Abstract
The T-cell Ig and mucin domain-containing molecules (TIMs) have emerged as promising therapeutic targets to correct abnormal immune function in several autoimmune and chronic inflammatory conditions. It has been reported that proinflammatory cytokine dysregulation and neutrophil-dominated inflammation are the main causes of morbidity in cystic fibrosis (CF). However, the role of TIM receptors in CF has not been investigated. In this study, we demonstrated that TIM-3 is constitutively overexpressed in the human CF airway, suggesting a link between CF transmembrane conductance regulator (CFTR) function and TIM-3 expression. Blockade of CFTR function with the CFTR inhibitor-172 induced an upregulation of TIM-3 and its ligand galectin-9 in normal bronchial epithelial cells. We also established that TIM-3 serves as a functional receptor in bronchial epithelial cells, and physiologically relevant concentrations of galectin-9 induced TIM-3 phosphorylation, resulting in increased IL-8 production. In addition, we have demonstrated that both TIM-3 and galectin-9 undergo rapid proteolytic degradation in the CF lung, primarily because of neutrophil elastase and proteinase-3 activity. Our results suggest a novel intrinsic defect that may contribute to the neutrophil-dominated immune response in the CF airways.
Cystic fibrosis (CF) is the most common lethal genetic disease in whites that is caused by mutations in the gene encoding the CF transmembrane conductance regulator (CFTR) chloride channel (1, 2). CF patients suffer from persistent pulmonary infections accompanied by chronic neutrophil-dominated inflammation that results in severe lung injury and ultimately death. Several mechanisms have been proposed to explain how CFTR mutations lead to chronic lung disease in CF including altered ion transport across the airway epithelium, dehydration of the airway surface layer (3), and increased production of proinflammatory cytokines in the CF airway, arguably caused by constitutive NF-κB activation (4). Regardless of the initial cause, hyperinflammation in the CF lung occurs early and continues throughout life (5). This inflammatory state is further amplified by bacterial infections, in particular, Pseudomonas aeruginosa (6, 7). The presence of bacteria promotes the airway epithelium to release proinflammatory cytokines such as the potent neutrophil chemoattractant IL-8 (8). This sustained inflammatory response recruits neutrophils in an attempt to resolve bacterial infection, yet paradoxically, chronic neutrophil stimulation also leads to neutrophil necrosis (9). As a consequence, neutrophils are not cleared effectively and neutrophil contents, including proteases, are released into the lung perpetuating the inflammatory cycle (10). Thus, a better understanding of the mechanisms involved in inducing and controlling inflammation is required for the successful design of novel intervention strategies in CF.
A growing body of evidence supports the critical role of T-cell Ig and mucin domain-containing molecules (TIMs) as modulators of the immune response in infection, autoimmunity, cancer, transplant tolerance, and kidney and liver aseptic injury (reviewed in Refs. 11, 12). Despite the initial discovery of the TIM gene family presence in a chromosomal region linked to airway hyperreactivity (13), the role of TIM receptors in airway inflammation is poorly understood. TIMs have been implicated in asthma (14–16), sarcoidosis (17), and pulmonary fibrosis (18), but their role in CF-related lung inflammation has not been investigated. TIM-3 is expressed in a variety of immune cells, including Th-1 (19), Th-17 (20), dendritic cells (21), NK cells (22), NKT cells (23), monocytes (23), macrophages (24, 25), and mast cells (26). There is mounting evidence that TIM-3 is a potent regulator of both the adaptive and innate immune responses; however, the mechanism depends on both cell type and specific disease. Indeed, TIM-3 has exhibited modulatory properties involved in tumor cell proliferation and immune evasion in nonimmune cells including endothelial cells (27, 28) and epidermal melanocytes (29). TIM-3 function has also been implicated in neutrophil recruitment and tissue injury in several in vivo models (30–32). Because neutrophil-dominated inflammation is one of the main causes of morbidity in CF (33), in this study, we investigated the expression of TIM-3 in bronchial epithelial cells and explored its role in the pathogenesis of CF lung disease.
Materials and Methods
Reagents
Unless stated otherwise, cell culture reagents were obtained from Life Technologies BRL (Karlsruhe, Germany). All other chemical reagents were purchased from Sigma-Aldrich (Dublin, Ireland) and were of the highest purity available.
Abs and recombinant proteins
34, 35). In brief, galectins were expressed using the pET expression system (Novagen, Madison, WI) in Escherichia coli BL21 (DE3) and purified with a lactose-agarose column (Seikagaku Kogyo, Tokyo, Japan) and dialyzed against PBS. Endotoxin was eliminated with Cellufine ETclean-L, a poly-ε-lysine–conjugated resin (Chisso, Tokyo, Japan). Galectin preparations used in this study were >95% pure as determined by SDS-PAGE, with <0.3 endotoxin units/ml (<0.03 ng/ml) detected by Limulus turbidimetric kinetic assays using a Toxnometer ET-2000 instrument (Wako, Osaka, Japan). Protein concentration was determined using a bicinchoninic acid assay reagent (Pierce Biotechnology) with BSA as a standard.
Cell culture
Immortalized human bronchial epithelial cells 16HBE14o¯ (HBE) (36) and immortalized bronchial epithelial cells from a (ΔF508/ΔF508) CF patient CFBE41o¯ (CFBE) (37) were kindly donated by Dr. D.C. Gruenert (University of California, San Francisco, San Francisco, CA). Cells were grown in flasks coated with fibronectin (1 mg/ml), collagen (1 mg/ml; BD Biosciences, Bedford, MA), and BSA (1 mg/ml) in complete media (MEM containing 10% [v/v] heat-inactivated FCS and 1% [v/v] penicillin/streptomycin) at 37°C in a humidified incubator with 5% CO2. The media was changed every other day. All experiments were conducted within 10 passages.
Bronchial brushing sample collection
Ten individuals were recruited for this study. Five patients (four male and one female patient) had CF confirmed by sweat testing or genotyping, and seven were non-CF control patients (four male and three female patients). The patient characteristics are summarized as follows (mean ± SD): CF patients: n = 5; age 21.6 ± 4.6 y; % forced expiratory volume in 1 sec, 50.6 ± 32.1; control non-CF patients: n = 7; age 50.14 ± 10.5 y. All patients (controls and CF) were undergoing diagnostic and/or therapeutic fiber-optic flexible bronchoscopy for clinical reasons. Full informed consent was obtained before the procedure according to a protocol approved by Beaumont Hospital Ethics Committee. Before withdrawal of the bronchoscope, an area 2 cm distal to the carina (medially located) in either the right or left main bronchus was selected and washed twice with 10 ml sterile 0.9% (w/v) NaCl. Subsequently, a sterile 10 × 1.2-mm bronchial brush (Olympus Medical Systems, Tokyo, Japan) was inserted through the appropriate port on the bronchoscope, and the chosen area was sampled with a brush by gently scraping the selected area. The brush was withdrawn and immediately placed into complete media. Brushes were gently agitated to dislodge cells into the medium, which was centrifuged at 200 × g for 5 min and cell pellets resuspended in 0.5 ml TRI reagent for RNA extraction and quantitative real-time PCR (QRT-PCR).
Bronchoalveolar lavage fluid sample collection
Bronchoalveolar lavage (BAL) fluid from children with CF (n = 9) or adults with CF (n = 14), non-CF bronchiectasis (n = 10), chronic obstructive pulmonary disease (COPD; n = 5), or healthy controls (n = 5) was collected from patients undergoing bronchoscopy for clinical reasons. The characteristics of the study population are summarized in Table I. Full informed consent from adult patients was obtained before the procedure according to a protocol approved by Beaumont Hospital Ethics Committee, full informed parental assent was obtained for all procedures, and ethical approval for the use of these samples was obtained from the Ethics Committee of Our Lady’s Children’s Hospital (Crumlin, Ireland). Bronchoscopy was performed via a laryngeal mask airway and the bronchoscope was directed to the lingula and right middle lobe. BAL was performed by instilling 1 ml/kg sterile normal saline per lobe. BAL samples from both sides were pooled. The characteristics of the study population are summarized in Table I. All BAL samples were centrifuged at 1000 × g for 10 min at 4°C, and cell-free supernatants were aliquoted and stored at −80°C for subsequent analysis.
