and activated with various concentrations of Ag for 10 min. Increased surface expression of LAMP1 was determined by flow cytometry. Means ± SE are shown (n = 4). (J) BMMCs were transfected with empty pEGFP vector or pEGFP-hGal3 to overexpress Gal3. Gal3 protein expression was analyzed by IB in whole-cell lysates with anti-Gal3 antibody. Actin is shown as a loading control. The arrows indicate the positions of endogenous Gal3, exogenous EGFP-hGal3 fusion protein, and actin. Nontransfected cells are labeled as WT. (K) BMMCs were transfected as in panel J, and surface expression of LAMP1 was determined by flow cytometry. Means ± SE are shown (n = 6). (L) Cytokine and chemokine mRNA expression. BMMCs were transduced as in panel D, sensitized with IgE, and activated with Ag (100 ng/ml) for 1 h. mRNAs encoding TNF-α, IL-6, IL-13, and CCL3 were quantified by RT-qPCR. The fold increase in mRNA expression after Ag stimulation was normalized to nonactivated pLKO.1 control cells. Means and SE are shown (n = 7). (M) PGD₂ production. BMMCs were transduced as in panel D, sensitized with IgE, and activated with Ag (100 ng/ml) for 25 min. The amount of PGD₂ in cell-free supernatants was determined by immunoassay. Means and SE are shown (n = 5). *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Cell activation and degranulation are accompanied by enhanced surface expression of LAMP1 secretory granule/lysosomal marker (32). In agreement with data evaluating β-glucuronidase release, Ag (TNP-BSA)-activated BMMCs with Gal3 KD showed significantly higher levels of surface LAMP1 than pLKO.1 controls (Fig. 2I). Similarly, we found that cells with Gal3 KD activated with DNP-HSA showed more LAMP1 expression than control cells, especially when activated with low concentrations of Ag (see Fig. S2 in the supplemental material). To further confirm the role of Gal3 in mast cell activation, we next examined mast cell activation in cells overexpressing Gal3. BMMCs were transiently transfected with a plasmid encoding Gal3 fused to enhanced green fluorescent protein (EGFP) (pEGFP-hGal3) or a control plasmid (Fig. 2J), and LAMP1 surface expression was analyzed in EGFP-positive cells. The data presented in Fig. 2K show that when activated with Ag (10 to 500 ng/ml), Gal3-overexpressing cells exhibited significantly decreased LAMP1 surface expression compared with control cells; this supports our findings of a negative regulatory role of Gal3 in FcεRI signaling.

Gal3 in BMMCs showed mainly cytoplasmic localization, with some in the nucleus and on the plasma membrane (see Fig. S3 in the supplemental material). Upon Ag activation, Gal3 localization did not change. To test whether Gal3 acts on mast cell degranulation via its lectin function, we attempted to block its lectin activity with lactose. However, as shown in Fig. S4A in the supplemental material, lactose had no effect on Ag-induced degranulation. Alternatively, BMMCs with Gal3 KD were incubated or not for 30 min with recombinant mouse Gal3 prior to activation. The data in Fig. S4B in the supplemental material show that externally added Gal3 was ineffective in restoring the Gal3 KD phenotype.

Mast cells are potent producers of cytokines, chemokines, and other soluble mediators involved in immune responses. Therefore, we also investigated the role of Gal3 in these processes. Compared with control pLKO.1 cells, BMMCs with Gal3 KD produced significantly more mRNA for tumor necrosis factor alpha (TNF-α), IL-6, and IL-13, as well as chemokine C-C motif ligand 3 (CCL3), upon Ag stimulation (Fig. 2L). Production of PGD₂ was also significantly increased in BMMCs with Gal3 KD (Fig. 2M).

Together, our data from gene expression knockdown and overexpression experiments in primary mast cells are consistent with the results from our shRNA screen in BMMCL and confirm the negative regulatory role of endogenous Gal3 in Ag-induced mast cell activation.

Negative regulatory role of Gal3 in F-actin dynamics, calcium response, and phosphorylation of signal transduction molecules.

Consistent with increased degranulation, BMMCs with Gal3 KD exhibited significantly higher calcium response to Ag than pLKO.1 control cells (Fig. 3A). To determine whether Gal3 mediates its effect on mast cell degranulation by affecting negative regulation mediated by the actin cytoskeleton, we analyzed the F-actin levels in Ag-stimulated BMMCs with Gal3 KD and control cells by confocal microscopy. As shown in Fig. 3B and C, Ag-stimulated BMMCs with Gal3 KD exhibited reduced levels of F-actin compared to control cells.

