We have shown previously that the pro-inflammatory cytokine TNF (tumour necrosis factor) could drive sLex (sialyl-Lewisx) biosynthesis through the up-regulation of the BX transcript isoform of the ST3GAL4 (ST3 β-galactoside α-2,3-sialyltransferase 4) sialyltransferase gene in lung epithelial cells and human bronchial mucosa. In the present study, we show that the TNF-induced up-regulation of the ST3GAL4 BX transcript is mediated by MSK1/2 (mitogen- and stress-activated kinase 1/2) through the ERK (extracellular-signal-regulated kinase) and p38 MAPK (mitogen-activated protein kinase) pathways, and increases sLex expression on high-molecular-mass glycoproteins in inflamed airway epithelium. We also show that the TNF-induced sLex expression increases the adhesion of the Pseudomonas aeruginosa PAO1 and PAK strains to lung epithelial cells in a FliD-dependent manner. These results suggest that ERK and p38 MAPK, and the downstream kinase MSK1/2, should be considered as potential targets to hamper inflammation, bronchial mucin glycosylation changes and P. aeruginosa binding in the lung of patients suffering from lung diseases such as chronic bronchitis or cystic fibrosis.
Mucins are the main components of mucus and represent a broad family of high-molecular-mass highly O-glycosylated proteins. They are synthesized by specialized cells and responsible for the visco-elastic and adhesive properties of the mucus. In human airways, the peptide diversity of mucins originates from the expression of several MUC genes and the major gel-forming mucins (MUC2, MUC5AC, MUC5B and MUC19) are secreted in the lung by goblet cells and mucous cells of submucosal glands, where they play an essential role in bronchial epithelium protection. Mucin-type O-glycans are the primary constituents of mucins and extremely diverse O-glycan chains cover the apomucin protein backbone. These glycan chains are synthesized by Golgi-localized glycosyltransferases and sulfotransferases whose relative expression determines the final structures of O-linked chains. Different carbohydrate antigenic structures, such as histo-blood group ABH or Lewis antigens, have been identified at the non-reducing end of human bronchial mucins  where they can serve as ligands for carbohydrate-binding molecules.
Numerous studies have shown that inflammation can modify mucin expression and glycosylation. In particular, bronchial mucins from severely infected patients suffering from either chronic bronchitis or CF (cystic fibrosis) are more sialylated and contain more sLex (sialyl-Lewisx or Neu5Acα2-3Galβ1-4[Fucα1-3]GlcNAc) carbohydrate structures than mucins from less-infected patients . In parallel, sLex and its 6-sulfated derivative 6-sulfo-sLex are ligands for Pseudomonas aeruginosa that are involved in the morbidity and early death of CF patients [3,4]. In CF patients, inflammation-induced glycosylation changes could therefore provide additional binding sites for bacteria and/or inflammatory cells.
Modifications in terminal glycosylation and sulfation of O-glycan chains are mostly linked to the transcriptional regulation of the corresponding glycosyltransferase and sulfotransferase genes. It is well documented that inflammation, via pro-inflammatory cytokines, modifies the expression and the activity of fucosyl-, sialyl- and sulfo-transferases involved in terminal glycan chain biosynthesis. In particular, we and others have shown, that in human bronchial mucosa and lung epithelial cells, TNF (tumour necrosis factor) induced the expression of the mucin-associated sLex and 6-sulfo-sLex epitopes [5–7]. We recently identified ST3Gal IV as the main α2,3-sialyltransferase responsible for the biosynthesis of sLex and 6-sulfo-sLex in bronchial mucosa . The ST3GAL4 (ST3 β-galactoside α-2,3-sialyltransferase 4) gene that encodes the sialyltransferase ST3Gal IV is localized on chromosome 11(q23–q24) and distributed over 14 exons. It produces at least five transcripts by a combination of alternative splicing and alternative promoter utilization, which code for identical protein sequences except at the 5′-ends . We have shown that, in both human bronchial explants and A549 lung cancer cells, TNF regulates the expression of the BX-specific transcript of ST3GAL4 .