QRT-PCR
Gene expression was analyzed by two-step QRT-PCR. Total RNA was extracted from HBE and CFBE (1 × 105) cells using TRI reagent following manufacturer’s instructions. Contaminating DNA was removed using DNase I (Qiagen, Crawley, West Sussex, U.K.) treatment. RNA (0.2–1 μg) was reversed-transcribed in 20 μl reaction volume (42°C, 30 min; 95°C, 5 min) using QuantiTect Reverse Transcription Kit (Qiagen) using a PTC-200 thermocycler (MJ Research, Cambridge, MA). cDNA (2 μl) was amplified with SYBR Green I Master mix (Roche, Basel, Switzerland) using the LightCycler 480 PCR system (Roche). Previously described primers for TIM-3 (23), galectin-9 (38), and GAPDH (39) were obtained from MWG Biotech (Ebersberg, Germany). PCR was performed according to the manufacturer’s instructions using the following protocol: preincubation (95°C for 3 min); amplification, 40 cycles consisting of denaturation, annealing, elongation (10 s at 95°C; 10 s at 57, 56, or 55°C, respectively, for TIM-3, galectin-9, or GAPDH; and 72°C for 10 s); melting curve analysis (95°C for 5 sec, 65°C for 1 min, and 97°C for 5 continuous acquisitions/°C); and final cooling step to 40°C. All samples were carried out in duplicate 20-μl reactions in 96-well plates, and a negative control with no cDNA template was included in every run. Specificity of the products was confirmed by visual inspection of melting curves and products run on a 1.2% (w/v) agarose gel. The relative expression of the gene was determined using the 2−ΔΔCt method with GAPDH as an internal control.
Laser-scanning cytometry
TIM-3 surface expression on airway epithelial cells was measured by laser-scanning cytometry (LSC). Cells (1 × 105) were seeded in eight-well Lab-Tek chamber glass slides (Nalgene Nunc, Roskilde, Denmark) and incubated overnight in complete medium. The medium was then aspirated and cells were fixed with methanol for 10 min at room temperature. Slides were washed with PBS (× 3) and blocked with 2% (w/v) BSA in PBS for 15 min at room temperature. After washing with PBS, cells were incubated at 4°C in the dark with primary Ab (1:20 in PBS) for 30 min. Then slides were washed and incubated with FITC-labeled secondary Ab (1:40 in PBS) for 30 min at 4°C in the dark. Control chambers were probed with secondary Ab only. Slides were washed with PBS; then cells were permeabilized and the nuclei stained in a single step using a 1:1 ratio of permeabilizing buffer (0.1% [w/v] sodium citrate, 0.1% [w/v] Triton X-100 in PBS) and 0.1 μg/ml solution of propidium iodide (Molecular Probes, Leiden, The Netherlands) in PBS. Slides were washed with PBS, and TIM-3 expression was quantified on a CompuCyte laser-scanning cytometer (CompuCyte, Westwood, MA). Cell nuclei were identified by propidium iodide fluorescence (588 ± 10 nm), and TIM-3 surface expression was detected by FITC fluorescence (530 ± 20 nm) and quantified. At least 3 × 103 cells were counted in triplicate in each well.
Immunocytochemistry and confocal microscopy
HBE and CFBE cells (1 × 105) were grown in complete media in an eight-well chamber slide or on 13-mm-diameter glass coverslips (VWR International, Dublin, Ireland). After removing media and washing with PBS, cells were fixed with either methanol or 4% (w/v) paraformaldehyde at room temperature for 20 min. Cells were washed again with PBS (× 3) and blocked with 2% (w/v) BSA in PBS for 15 min at room temperature. Then cells were probed with the anti–TIM-3 Ab (1:50 in PBS) for 30 min at 4°C in the dark. After washing, cells were probed again with the corresponding FITC-labeled secondary Ab (1:100 in PBS). Cells were washed (3×) and mounted on a glass slide using Vectashield fluorescence mounting media (Vector Laboratories, Orton Southgate, Peterborough, U.K.). Controls for this experiment included secondary Ab only. Cells were visualized by confocal microscopy using an LSM510 confocal microscope (Zeiss, Welwyn Garden City, U.K.). Images were captured at 40× magnifications under oil immersion.
Cell surface biotinylation analysis
Cell surface proteins from HBE and CFBE cells were isolated using a cell surface biotinylation kit (Pierce Biotechnology) according to the manufacturer’s instruction. Four 90% confluent 75-cm2 flasks were quickly washed twice with 8 ml ice-cold PBS per flask. Ten milliliters of ice-cold PBS containing 250 μg/ml EZ-Link sulfo-NHS-LC-biotin (Pierce Biotechnology) was added, and the cells were incubated with gentle agitation at 4°C for 30 min. The biotinylation reaction was stopped by addition of 500 μl quenching solution per flask; all contents were pooled and flasks were rinsed with a single 10-ml volume of TBS. The cells were pelleted after centrifugation at 500 × g for 3 min at 4°C. Cells were resuspended in 500 μl of the kit lysis buffer supplemented with protease inhibitors (100 μg/ml PMSF, 1 mM Na3VO4, 1 μg/ml aprotinin) and incubated for 30 min on ice and sonicated using five 1-s bursts every 10 min using a Vibra-Cell VC 130 PB ultrasonic processor (Sonics & Materials, Newtown, CT). Cells were vortexed every 5 min for 5 s to improve solubilization efficiency. The cell lysate was clarified by centrifugation at 10,000 × g for 2 min at 4°C. The solubilized biotinylated proteins were isolated on immobilized NeutrAvidin agarose columns after 1-h incubation at room temperature with end-over-end mixing by rotation. The column was washed with 500 μl of the kit wash buffer supplemented with protease inhibitors followed by centrifugation (1000 × g for 1 min), and the flow-through was discarded (4×). The bound biotinylated proteins were eluted from the column by incubating at 95°C for 5 min with 400 μl sample buffer containing 50 mM DTT, followed by 2000 × g for 2 min centrifugation. The isolated proteins were stored at −20°C.
Gel electrophoresis and Western blot analysis
Proteins were separated by SDS-PAGE on 10 or 12.5% (w/v) polyacrylamide gels. Proteins were transferred to a nitrocellulose membrane by semidry transfer at 150 mA for 90 min. Membranes were blocked with 5% (w/v) nonfat powdered milk in PBS containing 0.1% (v/v) Tween 20 (PBST) for 1 h at room temperature. Blots were incubated overnight at 4°C in blocking buffer containing Abs against TIM-3 (1:1000), phosphotyrosine (1:1000), or X-actin (1:1000) as a loading control marker where required. Subsequently, nitrocellulose membranes were washed for 30 min in PBST buffer (2×), probed with corresponding HRP-conjugated secondary Ab (1:1000) in PBST for 1 h, and then washed again. Blots were developed with Immobilon western chemiluminescent HRP substrate (Millipore, MA) and visualized on the Syngene G:Box chemi XL gel documentation system (Synoptics, Cambridge, U.K.). Protein band size was determined by loading SeeBlue Plus2 Prestained molecular mass marker (Invitrogen, Bioscience, Ireland) on each gel.