FIG 3.

Calcium mobilization and F-actin dynamics. (A) Intracellular calcium mobilization. BMMCs with Gal3 KD or pLKO.1 control cells were sensitized with IgE and loaded with Fura 2-AM. The cells were then activated with Ag at a concentration of 10 ng/ml (left) or 500 ng/ml (right), and increases in calcium response were monitored by measuring fluorescence ratios. The arrows indicate the time points of Ag addition, and the bars above the curves represent statistically significant differences between cells with Gal3 KD and pLKO.1 controls. Means ± SE are shown (n = 6). (B) F-actin response. IgE-sensitized BMMCs as in panel A were activated with Ag (100 ng/ml) for the indicated times. After fixation, the cells were permeabilized and stained for F-actin with AF488-conjugated phalloidin. The images were acquired by confocal microscopy. Scale bars, 10 μm. (C) Statistical evaluation of F-actin fluorescence in cells, as in panel B; 5 to 20 cells were analyzed in each experiment. Means and SE were calculated from the results of 4 independent experiments. *, P < 0.05.

Next, we examined global tyrosine phosphorylation using confocal microscopy. Nonactivated BMMCs with Gal3 KD and pLKO.1 control cells showed comparable levels and localization of tyrosine-phosphorylated proteins (Fig. 4A, left, and B). However, Ag-activated cells with Gal3 KD showed more intense phosphotyrosine staining than control pLKO.1 cells (Fig. 4A, right, and B). We also examined phosphorylation of key signaling molecules in the FcεRI pathway by immunoblotting. Analysis with antibodies specific for phosphorylated SYK, PLCγ1, and AKT (Fig. 4C) after 5 min of activation with various concentrations of Ag (10 or 100 ng/ml) and antibodies specific for pJNK (Fig. 4D) after 15 min of activation with Ag (100 ng/ml) showed significantly upregulated phosphorylation of the target proteins in activated BMMCs with Gal3 KD compared with pLKO.1 controls. No significant differences in phosphorylation of ERK were found between these two cell types after FcεRI triggering (Fig. 4C). In nonactivated cells with Gal3 KD, the phosphorylation of AKT and ERK was significantly increased, suggesting that Gal3 also affects signaling pathways in the resting cells. These results demonstrate that Gal3 is involved in downregulation of early FcεRI activation events, including phosphorylation of key signaling molecules and calcium response.

FIG 4.

Early activation events in BMMCs with Gal3 KD. (A) Global tyrosine phosphorylation in BMMCs with Gal3 KD (top) or pLKO.1 controls (bottom). Cells were sensitized with IgE and either not activated (Control) or Ag activated (Ag) (100 ng/ml) for 5 min. After activation, the cells were fixed, permeabilized, and stained with anti-IgE, followed by AF568-mouse Ig-specific secondary-antibody conjugate (red) and anti-pY antibody followed by AF488-rabbit Ig-specific secondary antibody conjugate (green). Bars, 2 μm. (B) Evaluation of differences in pY fluorescence in cells as in panel A. Each dot represents one cell. Means are indicated by red bars. A statistically significant difference between BMMCs with Gal3 KD and pLKO.1 controls is indicated. (C) Phosphoprotein analysis. BMMCs with Gal3 KD or pLKO.1 controls were sensitized with IgE and activated with various concentrations of Ag for 5 min. SDS-PAGE-separated whole-cell lysates were analyzed with various phosphosite-specific antibodies (pPLCγ1, pSYK, pAKT, and pERK), as indicated. For loading controls the membranes were analyzed by immunoblotting with the corresponding protein-specific antibodies. GRB2 is shown as another loading control. (Left) Representative immunoblots from at least 3 experiments. (Right) Densitometry analysis of the corresponding immunoblots, in which signals from tyrosine-phosphorylated proteins in activated cells were normalized to the signal in nonactivated pLKO.1 control cells and loading control proteins. Means and SE were calculated from the results of three independent experiments. The statistical significance of differences between cells with Gal3 KD and pLKO.1 controls is also shown. (D) BMMCs were processed and analyzed as in panel C, with the modification that cells were activated with Ag (100 ng/ml) for various times. Phosphorylation of JNK1/2 was determined by immunoblotting with pJNK antibody. For loading controls, the membranes were analyzed by immunoblotting with JNK1- and GRB2-specific antibodies. Means ± SE are shown (n = 3). *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Gal3 regulates IgE internalization and trafficking to endolysosomes.