In the present study, we show that the TNF-induced up-regulation of the ST3GAL4 BX transcript is mediated by MSK1/2 (mitogen- and stress-activated kinase 1/2) through the ERK (extracellular-signal-regulated kinase) and p38 MAPK (mitogen-activated protein kinase) pathways, and increases P. aeruginosa adhesion on NCI-H292 lung epithelial cells in a FliD-dependent manner. These results shed light on the mechanism by which severe inflammation can induce sLex and 6-sulfo-sLex overexpression in bronchial mucosa and might allow the identification of new targets to reduce P. aeruginosa infection.
MATERIALS AND METHODS
Antibodies and reagents
Anti-p65 (sc-7151x), anti-β-tubulin (H-235), anti-ERK (K-23), anti-phospho-Erk (E-4) antibodies were from Santa Cruz Biotechnology. Anti-JNK (c-Jun N-terminal kinase), anti-phospho-JNK (Thr183/Tyr185), anti-p38 and anti-phospho-p38 (Thr180/Tyr182) antibodies were from Cell Signaling Technology. The anti-sLex Heca-452 antibody was from BD Biosciences. The anti-(histone H2B) (07-371) antibody was from Millipore. The secondary antibodies were HRP (horseradish peroxidase)-conjugated goat anti-mouse, goat anti-rat or goat anti-rabbit IgG Sigma–Aldrich. TNF was from AbCys SA. U0126, SB203580, SP600125 and MG132 were obtained from Calbiochem Merck Chemicals and diluted in DMSO. SB747651A was purchased from Axon MedChem and diluted in water.
Cell and human bronchial explant culture
The lung epithelial carcinoma cell line A549 was purchased from the A.T.C.C. (cell line CCL-185). The mucoepidermoid pulmonary carcinoma cell line NCI-H292 (A.T.C.C. cell line CRL-1848) was kindly provided by Dr Jean-Marc Lo Guidice (EA4483, Lille, France). Cell lines were maintained in DMEM (Dulbecco's modified Eagle's medium, Gibco Invitrogen) supplemented with 10% FBS and 100 mg/ml penicillin/streptomycin (Gibco Invitrogen).
In accordance with the Declaration of Helsinki (2000) of the World Medical Association, after ethics committee approval (Centre Hospitalier Régional Universitaire de Lille, Lille, France) and informed consent from each patient had been obtained, tissues were collected from macroscopically healthy areas of the bronchial tree of patients undergoing surgery for bronchial carcinoma. They were immersed in DMEM, immediately transported on ice to the laboratory and then processed for mucosa isolation. Mucosa (1–2 cm2) were cut into 1–2 mm2 pieces and suspended in CMRL-1066 medium (Gibco Invitrogen).
Extraction of nuclear and cytoplasmic proteins and p65 nuclear translocation analysis
A549 cells were pre-incubated for 1 h in complete DMEM supplemented with DMSO (vehicle) or MG132 (10 μM) and then with or without TNF (40 ng/ml) for 20 min. After two washes in PBS, cells were lysed on ice in a hypotonic buffer [10 mM Hepes, 1.5 mM MgCl2 and 10 mM KCl (pH 7.9)] supplemented with 0.125% NP40 (Nonidet P40) and protease cocktail inhibitors (Roche). The lysates were centrifuged at 10000 g for 5 min at 4°C. The supernatants corresponding to the cytosolic fractions were collected and the pellets, corresponding to the nuclear fractions, were lysed with hypertonic buffer [20 mM Hepes, 1.5 mM MgCl2, 0.2 mM EDTA, 0.5 M NaCl and 25% glycerol (pH 7.9)] supplemented with protease cocktail inhibitors on ice for 2 h. The protein concentrations were determined with the Micro BCA Protein Assay Reagent kit (Bio-Rad Laboratories). Aliquots (20 μg) of the cytoplasmic and nuclear protein fractions were boiled for 10 min in reducing Laemmli sample buffer and resolved by SDS/PAGE (8% mini-gels). After transfer on to a nitrocellulose membrane, blocking was performed using TBS containing 5% (w/v) non-fat dried milk and 0.05% Tween 20 for 1 h at room temperature (20°C). Incubations with anti-p65 (sc-7151x), anti-(histone H2B) (07-371) and anti-β-tubulin antibodies were performed for 1 h at room temperature in TBS/5% non-fat dried milk/0.05% Tween 20. After three washing steps in TBS, membranes were incubated with HRP-conjugated goat anti-mouse, anti-rat or anti-rabbit antibodies diluted 1:5000 for 1 h at room temperature in TBS/0.05% Tween 20. Membranes were finally washed three times for 10 min in TBS/0.05% Tween 20 and detection was achieved using enhanced chemiluminescence (ECL+® advance Western blotting detection reagents, Amersham Biosciences).