Inhibition of CFTR function
HBE cells (1 × 105) were serum starved overnight and then treated with 10 μM CFTR inhibitor-172 (CFTRinh172; Calbiochem, Merck Chemicals, Nottingham, U.K.) in low-serum medium for 48 h. Control cells were treated with vehicle 0.1% (v/v) DMSO. Cells were then collected in 0.5 ml TRI reagent for subsequent QRT-PCR analysis.
TIM-3 expression under inflammatory conditions
HBE and CFBE cells (1 × 105) were grown in complete medium in Lab-Tek chamber glass slides overnight. The medium was aspirated, and cells were washed with MEM and serum starved for 2 h. Serum-free medium was replaced by fresh MEM supplemented with 10 μg/ml LPS. The LPS concentration used has previously been reported to be within physiological levels in the CF lung and has been found to induce IL-8 production in CFBE cells (40). MEM was used as negative control. Cells were incubated for 24 h at 37°C followed by washing with PBS. Subsequently, TIM-3 expression was quantified by LSC as described earlier.
Cell viability assay
Cell viability was determined by MTT assay carried out in 96-well plates (4 × 104 cells/well). An MTT stock solution of 5 mg/ml was prepared in PBS and filter sterilized. Single-use aliquots were stored at −20°C. After treatment, 10 μl/100 μl medium was added to all wells. Cells were incubated at 37°C for 4 h. The plate was centrifuged at 700 × g for 5 min to pellet the purple formazan crystals. The supernatant was carefully removed, and 100 μl acid isopropanol solution (HCl 0.1N, 10% [v/v] Triton X-100 in anhydrous isopropanol) was added to dissolve the formazan crystals by pipetting. Plates were read at 570 nm, and viability was expressed as a percentage of the control cells.
TIM-3 immunoprecipitation
CFBE or HBE cells were grown to confluence in 10-cm dishes coated with fibronectin (1 mg/ml), collagen (2.9 mg/ml), and BSA (1 mg/ml). Cells (two plates/treatment) were washed with PBS and incubated at 37°C for 15 min with 2 ml full media containing 500 nM galectin-9 per plate or fresh full media as a control. Cells were washed with 5 ml ice-cold PBS (2×) and scrapped off in 1 ml Lamberth's break buffer (10 mM KCl, 3 mM NaCl, 4 mM MgCl2, 10 mM PIPES, pH 7) supplemented with protease inhibitors Complete Mini tablets and phosphatase inhibitors PhosStop Mini tablets (Roche) at 4°C. Cell lysates were sonicated on ice for 5 s at 20-W output with 1-min resting intervals (3×). The lysates were clarified by centrifugation at 240 × g at 4°C for 10 min. Clarified lysates were ultracentrifuged at 100,000 × g for 1 h at 4°C to obtain a crude cell membrane fraction. The crude membrane pellet was resuspended in 1 ml radio immunoprecipitation assay buffer (150 mM NaCl, 50 mM Tris-HCl, pH 8, 0.5% [w/v] sodium deoxycholate, 1% Triton X-100 [v/v], 0.1% [w/v] SDS) containing protease and phosphatase inhibitors, and homogenized using a 1-ml syringe and a Microlance 3, 21G 1.5-inch needle (BD, Oxford, U.K.). The resuspended crude membrane lysate was clarified by centrifugation, and the protein concentration was determined by bicinchoninic acid in a 5-μl aliquot. Equal amounts of crude membrane extract (500 μg) in a total volume of 1 ml were used for immunoprecipitation. Samples were precleared with 6 μg normal goat IgG (Santa Cruz Biotechnology) and 50 μl protein-G Dynabeads (Invitrogen) for 1 h at 4°C with rotation. The beads were collected magnetically, washed with 1 ml PBS (× 3), and then beads were boiled in 20 μl of 2× sample buffer at 95°C for 10 min. This sample was termed IgG immunoprecipitation control. The precleared samples were incubated with 6 μg goat anti–TIM-3 Ab for 2 h at 4°C with rotation, and the immunocomplex was captured by incubation with 50 μl protein-G Dynabeads as described earlier. The activation of TIM-3 by galectin-9 was analyzed by Western blot of immunoprecipitated cell membranes probed for phosphotyrosine. Blots were stripped with Restore Western blot stripping buffer (Pierce Biotechnology) according to manufacturer’s instructions and reprobed with rabbit anti–TIM-3 Ab to confirm equal amounts of TIM-3 being immunoprecipitated.
Galectin-9 treatment
CFBE cells (1 × 105
Determination of IL-8 and galectin-9 levels by ELISA
IL-8 protein concentration in cell supernatants (100 μl) was determined by sandwich ELISA according to manufacturer’s instructions. BAL samples were diluted 1:10 in PBS (final volume, 100 μl) for IL-8 quantification by ELISA to ensure that the values were within the range of detection. Concentrations of galectin-9 in BAL were determined using a custom-made sandwich ELISA as previously described (35, 41). In brief, 96-well plates (Nunc, Naperville, IL) were coated with an anti-human galectin-9 mAb (9S2-3; GalPharma), blocked with 3% (w/v) BSA containing 0.05% (v/v) Tween 20 in PBS, and then incubated for 1 h at 37°C with BAL (100 μl) or known concentrations of recombinant human galectin-9. After several washings, bound galectin-9 was recognized by polyclonal anti-human galectin-9 Ab conjugated with biotin using EZ-Link Sulfo-NHS-Biotin reagent (Pierce Biotechnology). Quantification was performed using HRP-streptavidin (Invitrogen, Tokyo, Japan) and the colorimetric substrate tetramethylbenzidine (KPL, Gaithersburg, MD). The OD was recorded using a microplate spectrophotometer (Bio-Rad, Tokyo, Japan). The galectin-9 ELISA limit of detection was 15.6 pg/ml.
Polarized cell culture
CFBE cells were grown as polarized cultures at a liquid–liquid interface as previously described (42). In brief, cells (7 × 104) were seeded onto 1.0-μm polyethylene hanging cell culture inserts (Millipore) and maintained in full media. Transepithelial electrical resistance measurements were taken every 48 h using the EVOM epithelial voltohmmeter (World Precision Instruments, Stevenage, U.K.) according to manufacturer’s instructions. Cells with transepithelial electrical resistance >1000 Ω.cm2 were used for experiments. Cells were serum starved overnight before treatment with galectin-9 (50 nM) in low-serum media for 24 h. IL-8 levels in both apical (0.2 ml) and basolateral (1.2 ml) supernatants were measured by IL-8 ELISA.
Proteolytic degradation of recombinant TIM-3
Pooled CF or pooled non-CF bronchiectasis BAL (10 μl) was incubated with rhTIM-3 at 37°C. Non-CF BAL with nondetectable neutrophil elastase (NE) activity or PBS was used as a negative control. Samples were collected at specific time points (0–30 min, 1–24 h). The reaction was immediately stopped by the addition of sample loading buffer (2×) followed by 10-min incubation at 95°C. Samples were stored at −20°C for subsequent Western blot analysis. In some experiments, CF BAL samples were preincubated for 1 h at 4°C with 1 μl of 1.1 mg/ml aprotinin, 5 mg/ml soya bean trypsin inhibitor (SBTI), 0.2 M Pefabloc SC, 5 mg/ml E-64 (Calbiochem), 10 mg/ml pepstatin A, 0.5 M EDTA, 1 mg/ml α-1-antichymotrypsin (A1CT; Calbiochem), 2.5 mM GM6001 (Millipore), 10 mg/ml benzamidine, 20 mg/ml o-phenanthroline, 10 mg/ml TNF-α protease inhibitor-1 (Calbiochem), 2 mg/ml phosphoramidon (Calbiochem), 3 mM N-(methoxysuccinyl)-Ala-Ala-Pro-Val-chloromethyl ketone (CMK), 2 mg/ml N-p-tosyl-l-phenylalanine CMK, 10 mg/ml Nα-tosyl-l
Cleavage of recombinant TIM-3 by serine proteases and determination of N-terminal sequence of proteolytic fragments
rhTIM-3 (100 ng) was incubated with 10−7 M, 5 × 10−8 M, or 10−8 M NE, cathepsin G (Cath G; Elastin Products, Owensville, MO), or proteinase-3 (PR3; Athens Research and Technology) for 2 h at 37°C in 10 μl of 0.1 M HEPES buffer, pH 7.5, containing 0.5 M NaCl, 0.05% (w/v) Brij-35. All reactions were stopped by addition of 2× loading sample buffer and boiling at 95°C for 10 min. Degradation of rhTIM-3 was analyzed by Western blotting. For N-terminal sequencing analysis of proteolytic fragments, samples were separated by SDS-PAGE using a 10% (w/v) polyacrylamide gel and transferred to polyvinylidene difluoride membrane. The membrane was rinsed in methanol for 5 s and fixed in Ponceau staining buffer (40% [v/v] isopropanol, 10% [v/v] acetic acid, 50% [v/v] water, and 0.1% (w/v) Ponceau S) for 20 min at room temperature. Proteolytic fragments were visualized after three 5-min consecutive washings in destaining buffer (45% [v/v] isopropanol, 45% [v/v] acetic acid, 10% [v/v] water). Bands corresponding to proteolytic fragments were excised from the membrane and analyzed by N-terminal sequence Edman degradation by Altabioscience (Birmingham, U.K.).