Previous studies have shown that T cells lacking Gal3 exhibited enhanced T cell receptor (TCR) activation events and that these changes were connected to the impaired downregulation and internalization of TCRs (33). FcεRI is rapidly internalized upon its aggregation with IgE-multivalent Ag complexes (34). Therefore, we next compared internalization of the Ag-aggregated FcεRI in BMMCs with Gal3 KD and control cells. To quantitatively determine intracellular IgE, we used flow cytometry to measure the fraction of total IgE remaining on the cell surface after acid wash stripping. We found that upon activation with Ag at concentrations of 500 ng/ml and 1 μg/ml, BMMCs with Gal3 KD showed less acid-resistant and intracellular IgE than pLKO.1 controls (Fig. 5A).

FIG 5.

Gal3-regulated IgE internalization and trafficking to endolysosomes. (A) IgE internalization in BMMCs with Gal3 KD and control pLKO.1 cells. The cells were sensitized with IgE and activated with various concentrations of Ag (100 ng/ml, 500 ng/ml, or 1 μg/ml). At various time intervals after triggering (0, 5, 15, or 30 min), surface IgE was removed by acid stripping, and the cells were fixed and permeabilized. IgE levels were quantified with anti-IgE antibody by flow cytometry. Means ± SEs from the results of 3 independent experiments are shown. (B) Representative confocal images of IgE and LAMP1 localization in BMMCs transduced as in panel A (left, Gal3 KD; right, pLKO.1 controls). The cells were sensitized with IgE and not activated (0 min) or activated with Ag (100 ng/ml) for various times (5, 15, or 30 min), fixed, permeabilized, and stained with anti-IgE, followed by AF568–mouse Ig-specific secondary antibody conjugate (red) and anti-LAMP1-specific antibody and then by AF488-secondary antibody conjugate (green). Bars, 5 μm. (C) Quantitative analysis of IgE distribution in nonactivated (0 min) and Ag-activated (30 min) cells from experiments, as in panel B. IgE fluorescence was analyzed as described in Materials and Methods and plotted as a percentage of cytoplasmic IgE. Each point represents one cell from 2 independent experiments; the red bars represent the means. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Internalized IgE-FcεRI complexes have been shown to be delivered into LAMP1-positive endolysosomes (35). To determine the involvement of Gal3 in this process, we next examined trafficking of IgE-FcεRI complexes. In nonactivated cells, IgE staining was almost exclusively associated with plasma membranes in both BMMCs with Gal3 KD and pLKO.1 control cells (Fig. 5B). Thirty minutes after Ag activation, a fraction of IgE in pLKO.1 cells moved into vesicular structures. In contrast, BMMCs with Gal3 KD showed IgE staining with a diffuse pattern, with IgE predominantly localized near the plasma membrane. We quantified the amount of IgE localized in the vesicles inside the cells after 30 min of Ag activation and found that significantly less IgE was associated with intracellular structures in cells with Gal3 KD than in controls (Fig. 5C). The combined data indicate that Gal3 facilitates internalization of IgE-FcεRI complexes upon Ag triggering and facilitates the trafficking of internalized IgE into the LAMP1-positive lysosomal compartments.

Gal3 is involved in regulation of FcεRI ubiquitination.

The internalization and sorting of FcεRI into endosomal compartments is a ubiquitin-dependent process (36). We speculated that aberrant IgE internalization and trafficking in BMMCs with Gal3 KD could be coupled with impaired ubiquitination of FcεRI. To test this hypothesis, we immunoprecipitated IgE-FcεRI complexes from BMMCs with Gal3 KD or pLKO.1 control cells and evaluated the ubiquitination of the FcεRI β and γ subunits upon Ag activation. We found that FcεRI β and γ subunits in BMMCs with Gal3 KD had less ubiquitination than those in control cells (Fig. 6A to C). We also prepared BMMCs overexpressing mGal3-Myc after lentiviral transduction and selection of puromycin-resistant cells, but the level of mGal3-Myc expression was rather moderate in several independent experiments and did not result in significant differences in ubiquitination of FcεRI β and γ subunits in these cells compared to cells transduced with empty pCDH vector (Fig. 6D to F). Interestingly, no significant differences in tyrosine phosphorylation of β and γ FcεRI subunits were found between BMMCs with Gal3 KD and pLKO.1 control cells (Fig. 6G and H). CBL, an E3-ubiquitin ligase, is rapidly phosphorylated after Ag stimulation and is involved in FcεRI ubiquitination (37, 38). Therefore, we also investigated whether phosphorylation of CBL is dependent on the presence of Gal3. As shown in Fig. 6I to J, similar levels of CBL phosphorylation were found in cells with Gal3 KD and pLKO.1 controls. Thus, changes in phosphorylation of CBL do not explain the observed defect in FcεRI ubiquitination in BMMCs with Gal3 KD.