Cleavage of native TIM-3 by NE and PR3 on CFBE cell outer surface
CFBE cells (1 × 105/well) were cultured in complete media in Lab-Tek chamber glass slides overnight. The media was then replaced by 1 ml low-serum media containing NE (10−7 M), PR3 (10−7 M), or a combination of both (10−7 M each). Cells were incubated for 2 h at 37°C and then washed and fixed for TIM-3 detection on the cell outer surface by LSC as described earlier.
NE activity assay
NE activity was determined using the chromogenic substrate N-(methoxysuccinyl)-Ala-Ala-Pro-Val p-nitroanilide specific for human NE. Samples (10 μl) were mixed with 90 μl of 3 mM substrate in assay buffer (0.5 M NaCl, 0.1% [v/v] Brij-35, 0.1 M HEPES, pH 7.5). OD was recorded at 405 nm for 5 min at 1-min intervals at 37°C using a Wallac 1420 Victor2 multilabel counter (PerkinElmer, Waltham, MA). NE activity in samples was calculated using an extinction coefficient of 9500 l × mol−1 × cm−1. NE activity was expressed as milliunits per milliliter (mU/ml) BAL defined as micromoles of peptide hydrolyzed per min per milliliter BAL. Samples were analyzed in duplicate.
Statistical analysis
Data were analyzed with GraphPad Prism 4.0 software package (GraphPad Software, La Jolla, CA). Unless specified otherwise, data are expressed as mean ± SD of at least three independent experiments in triplicate. Differences were calculated by two-tailed Mann–Whitney U test (two data sets) or Kruskal–Wallis test with post hoc Dunn’s multiple-comparison test (more than two data sets). The correlation between galectin-9 and NE activity in BAL was determined by one-tailed Spearman correlation analysis. Results were considered significant when *p < 0.05.
Results
TIM-3 and its ligand galectin-9 are constitutively upregulated in CF bronchial epithelial cells
Altered TIM-3 expression has been reported in several inflammatory conditions; however, how TIM-3 expression is affected by CF is not known. We used HBE and CFBE cells, and examined TIM-3 expression by QRT-PCR. TIM-3 was found to be present in both cell types, but there was a significant increased expression of TIM-3 in CF cells compared with normal HBE cells (p < 0.05; Fig. 1A). TIM-3 overexpression was confirmed by QRT-PCR in bronchial brushings from CF patients who were homozygous for the ΔF508 mutation (Fig. 1B). Next, we analyzed the expression of the TIM-3 ligand, galectin-9 (32). QRT-PCR analysis demonstrated that galectin-9 was also overexpressed (p < 0.05) in CFBE cells compared with HBE cells (Fig. 1C). However, no significant difference in galectin-9 expression was detected in bronchial brushings (Fig. 1D). Having observed that TIM-3 and its ligand galectin-9 are upregulated in CFBE cells, we further investigated the link between CF and TIM-3/galectin-9 expression in bronchial epithelial cells. To examine the effect of a lack of CFTR function in our experimental setting, we used the CFTRinh172. This thiazolidinone channel inhibitor has been reported to be highly selective for CFTR channel (43) and has been shown to induce CF-like inflammatory responses in HBE cells at a concentration of 10 μM without causing cytotoxicity (44). After 48-h treatment with CFTRinh172 (10 μM), TIM-3 expression was measured in HBE cells by QRT-PCR. A significant (p < 0.05) upregulation of TIM-3 was observed compared with vehicle control (0.1% v/v DMSO; Fig. 1E). Pharmacological inhibition of CFTR function revealed a similar effect on galectin-9 expression (Fig. 1F). These results suggest that the TIM-3/galectin-9 axis is constitutively upregulated in CF bronchial epithelium because of lack of CFTR function.
TIM-3 and galectin-9 mRNA are upregulated in CF. A, TIM-3 mRNA expression was measured in the bronchial epithelial cell line CFBE41o− (CFBE) compared with its non-CF counterpart 16HBE14o− (HBE). B, Differential TIM-3 mRNA expression in bronchial brushings from healthy controls (n = 7) and CF patients (n = 5). C, Galectin-9 mRNA expression was measured in CFBE cells compared with HBE cells. D, Differential galectin-9 expression in bronchial brushings from healthy controls (n = 7) and CF patients (n = 5). HBE cells were untreated (control) or treated with 10 μM CFTRinh172 for 48 h, and TIM-3 (E) or galectin-9 (F) expression determined. The differential expression levels were analyzed by QRT-PCR relative to non-CF controls and normalized to GAPDH expression. Bars show mean ± SD; n = 3; statistical significance analyzed by Mann–Whitney U test. *p < 0.05. Results are representative of three independent experiments.
TIM-3 protein is overexpressed in CF and upregulated by LPS
Next, we investigated TIM-3 expression at the protein level by LSC. This method was chosen over flow cytometry to quantify TIM-3 cell surface expression to avoid artifacts caused by epithelial layer disruption. TIM-3 expression was detected in both cell types; however, CFBE cells expressed about 50% more TIM-3 than HBE cells (Fig. 2A). The expression of TIM-3 in HBE and CFBE cells was further confirmed by Western blot analysis (Fig. 2B). TIM-3 is a 33-kDa protein that appears as two diffuse bands of 50 and 64 kDa because of glycosylation (45). Of note, the 64-kDa band corresponding to the fully glycosylated form was also overexpressed in CFBE cells. This mature glycosylated form of TIM-3 has been reported to be present on cell surface and capable of acting as a functional receptor (45).
TIM-3 protein is overexpressed in CF and upregulated by LPS. A, Quantification of TIM-3 surface expression in unstimulated HBE and CFBE cells by LSC. A minimum of 3000 cells were counted. Bar graph shows mean fluorescence intensity (MFI) peak values for each point. Data are shown as mean ± SEM; n = 3. B, Representative immunoblot of TIM-3 protein levels in HBE and CFBE whole-cell lysates. Glycosylated TIM-3 was detected as diffuse bands at 50 and 64 kDa. An immunoblot for actin served as a protein loading control. C, HBE (white bars) and CFBE (black bars) cells (105) were treated with LPS (10 μg/ml) for 24 h or low-serum media as control and TIM-3 cell surface expression analyzed by LSC. Statistical significance was calculated by Mann–Whitney U test. *p < 0.05. All results are representative of a minimum of three independent experiments.
Additional studies were conducted to determine whether inflammatory stimuli could further influence TIM-3 expression in bronchial epithelial cells. Cells were treated with LPS (10 μg/ml), and TIM-3 expression was measured by LSC after 24-h treatment. LPS (10 μg/ml) induced a 2-fold increase in TIM-3 expression in HBE cells and a 3-fold increase in CFBE cells (Fig. 2C). These data revealed that inflammatory stimuli present in the CF lung can modulate TIM-3 overexpression in bronchial epithelial cells. Furthermore, TIM-3 expression is more susceptible to LPS-induced upregulation in CF than in non-CF bronchial epithelial cells.
Galectin-9 phosphorylates TIM-3 on the cell outer membrane of CFBE cells
To establish whether TIM-3 can act as a membrane receptor in bronchial epithelial cells, we examined TIM-3 localization by fluorescence confocal microscopy in HBE and CFBE cells. This investigation revealed that TIM-3 was clearly localized to the membrane (Fig. 3A, white arrows), further corroborating TIM-3 membrane expression as shown by LSC (Fig. 2A). Membrane expression of TIM-3 was additionally confirmed in both cell types by biotinylation of the outer membrane cell surface (Fig. 3B). We next investigated whether the TIM-3 protein expressed in bronchial epithelial cells was a functional receptor for galectin-9. For this purpose, CFBE or HBE cells were treated with galectin-9, a TIM-3 ligand known to activate TIM-3 by phosphorylation of the tyrosine motif Y265 in the cytosolic domain (45). Because galectin-9–induced TIM-3 signaling has never been documented in CFBE or other epithelial cells, we used 500 nM galectin-9 because it proved nontoxic to the cells (data not shown), and this dose has been reported to induce TIM-3–mediated responses in Th cells (32). After 15-min treatment with galectin-9, TIM-3 was immunoprecipitated from cell membrane extracts, and Western blot analysis revealed that TIM-3 exhibited a substantial increase in the level of tyrosine phosphorylation compared with untreated controls (Fig. 3C, upper panel). Interestingly, a background level of phosphorylation was detected in untreated samples, suggesting constitutive activation of TIM-3, possibly because of constitutively produced galectin-9 (Fig. 1C).
TIM-3 expressed on bronchial epithelial outer cell membrane is a functional receptor. A, HBE or CFBE cells (105/well) were grown on coverslips and probed with anti–TIM-3 followed by FITC-labeled secondary Ab (left panel). Controls included secondary Ab only (right panel). Cell nuclei were visualized by DAPI (blue), and expression of TIM-3 was detected by green fluorescence. White arrows indicate cell surface localization. Images for each Ab treatment were captured using identical image capture parameters at 40× magnification under oil immersion. Images are representative of three independent experiments. B, Localization of TIM-3 in the outer cell membrane of bronchial epithelial cells. Cell surface proteins were isolated using a cell surface biotinylation kit. rhTIM-3 was used as a positive control in the immunoblot. C, Phosphorylation of TIM-3 following galectin-9 (Gal-9) stimulation. HBE or CFBE cells were either left untreated or treated with 500 nM Gal-9 for 15 min. TIM-3 was immunoprecipitated (IP) from cell membranes with goat polyclonal anti–TIM-3 Ab. Normal goat-IgG was used as a control of nonspecific binding. Activation of TIM-3 was analyzed by Western blotting (WB) using a mouse monoclonal anti-phosphotyrosine Ab (p-tyrosine; upper panel). Blots were stripped and reprobed with rabbit anti–TIM-3 Ab to confirm equal levels of immunoprecipitated TIM-3 (lower panel). The image shown is representative of two independent experiments.
Galectin-9 induces IL-8 production via TIM-3 in CFBE cells
Next, we investigated whether TIM-3 can modulate the immune response in the CF lung. TIM-3 has been involved in regulation of neutrophil recruitment via modulation of neutrophil chemokine expression (30). Therefore, we examined IL-8 production in CFBE cells on galectin-9 stimulation. Galectin-9 (50 nM) promoted IL-8 production in CFBE cells as measured in cell culture supernatants after 24 h (Fig. 4A). To determine whether IL-8 production was due to a TIM-3 signaling on galectin-9 ligation, we exposed cells to galectin-9 pretreated with lactose (30 mM) or rhTIM-3 (100 ng/ml), which has been shown to prevent galectin-9 binding to TIM-3 (46). Blockade of the TIM-3–galectin-9 interaction by lactose or rhTIM-3 suppressed IL-8 production in CFBE cells (Fig. 4B). These results suggest that galectin-9 is a natural agonist for TIM-3 in bronchial epithelial cells, inducing production of IL-8. Subsequently, we set out to establish the localization of TIM-3 in polarized cells. Apical treatment of CFBE cells with galectin-9 resulted in a 2-fold increase in IL-8 release, demonstrating that TIM-3 is localized apically in CFBE cells (Fig. 4C). Basolateral treatment also induced IL-8 release but to a lesser extent (1.5-fold increase), suggesting that TIM-3 expression is also localized basolaterally.
Galectin-9 (gal-9) induces IL-8 production via TIM-3 in CFBE cells. A, CFBE cells (105) were treated with 50 nM gal-9 or left untreated (ctrl). After 24 h, IL-8 levels were determined in cell culture supernatant by ELISA. Statistical significance was analyzed by Mann–Whitney U test. B, Blockade of TIM-3/gal-9 interaction abrogates IL-8 production. CFBE cells (105) were treated with 50 nM gal-9 alone or in combination with 30 mM lactose (gal-9 + lact), or 100 ng/ml rhTIM-3 (gal-9 + rhTIM-3). IL-8 levels were determined in cell culture supernatants by ELISA and expressed as percentage increase compared with control. Results are expressed as mean ± SD; n = 3. Statistical significance was analyzed by Kruskal–Wallis test. C, CFBE cells were grown at a liquid–liquid interface as polarized monolayers. Cells were treated with 50 nM gal-9 (black bars) or left untreated (white bars). After 24 h, IL-8 levels were determined in the apical and basolateral compartment by ELISA. Statistical significance was analyzed by Mann–Whitney U test. *p < 0.05. Results are representative of at least three independent experiments for monolayers or two experiments for polarized cells.
Galectin-9 is degraded in CF BAL
Prompted by the results of in vitro experiments suggesting an upregulation of galectin-9 in CF (Fig. 1C), we next measured the levels of galectin-9 in CF BAL by ELISA to determine whether galectin-9 function has relevance in vivo. Galectin-9 was not detected in any of the CF BAL samples tested, whereas galectin-9 was found in healthy controls (Fig. 5A). COPD and non-CF bronchiectasis BAL samples served as inflammatory controls and showed varied levels of galectin-9 ranging from 0–1000 pg/ml (Fig. 5A). Because BAL fluid is considered to be from 25- to 100-fold diluted compared with epithelial lining fluid (39, 44), these values indicate that galectin-9 can be present in the nanomolar range on the lung surface in vivo in conditions with a marked neutrophilic component such as COPD or non-CF bronchiectasis, but not in adult CF.
Galectin-9 is degraded in CF BAL. A, Galectin-9 levels were measured by ELISA in BAL from non-CF bronchiectasis (Bron; n = 10), adult CF (n = 14), COPD (n = 5), and healthy controls (HC; n = 5). Statistical significance calculated by Kruskal–Wallis test. *p < 0.05 compared with adult CF BAL. B, Galectin-9 levels were measured in CF patients aged 0–2 y (n = 4), 3–8 y (n = 5), and >18 y (n = 14). Statistical significance calculated by Kruskal–Wallis test. *p < 0.05 compared with 0–2 y age group. C, Correlation of galectin-9 levels and NE activity in BAL analyzed by one-tailed Spearman correlation (r = −0.9384; p = 0.0002). Horizontal bars represent median values.