FIG 6.

Gal3 promotes ubiquitination of FcεRI β and γ subunits without affecting phosphorylation of FcεRI and CBL. (A to C) BMMCs with Gal3 KD or pLKO.1 controls were sensitized with IgE and activated with Ag (100 ng/ml) for the indicated times. (A) IgE-FcεRI complexes were immunoprecipitated with IgE-specific polyclonal antiserum. Immunoprecipitates (IPs) were size separated by SDS-PAGE, and ubiquitination was determined by IB with an antibody specific for ubiquitinated proteins (Ub). The positions of molecular weight standards and the FcεRI β and γ subunits are indicated on the left and right, respectively. (B and C) Quantification of FcεRI β (B) and γ (C) subunit ubiquitination from data in panel A. Densitometry data for the subunit ubiquitination were normalized to the number of β subunits used as loading controls and nonactivated pLKO.1 controls. (D to F) BMMCs were transduced with mGal3-Myc or empty pCDH vector, puromycin-resistant cells were selected, and FcεRI subunit ubiquitination was analyzed as in panel A. Expression of Gal3 (endogenous and exogenous) and mGal3-Myc was detected with anti-Gal3 and anti-Myc antibodies, respectively. GRB2 was also analyzed as another loading control. (E and F) Quantification of FcεRI β (E) and γ (F) subunit ubiquitination from data in panel D was performed as described above, except that densitometry data were normalized to nonactivated empty pCDH controls. (G and H) The FcεRI IPs were obtained as in panel A, and samples were analyzed for tyrosine phosphorylation of the FcεRI β and γ subunits with the phosphotyrosine-specific antibody PY-20–HRP (pY). Quantification of the subunit phosphorylation was performed by densitometry of the blots as in panel G, and data were normalized to the number of FcεRI β subunits and pLKO.1 controls. (I and J) Phosphorylation of CBL in BMMCs with Gal3 KD or pLKO.1 controls. Cells were activated as in panel A, and whole-cell lysates were size separated by SDS-PAGE and analyzed by immunoblotting with pCBL-specific antibody; because of poor performance of anti-CBL, GRB2 was used as a loading control (I). Quantification of pCBL was performed by densitometry of the blots as in panel I, and data were normalized to the amount of GRB2 and nonactivated pLKO.1 controls. (B, C, E, F, H, and J) Means ± SE calculated from the results of at least 3 independent experiments. *, P < 0.05; **, P < 0.01.

Regulatory role of Gal3 in mast cell adhesion, motility, and chemotaxis.

Gal3 affects cell adhesion, motility, and migration in various cell types (39, 40). Previous studies have shown adhesion of BMMCs to fibronectin (41). Therefore, we investigated whether Gal3 KD interferes with the adhesion and motility of mast cells on fibronectin-coated surfaces. We found that Gal3 had no effect on binding of nonactivated cells to fibronectin-coated surfaces (Fig. 7A). However, after Ag activation, cells with Gal3 KD showed significantly decreased binding to fibronectin compared to pLKO.1 control cells.

FIG 7.

Involvement of Gal3 in BMMC motility, adhesion, and chemotaxis. (A) Cell adhesion to fibronectin-coated surfaces was analyzed in BMMCs with Gal3 KD or pLKO.1 controls. The cells were sensitized with IgE, loaded with calcein, and left to attach to a fibronectin-coated surface for 30 min. After activation with Ag (100 ng/ml) or PGE₂ (100 nM) for 30 min, the fraction of attached cells was determined. Means and SE are shown (n = 5). (B) Activation of β1-integrin in cells sensitized with IgE and, after washing, activated with Ag (100 ng/ml) or PGE₂ (100 nM) for 3 min. Surface expression of activated β1-integrin was determined with an antibody specific for the open conformation of β1-integrin and by flow cytometry. Means and SE are shown (n = 7). (C) Detection of total surface β1-integrin in cells sensitized with IgE and activated with Ag (100 ng/ml) or PGE₂ (100 nM) for 3 min. Surface expression of β1-integrin was determined by flow cytometry. Means and SE are shown (n = 7). (D) Motility was evaluated in cells attached to a fibronectin-coated surface for 15 min and activated with PGE₂ (100 nM). Cell movement was recorded at 1-min intervals. Representative cell trajectories over 30 min are indicated in colors. Bars, 50 μm. (E) Quantification of average velocity in relative units (pixels per min) in cells analyzed as in panel B. Each point represents one cell; the data are from two independent experiments; the red bars represent means. (F and G) Migration of IgE-sensitized BMMCs with Gal3 KD or pLKO.1 controls toward Ag (100 ng/ml) (F) or PGE₂ (100 nM) (G). (H) Migration of nonsensitized BMMCs with Gal3 KD or pLKO.1 control toward supernatants obtained from IgE-sensitized BMMCs with Gal3 KD or pLKO.1 controls activated for 6 h with Ag (TNP-BSA; 100 ng/ml) or without Ag (Control). Means and SE are shown (n = 3). *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Upon Ag stimulation of mast cells, β1-integrin, a receptor for fibronectin, adopts an activated conformation with higher affinity for fibronectin (42). To determine whether defects in adhesion and motility of BMMCs with Gal3 KD on fibronectin are related to changes in β1-integrin activation, we tested the levels of activated β1-integrin by flow cytometry with antibody against the 9EG7 epitope. Our data show that nonactivated BMMCs with Gal3 KD expressed almost doubled the levels of activated β1-integrin compared with control cells (Fig. 7B). After Ag activation, levels of activated β1-integrin increased in pLKO.1 control cells but remained unchanged in BMMCs with Gal3 KD. One explanation of these findings could be deregulated β1-integrin internalization in the absence of Gal3. Therefore, we examined total β1-integrin on the cell surface. We found that BMMCs with Gal3 KD expressed significantly larger amounts of β1-integrin on the plasma membrane and, when activated with Ag, β1-integrin rapidly internalized to a greater extent in BMMCs with Gal3 KD than in control pLKO.1 cells (Fig. 7C).

To examine motility, we activated the cells with PGE₂, a potent inducer of mast cell migration (43). Figure 7D shows representative trajectories of PGE₂-activated BMMCs with Gal3 KD or control cells recorded in 1-min intervals for a total of 30 min. Trajectories of BMMCs with Gal3 KD showed impaired motility compared with control cells. When quantified, the differences were significant (Fig. 7E).

Next we examined whether Gal3 is involved in fibronectin-independent chemotaxis toward Ag and PGE₂. As shown in Fig. 7F and G, BMMCs with Gal3 KD showed significantly increased migration toward Ag and PGE₂, respectively, compared with their control counterparts. In the absence of chemoattractants, no differences in migration were observed between the two cell types. To find out how the production of cytokines and chemokines by activated BMMCs accounts for migration of BMMCs with Gal3 KD toward Ag, we let the cells migrate toward supernatants from Ag-activated BMMCs with normal Gal3 expression or with Gal3 KD (Fig. 7H). We recorded increased migration of BMMCs toward supernatants from both nonactivated and Ag-activated cells with Gal3 KD. The combined data suggest that Gal3 is involved in stabilizing β1-integrin on the plasma membranes of mast cells and facilitates proper cell adhesion and motility on fibronectin. On the other hand, Gal3 inhibits fibronectin-independent cell chemotaxis toward Ag, in part by regulating the release of chemoattractants from mast cell products.

DISCUSSION

In this study, we utilized RNAi screening to gain new insights into FcεRI signaling in mast cells. We focused on genes with unknown or poorly understood functions in mast cell activation. Based on the screen, we identified several potential negative and positive regulators of mast cell degranulation and focused on understanding the regulatory mechanism of one of them, Gal3.