Galectin-9 has been shown to be rapidly degraded by NE (47). Because this protease is abundantly present in CF BAL (48), we measured galectin-9 levels in CF infants and children’s BAL with low levels of neutrophil infiltration and, therefore, low levels of neutrophil-derived proteases. We found that galectin-9 levels decline with age (Fig. 5B) and correlate inversely with NE (Fig. 5C). These data collectively suggest that galectin-9 in the lung undergoes degradation by proteases, mainly neutrophil-derived elastase.
TIM-3 is degraded in CF BAL
Because TIM-3 was not detected in CF BAL by Western blot analysis (data not shown), we next investigated the effect of the proteolytic burden of the CF lung on the integrity of the galectin-9 receptor. rhTIM-3 (300 ng) was completely degraded after 24-h incubation at 37°C in CF BAL compared with the PBS control, as evidenced by the disappearance of the 64-kDa band on Western blot analysis (Fig. 6A). In contrast, TIM-3 was only partially degraded in non-CF bronchiectasis BAL (Fig. 6A), possibly because of lower abundance of proteases compared with CF BAL (Table I). A time course of TIM-3 proteolytic degradation showed that after only 10-min incubation, proteolytic degradation of the 64-kDa band commenced and TIM-3 was rapidly cleaved into smaller fragments of ∼60 and 36 kDa (Fig. 6B). Longer incubation times showed a further fragmentation of TIM-3, possibly suggesting the involvement of various proteases at different stages of degradation. The 64-kDa band corresponding to the fully glycosylated form of TIM-3 was almost completely degraded after 2 h. After 24 h, virtually no TIM-3 fragments were detected by Western blot. In contrast, rhTIM-3 was stable for 24 h in control non-CF BAL where NE activity was not detected and in PBS controls (data not shown).
TIM-3 is degraded in CF BAL. TIM-3 degradation was analyzed by Western blotting. A, rhTIM-3 (300 ng) was incubated for 24 h at 37°C with CF or non-CF bronchiectasis (Bron) BAL (10 μl). PBS was used as a negative control. B, rhTIM-3 (100 ng) were incubated for the indicated time points at 37°C with 10 μl CF BAL (upper panel) or non-CF BAL as a control (lower panel). Identification of proteases degrading TIM-3 in CF BAL. CF BAL (20 μl) was preincubated with protease inhibitors at 4°C for 1 h. Subsequently, samples were incubated with rhTim-3 (100 ng) at 37°C for 2 h (C–E). C, Nonspecific inhibitors (aprotinin [Apro], SBTI, Pefabloc [PEFA], pepstatin A [Pep A], GM6001, E64). D, Metalloprotease inhibitors (GM6001, o-phenanthroline [OP], TNF-α protease inhibitor-1 [TAPI-1], EDTA, phosphoramidon [Phos]). E, Serine protease inhibitors (CMK, A1CT, N-p-tosyl-l-phenylalanine CMK [TPCK], benzamidine [BENZ], Nα-tosyl-l-lysine CMK hydrochloride [TLCK]). F, Natural neutrophil serine protease inhibitors (A1AT, elafin [ELA], SLPI). Results are representative of three independent experiments. Lanes illustrated in B and F were run on the same gel but were noncontiguous.
Our next aim was to identify the proteases implicated in TIM-3 degradation. Because proteolytic degradation of TIM-3 has never been reported, a systematic approach was adopted to categorize the proteases involved in TIM-3 fragmentation. CF pooled BAL aliquots were preincubated with specific protease inhibitors for each class of enzyme for 1 h at 4°C before adding rhTIM-3. After 2-h incubation at 37°C, samples were analyzed by Western blot to identify which protease inhibitor could prevent degradation of TIM-3. First, nonspecific protease inhibitors targeting the main protease families were used: E64 is an inhibitor of cysteine proteases, pepstatin A an aspartic protease inhibitor, GM6001 a general metalloprotease inhibitor, and Pefabloc a potent serine protease inhibitor. Only Pefabloc, and to a lesser extent pepstatin A, prevented the degradation of full-length TIM-3, implicating a serine and an aspartic protease in TIM-3 cleavage (Fig. 6C). aprotinin and SBTI are inhibitors of the serine protease chymotrypsin and trypsin families, respectively, but only SBTI appeared to have a moderate inhibitory effect on TIM-3 degradation. Because a nonidentified metalloprotease has been implicated in TIM-1 shedding (49), a wider range of metalloprotease inhibitors were tested in comparison with GM6001. The metalloprotease inhibitors used included o-phenanthroline, TNF-α protease inhibitor-1, EDTA, and phosphoramidon. However, none of the metalloprotease inhibitors was capable of inhibiting the rapid degradation of the 64-kDa band corresponding to full-size TIM-3 (Fig. 6D). In addition to Pefabloc, several serine protease inhibitors were used to classify more precisely the serine protease involved in TIM-3 cleavage (Fig. 6E). N-methoxysuccinyl-alanine-alanine-proline-valine-CMK, a potent NE inhibitor, completely prevented TIM-3 cleavage. In contrast, the trypsin-like serine protease inhibitors had little (N-p-tosyl-l-phenylalanine CMK) or no effect (A1CT). Similarly, the chymotrypsin-like serine protease inhibitor benzamidine had no effect and Nα-tosyl-l-lysine CMK hydrochloride had only a moderate effect in preventing TIM-3 degradation. This set of experiments revealed neutrophil serine proteases as the most likely group of enzymes involved in TIM-3 degradation in CF BAL. Neutrophil serine proteases comprise NE, PR3, and Cath G. The natural serine protease inhibitors A1AT, elafin, and SLPI were used to further identify the neutrophil-derived serine protease mainly responsible for the rapid cleavage of TIM-3 (Fig. 6F). A1AT blocked TIM-3 degradation completely. A1AT is a potent inhibitor of NE; therefore, these results suggest that NE is implicated in TIM-3 degradation. Elafin also blocked TIM-3 degradation completely. The fact that elafin prevented TIM-3 degradation indicates that, in CF BAL, Cath G does not play a major role in the initial cleavage of TIM-3 as elafin inhibits NE and PR3 but not Cath G. This result was further confirmed by the lack of an effect of A1CT, which also inhibits Cath G. In contrast, SLPI only partially inhibited TIM-3 degradation, suggesting that PR3 can also cleave TIM-3 as SLPI inhibits NE and Cath G, but not PR3. Collectively, these results suggest that NE and PR3 are the main serine proteases involved in TIM-3 rapid cleavage in the CF lung. However, this finding does not exclude the involvement of other proteases in further degradation of TIM-3 serine protease derived fragments, alone or in combination with serine proteases.
TIM-3 is degraded by neutrophil serine proteases
The data obtained with the serine protease inhibitors provided evidence for a role for neutrophil-derived serine proteases in TIM-3 degradation in the CF lung. To further confirm this role, we examined the ability of purified human neutrophil serine proteases to cleave TIM-3 in vitro. A dose–response experiment was carried out with rhTIM-3 incubated for 2 h at 37°C with physiologically relevant concentrations of each protease (10−7–10−8 M). Western blot analysis of the samples showed that all serine proteases, but in particular NE and PR3, cleaved TIM-3 in a dose-dependent manner (Fig. 7A). Interestingly, each protease produced fragments of different sizes suggesting different TIM-3 cleavage sites for each protease. NE produced a single band corresponding to a fragment between 36 and 50 kDa; Cath G produced a clear fragment of ∼36 kDa and a weak band corresponding to a fragment of ∼40 kDa. PR3 generated a band of ∼50 kDa and a weaker band of 36 kDa. To identify specific TIM-3 cleavage sites, we analyzed the proteolytic products by N-terminal sequencing Edman degradation. NE was found to cleave TIM-3 between Ser-186 and Arg-187, whereas PR3 cleaved TIM-3 between Ile-199 and Arg-200 (Fig. 7B). Both NE and PR3 were capable of cleaving TIM-3 in the extracellular region of TIM-3, which would result in receptor inactivation in vivo. These TIM-3 fragments could then undergo further degradation by additional proteases present in BAL. As the experiments described so far were conducted in vitro using recombinant TIM-3, we next evaluated the ability of NE and PR3 to cleave native TIM-3 expressed on the cell surface of CF bronchial epithelial cells. CFBE cells were treated with 10−7 M of NE, PR3, or a combination of both for 2 h at 37°C in low-serum media. The remaining levels of TIM-3 on the cell outer surface were measured by LSC. Both NE and PR3 decreased the levels of TIM-3 cell outer surface expression. Furthermore, simultaneous addition of both proteases resulted in an almost complete depletion of TIM-3 protein expression on the surface of CFBE cells (Fig. 7C). These results confirm the ability of neutrophil-derived serine proteases to degrade native TIM-3 in vivo (Fig. 8).
TIM-3 is degraded by neutrophil serine proteases. A, Effect of purified neutrophil serine proteases on TIM-3 cleavage. TIM-3 (100 ng) was incubated with purified NE, cath G or PR3, or untreated (ctrl) at 37°C for 2 h. TIM-3 degradation was analyzed by Western blotting. Image is representative of two independent experiments. B, Schematic representation of NE and PR3 cleavage sites in the amino acid sequence of TIM-3. rhTIM-3 proteolytic fragments were analyzed by N-terminal sequencing Edman degradation to identify the cleavage site for each protease. Arrowheads represent cleavage sites by NE and PR3. NE cleaves TIM-3 at Ser-186 to Arg-187 and PR3 at Ile-199 to Arg-200. C, TIM-3 is cleaved on the cell surface of CF bronchial epithelial cells by NE and PR3. CFBE cells were treated with 10−7 M of NE, PR3, or a combination of both for 2 h at 37°C in MEM containing 1% (v/v) FCS. Levels of TIM-3 on the cell surface were measured by LSC. A minimum of 3000 cells were counted. Bar graph shows mean fluorescence intensity (MFI) peak values for each point. Data are shown as mean ± SEM; n = 4. Statistical significance calculated by Kruskal–Wallis test. *p < 0.05 compared with untreated control. Results are representative of two independent experiments.
Schematic representation of TIM-3/galectin-9 signaling in the lung. A, Following bacterial infection, TIM-3 receptor and galectin-9 are upregulated in response to LPS within the normal lung (1). Galectin-9 binds to TIM-3 expressed on the surface of bronchial epithelial cells inducing phosphorylation and production of IL-8. Neutrophils are then recruited to the lung to fight the bacterial infection (2). B, In the CF lung, galectin-9 and TIM-3 are constitutively overexpressed, an effect further increased by the presence of LPS (1). In CF newborns, TIM-3 becomes constitutively activated in the airway lumen, resulting in early neutrophil recruitment because of IL-8 production (2). Within the adult CF lung, the TIM-3/galectin-9 signaling axis required for proinflammatory and anti-inflammatory responses are disturbed because of proteolytic degradation (3), contributing to the perpetuation of neutrophil-driven inflammation (4).
Discussion
To our knowledge, we report for the first time that TIM-3 and its ligand, galectin-9, are constitutively overexpressed in CF airway epithelial cell surface, an observation further confirmed in CF patient samples. This finding implies a novel role for CFTR in TIM-3 expression because pharmacological inhibition of CFTR in normal cells induced an upregulation of TIM-3 and its ligand galectin-9. We also established that TIM-3 is a functional receptor capable of modulating the inflammatory response in bronchial epithelial cells. Cells stimulated with physiological relevant levels of galectin-9 induced IL-8 production by CF bronchial epithelial cells, indicating that TIM-3 may initiate the neutrophilic dominated inflammation in the CF lung. In addition, we demonstrated that the expression of TIM-3 can be modulated by LPS, which underscores the importance of TIM-3 under inflammatory conditions. Interestingly, both galectin-9 and TIM-3 undergo rapid proteolytic degradation in the CF lung because of serine protease activity. These data suggest that the dysregulation of TIM-3/galectin-9 signaling may play an important role in the pathogenesis of lung disease (Fig. 8). Furthermore, constitutive upregulation of this receptor and its ligand may reflect a proinflammatory state in CF bronchial epithelial cells (50, 51). Increased TIM-3 and galectin-9 expression in the lung may explain the early neutrophil airway infiltration observed in CF newborns (52) and in aseptic CF animal models (5). Of clinical relevance, we demonstrated that both TIM-3 and galectin-9 undergo rapid proteolytic degradation by serine proteases in CF BAL, which could impact on the previously described role of galectin-9, inducing resolution of inflammation via apoptosis of immune cells (32, 53).
TIM-3 expression has been previously reported in lung tissue (45) and in murine bronchial epithelial cells (11), data that are in line with our results on the expression of TIM-3 in human bronchial epithelial cells and bronchial brushings. In addition, we also provide evidence for the first time, to our knowledge, that both TIM-3 and its ligand galectin-9 are constitutively upregulated in unstimulated CF bronchial epithelial cells. Of interest, this also constitutes what we believe is the first report of TIM-3 overexpression linked to a specific genetic mutation. Having established a link between lack of CFTR and TIM-3 upregulation in vitro and ex vivo, we demonstrated a direct association between CFTR function and TIM-3 expression. Pharmacological inhibition of CFTR with CFTRinh172 resulted in an upregulation of TIM-3 and its ligand galectin-9 expression. It has been proposed that wild type CFTR cell membrane expression suppresses NF-κB–mediated inflammation (51), and that functional CFTR is required for such an inhibitory effect (50, 51). Inhibition of CFTR activity after CFTRinh172 treatment caused an increase in NF-κB activity in HBE cells, whereas activation of CFTR activity after forskolin treatment reversed the effect (51). Given that NF-κB has been predicted to have several binding sites in the TIM-3 promoter region (54), the mechanism behind the upregulation of TIM-3 after CFTR pharmacological inhibition could involve modulation of NF-κB activity.
The constitutive overexpression of TIM-3 and galectin-9 prompted us to investigate the role of TIM-3 in the CF lung. TIM-3 function in epithelial cells has not been studied to date; therefore, we sought to characterize in detail TIM-3 expression in bronchial epithelial cells before investigating possible TIM-3/galectin-9 signaling mechanisms. First, we demonstrated that TIM-3 was present on the surface of epithelial cells, which would enable the binding of galectin-9. Confocal imaging revealed a clear membrane localization of TIM-3 in CFBE and HBE cells. Similar cellular distribution was observed in HEK293T cells transiently transfected with an expression vector for mouse TIM-3 (24). We also analyzed TIM-3 expression under proinflammatory conditions. Cells were treated with LPS (10 μg/ml) from P. aeruginosa, a well-characterized component of CF BAL capable of inducing an inflammatory response in HBE and CFBE cells (40). This high concentration of LPS has physiological relevance because it has been estimated that CF lung surface can contain LPS levels as high as 40–70 μg/ml (40). LPS treatment stimulated TIM-3 expression in HBE cells and particularly increased TIM-3 levels in CFBE cells. In addition, TIM receptor expression has been shown to be modulated by other bacterial toxins, for instance, TIM-4, which has been reported to be upregulated by Staphylococcal enterotoxin B (55) and cholera toxin (56). Interestingly, LPS treatment downregulated TIM-3 expression in dendritic cells (57), whereas LPS exposure upregulated TIM-3 expression in endothelial cells (28), indicating a cell-type–specific response. Of note, another TLR-4 agonist, the alarmin high-mobility group box 1 (HMGB1), also upregulated TIM-3 expression in endothelial cells in an NF-κB and AP-1–dependent fashion (28). To our knowledge, we have also presented for the first time a functional role for TIM-3 in bronchial epithelial cells. We have demonstrated that galectin-9 signaling via TIM-3 receptor on CFBE cells induced IL-8 production in vitro. Furthermore, the prominent apical distribution of functional TIM-3 suggests that galectin-9 present within the airway lumen can initiate proinflammatory responses via interaction with TIM-3. This latter result could impact on early recruitment of neutrophils in vivo and the described exuberant neutrophil infiltration that perpetuates the inflammatory response in the CF lung (58, 59).