RNAi screening in differentiated immune cells has historically been complicated by difficulties with transfection. To overcome this problem, we established a screening protocol based on lentiviral shRNA delivery. We used BMMCL, a mouse mast cell line that preserved a high growth rate following lentiviral transduction and exhibited stable expression of mast cell surface receptors, including FcεRI and KIT (data not shown). Among several functional assays that we examined, we chose Ag-induced degranulation for its robustness and specificity. The performance of the assay is considered acceptable for screening by high-throughput standards, with a Z′ factor of 0.37; Z′ factors of cell-based RNAi screens are generally lower than the optimal values of 0.5 to 1 (30). To improve evaluation of the results, we included a control-based Z-normalization step to compare degranulation assays whose magnitudes showed inherent variation between experiments. Another crucial step in data analysis was proper quality control filtering of technical replicates. The major indicator of per plate data quality was the inhibition of degranulation induced by positive-control shRNA targeting SYK. Following shRNA transductions, among 144 genes, we identified 15 candidates that showed reproducibility and consistency across screen runs. To our knowledge, none of these genes except Plscr1 and Lgals3 have been studied in mast cells yet. Plscr1 was shown to play a role in degranulation of a rat basophilic leukemia cell line (44), and Lgals3 was found to be a positive regulator of mast cell degranulation in studies with BMMCs from mice with Gal3 knockout (see below). In further analysis, we focused on Lgals3, encoding Gal3, which showed strong effects on degranulation by scoring with all three shRNAs used in the screen. The inhibitory effect of Gal3 was confirmed in primary BMMCs.

Gal3 belongs to the family of galectins, animal lectins that bind β-galactosides and have a unique structure containing a highly conserved C-terminal carbohydrate recognition domain and a proline-rich N-terminal sequence. The N-terminal sequence facilitates oligomerization of Gal3 up to pentamers and can be further organized into complex structures called lattices (45, 46). Gal3 plays various roles in immune processes during parasitic infections and allergic inflammation, affecting cytokine production, participating in recruitment of immune cells to inflammatory sites, and regulating Th2 responses (47–50). Gal3 was originally identified as an IgE-binding protein (51) and has been studied in several cell types, including BMMCs (31, 52). Here, we show that Gal3 is a negative regulator of FcεRI-induced cell degranulation at a wide range of Ag concentrations. The fact that most profound differences in degranulation were observed when the cells were activated with low doses of Ag suggests that Gal3 participates in setting the FcεRI activation threshold.

Previous studies showed decreased degranulation and production of IL-4 upon Ag stimulation of BMMCs derived from C57BL/6 mice with Gal3 KO. However, Gal3-deficient cells showed normal tyrosine phosphorylation of LYN, SYK, PLCγ1, and LAT1 (31), suggesting that Gal3 has no effect on early activation events after FcεRI triggering. In contrast, in our study, we found enhanced tyrosine phosphorylation of SYK, PLCγ1, JNK, and AKT but not FcεRI β and γ subunits. This indicates that Gal3 in BMMCs with Gal3 KD interferes with events immediately after tyrosine phosphorylation of the FcεRI subunits, and this leads to enhanced degranulation and calcium response. The observed discrepancy could not be explained simply by the genetic background of BMMCs used in this study (BALB/c) and the previous study with Gal3 KO cells, because we found enhanced degranulation in BMMCs with Gal3 KD derived from both BALB/c and C57BL/6 mice (our unpublished data). However, the observed discrepancy could be explained by the use of different methodologies of gene ablation, KO versus KD (53). During development, gene ablation can be compensated for by genetic mechanisms that are in many cases poorly understood (21, 54, 55). Previous studies in mice with Gal3 KO showed involvement of Gal3 in regulating cell differentiation and morphogenesis, which could contribute, among other things, to changes in mast cell physiology (49, 56). Inducible and conditional in vivo gene targeting of Gal3 in mast cells will be required to understand the basis of the different results obtained with BMMCs with Gal3 KO or KD. Moreover, differences in the methodology of BMMC derivation from bone marrow and in culture conditions may also significantly contribute to different mast cell phenotype alterations (57).

It has been suggested that Gal3 prolongs receptor signaling by restricting receptor movement within the membrane and by delaying its removal by constitutive endocytosis (58). Internalization of FcεRI, as well as some sphingolipids, is triggered by their extensive cross-linking and immobilization (59, 60). Recently, Gal3 was identified as a driver of endocytosis of clathrin-independent carriers, including glycosphingolipids and integrins (61). Based on our data, we propose that Gal3 facilitates FcεRI internalization. Our data suggest that aberrant trafficking of aggregated FcεRI in cells with Gal3 KD is accompanied by enhanced FcεRI-mediated signaling. Consistent with these findings, it has been shown that intracellular-trafficking pathways of plasma membrane receptors can determine the outcome of receptor signaling (62–64). In this connection it should be noted that we observed enhanced F-actin depolymerization in FcεRI-activated BMMCs with Gal3 KD compared to controls. This indicates that Gal3 is a positive regulator of actin polymerization, and this finding could explain enhanced degranulation in cells with Gal3 KD because F-actin is a negative regulator of FcεRI-mediated degranulation (65, 66).