The observation that galectin-9 was upregulated in CFBE cells prompted us to examine the levels of galectin-9 in CF BAL. Galectin-9 was not detected in any adult CF sample tested. However, galectin-9 was detected in other BAL samples from patients with respiratory diseases including COPD and non-CF bronchiectasis, which display neutrophilic infiltration to a lesser extent. Examination of the NE levels revealed that galectin-9 expression correlated inversely with NE activity. Furthermore, galectin-9 was detected in CF infants without neutrophil infiltration and, therefore, very low or nondetectable levels of NE activity. Subsequently, galectin-9 levels gradually decrease as patient age increases and disease progresses. As NE is abundantly present in CF BAL (39, 40), the data indicate that galectin-9 is degraded in older patients with chronic lung neutrophil infiltration. This finding is in agreement with previous reports of the ability of proteases, including NE, to rapidly cleave galectin-9 (47). In contrast, examination of the stability of the galectin-9 receptor in the CF lung demonstrated that TIM-3 is cleaved in CF BAL within 10 min and the fully glycosylated receptor is totally degraded in less than 2 h. Fragmentation patterns suggest that various proteases are involved in this degradation at different stages. A systematic study of the proteases involved in TIM-3 degradation revealed that serine proteases, specifically NE and PR3, are involved in the initial cleavage of TIM-3 in the extracellular domain, an effect that would lead to inactivation of the receptor. Physiologically relevant concentrations of NE and PR3 completely abolished TIM-3 presence on CFBE cell outer surface. The unopposed action of serine proteases has been shown to cleave a variety of cell surface receptors in the CF lung, including the phosphatidylserine receptor in macrophages (60), causing defective removal of apoptotic neutrophils and exacerbating inflammation. The chemokine receptor CXCR1 has also been shown to be cleaved by serine proteases, resulting in reduced bacterial killing by neutrophils. Moreover, CXCR1 proteolytic fragments also induced IL-8 production by bronchial epithelial cells, which led to further inflammation (61). Of major importance, the extremely high protease burden in the CF lung has also been shown to cause the degradation of the natural protease inhibitors A1AT, elafin and SLPI (62), exacerbating the deleterious effect of NE and other proteases.
Collectively, our data indicate that activation of TIM-3 by galectin-9 may trigger neutrophil recruitment caused by stimulation of chemokine production by bronchial epithelial cells. In an acute infection context in the normal lung, this initial response would have a protective effect as neutrophils could be recruited into the lung to resolve the infection. Indeed, i.p. LPS injection in mice has been shown to increase galectin-9 levels, and galectin-9 treatment promoted neutrophil recruitment (63). Subsequently, galectin-9 could also induce the termination of the local acute inflammatory response in the normal lung by selectively modulating apoptosis of these immune cells. TIM-3 has been shown to be a negative modulator of Th-1 type immune response (32, 64). This dual effect of TIM-3/galectin-9 on different cell types has previously been reported in dendritic cells (21) and endothelial cell (27) versus T cells. However, in a chronic inflammatory scenario, as observed in the CF lung, dysregulation of TIM-3 signaling may have a detrimental effect. The biological relevance of this observation would imply that a dysregulation of TIM-3 function in CF bronchial epithelial cells may contribute to the persistent neutrophil recruitment and acute inflammatory status of the CF lung. Initially, constitutively activated TIM-3/galectin-9 could result in chemokine production and neutrophil infiltration in the airways of CF newborns. Indeed, increased IL-8 and neutrophil airway infiltration before bacterial colonization has been previously documented (52, 65). As disease progresses, the presence of bacteria and LPS may amplify TIM-3–mediated neutrophil recruitment. Subsequently, on chronic inflammation and infection, the lack of TIM-3 and galectin-9 protein caused by proteolytic cleavage may result in a dysregulation of this signaling axis, which would then fail to terminate the inflammatory response because of lack of selective apoptosis of inflammatory immune cells (53), or expansion of immunosuppressive macrophages (34, 41) or regulatory T cells (66). Indeed, disruption of this mechanism using TIM-3 or galectin-9 blocking Abs has been reported to result in increased neutrophil infiltration and tissue damage in a model of liver ischemia/reperfusion (31), experimental adhesion formation (30), and multiple sclerosis (32). Conversely, administration of proteolytically stable recombinant galectin-9 ameliorated symptoms in a model of arthritis (34, 35, 66), diabetes (67), and multiple sclerosis (32). Galectin-9 has also shown to exhibit a protective effect in lung infection in a mouse model of hypersensitivity pneumonitis induced by Trichosporon asahii (41).
Overall, the data presented demonstrate a dysregulation of TIM-3/galectin-9 signaling in the CF lung with important therapeutic implications. Initially, the constitutive upregulation of both TIM-3 and galectin-9 would induce neutrophil influx, even in the absence of infection. This effect could be potentiated by infection caused by LPS-induced upregulation of this signaling axis. As the disease progresses and neutrophil infiltration becomes chronic, proteolytic degradation of TIM-3 and galectin-9 could impact on termination of the inflammatory response in the CF lung. The data presented suggest that manipulation of the TIM-3 signaling pathway may be of therapeutic value in CF and other lung diseases with a neutrophilic component, preferably in conjunction with antiprotease treatment such as aerosolized A1AT augmentation therapy (68, 69).
Disclosures
The authors have no financial conflicts of interest.
Acknowledgments
We thank Dr. Warren Thomas (Royal College of Surgeons in Ireland, Molecular Medicine Laboratories) for providing support on the Zeiss LSM710 confocal immunofluorescence microscope purchased through the National Biophotonics Imaging Platform funding, Dr. Sanjay H. Chotirmall for collection of adult patient BAL and brushing samples at Beaumont Hospital (Dublin), and Donna Clarke for technical assistance at the National Children’s Research Center, Our Lady’s Children’s Hospital (Dublin).
Footnotes
This work was supported by the Irish Health Research Board (Grant PHD/2007/11), the US Alpha One Foundation, the Medical Research Charities Group/Health Research Board, and the Program for Research in Third Level Institutes administered by the Higher Education Authority and Science Foundation Ireland.
Abbreviations used in this article:
- A1AT
- α-1-antitrypsin
- A1CT
- α-1-antichymotrypsin
- BAL
- bronchoalveolar lavage
- Cath G
- cathepsin G
- CF
- cystic fibrosis
- CFTR
- CF transmembrane conductance regulator
- CFTRinh172
- CFTR inhibitor-172
- CMK
- chloromethyl ketone
- COPD
- chronic obstructive pulmonary disease
- LSC
- laser-scanning cytometry
- NE
- neutrophil elastase
- PBST
- PBS containing 0.1% (v/v) Tween 20
- PR3
- proteinase-3
- QRT-PCR
- quantitative real-time PCR
- rhTIM-3
- recombinant human TIM-3
- SBTI
- soya bean trypsin inhibitor
- SLPI
- serine leukoprotease inhibitor
- TAPI-I
- TNF-α protease inhibitor-1
- TIM
- T-cell Ig and mucin domain-containing molecule.
- Received September 28, 2010.
- Accepted December 20, 2010.
- Copyright © 2011 by The American Association of Immunologists, Inc.
References
In this issue