Receptor ubiquitination modulates its internalization and sorting for degradation (67). In LAMP1-positive compartments, IgE-FcεRI complexes are committed for degradation, or in some cases, recycling of IgE to the cell surface can occur (68, 69). Our data show that Gal3 is one of the regulators participating in proper sorting of the aggregated FcεRI to LAMP1-positive compartments. We note that Gal3 is not likely to affect the rate of IgE degradation at early stages of activation (up to 15 min), since the same levels of FcεRI were obtained by immunoprecipitation experiments. Furthermore, we found that Gal3 does not affect FcεRI ITAM phosphorylation. This is consistent with a finding that phosphorylation of FcεRI ITAMs is dispensable for receptor ubiquitination (70). We propose that Gal3 promotes ubiquitination of FcεRI and in this way contributes to downregulation of some of the downstream FcεRI-mediated signaling pathways. Previous studies have shown that overexpression of CBL ubiquitin ligase in mast cells inhibits Ag-mediated signaling (71). We analyzed CBL tyrosine residue 700, which is phosphorylated by SRC family kinases and SYK (72), and found that FcεRI ubiquitination promoted by Gal3 did not require enhanced CBL phosphorylation at this site. We cannot exclude the possibility that Gal3 affects phosphorylation of other sites on CBL or accessibility of CBL to FcεRI signalosomes or other components of the CBL interactome, such as SHIP or CIN85 (73). How exactly Gal3 regulates ubiquitination of FcεRI remains to be elucidated.

Gal3 functions are mainly attributed to its glycan binding activity (46, 61, 74). However, we did not observe any effect of lactose, an inhibitor of Gal3 lectin binding, on BMMC degranulation and adhesion to fibronectin (data not shown). Similarly, the effect of Gal3 KD was not reversed by treatment of the cells with recombinant Gal3. This suggests that Gal3 functions in mast cells, as documented in this work, by its intracellular interactions, which are not mediated by its lectin binding activity or are not accessible to externally added lactose. In this regard, it should be noted that Gal3 has been shown to interact with several nonglycosylated molecules through protein-protein interactions (75, 76).

We have also demonstrated that Gal3 is involved in regulating surface expression and internalization of β1-integrin in BMMCs. Our data show that Gal3 participates in regulation of the level of β1-integrin on the cell surface in resting cells and stabilizes the β1-integrin on the surface upon Ag activation. Similarly, Gal3 was shown to interact with β1-integrin and drive its internalization (61, 77). Rapid displacement of β1-integrin from the cell surface observed after Ag activation of BMMCs with Gal3 KD can lead to inefficient interactions between the plasma membrane and fibronectin-coated surfaces. The motility of the cells upon PGE₂ stimulation could be affected by Gal3 at the level of β1-integrin activation, as β1-integrin was not internalized upon PGE₂, but its activation was affected by the expression level of Gal3. These defects were specific to fibronectin-mediated events, since the general ability of BMMCs with Gal3 KD to migrate toward chemoattractants was increased. It is possible that enhanced production of cytokines and PGD₂ observed in cells with Gal3 KD acts by autocrine mechanisms on enhanced chemotaxis toward Ag, as has been described previously (78). The upregulated migration of mast cells toward supernatants from activated BMMCs with Gal3 KD confirms the contribution of mast cell-produced chemoattractants. However, since PGE₂ alone is not known to trigger mast cell degranulation or cytokine production (79), the upregulated migration of the cells with Gal3 KD toward PGE₂ suggests involvement of other mechanisms. This conclusion is supported by elevated phosphorylation of AKT and ERK in nonactivated cells with Gal3 KD.

In conclusion, our data obtained in shRNA-based RNAi screening in activated mast cells allowed us to identify several potential new regulators of FcεRI-mediated mast cell degranulation. Detailed analysis of one of them, Gal3, showed that Gal3 in Ag-activated mast cells is involved in regulation of F-actin depolymerization, surface FcεRI internalization, and trafficking, and through these activities could affect a variety of mast cell signaling pathways.