MAPK phosphorylation analysis
A549 and NCI-H292 cells were pre-incubated for 1 h in complete DMEM supplemented with DMSO (vehicle), U0126 (20 μM), SB203580 (20 μM), SP600125 (20 μM) or SB747651A (20 μM) and then with or without TNF (40 ng/ml) for 20 min. Cells were washed three times in PBS, scraped, collected by centrifugation (2000 g for 5 min at 4°C) and total proteins were extracted as described previously , with the additional use of phosphatase inhibitor cocktail in lysis buffer (Thermo Fischer Scientific). A 20 μg aliquot of protein was boiled for 10 min in reducing Laemmli sample buffer and resolved by SDS/PAGE (10% mini-gels). After transfer on to a nitrocellulose membrane, blocking was performed using TBS/5% non-fat dried milk/0.05% Tween 20 for 1 h at room temperature. Incubations with anti-ERK, anti-phospho-ERK, anti-p38, anti-phospho-38, anti-JNK, anti-phospho-JNK and anti-β-tubulin antibodies were performed for 1 h at room temperature in TBS/5% non-fat dried milk/0.05% Tween 20. After three washing steps in TBS, membranes were incubated with HRP-conjugated goat anti-mouse or goat anti-rabbit antibodies diluted 1:5000 for 1 h at room temperature in TBS/0.05% Tween 20. Membranes were finally washed three times for 10 min in TBS/0.05% Tween 20 and detection was achieved using enhanced chemiluminescence (ECL+® advance Western blotting detection reagents).
Effect of inhibitors on ST3GAL4 transcript isoform BX expression
A549 cells were pre-incubated for 1 h in complete DMEM supplemented with DMSO (vehicle), MG132 (10 μM), U0126 (20 μM), SB203580 (20 μM) and SP600125 (20 μM) alone or in combination with SB747651A (20 μM) and then with or without TNF (40 ng/ml) for 12 h. Human bronchial mucosa explants were pre-incubated for 1 h in CMRL-1066 supplemented with DMSO (vehicle) or U0126/SB203580 (20 μM) and then with or without TNF (40 ng/ml) for 12 h. Cells were then scraped, collected by centrifugation (2000 g for 5 min at 4°C) and washed twice with PBS prior to RNA extraction. Pieces of bronchial mucosa were harvested and washed twice with PBS before RNA extraction for the study of ST3GAL4 transcript isoform BX expression.
RNA extraction, cDNA synthesis and qPCR (quantitative PCR) analysis
Total RNA was isolated from A549 cells and human bronchial mucosa using the NucleoSpin_RNA II extraction kit (Macherey–Nagel). For bronchial mucosa explants, tissue dissociation was first achieved using the gentleMACS dissociator (Miltenyi Biotec, RNA01_01 programme). RNA was then subjected to reverse transcription in the presence of a 1:1 ratio of oligo(dT) and hexameric random primer (AffinityScript qPCR cDNA Synthesis kit, Stratagene) for cDNA synthesis in a final volume of 20 μl according to the manufacturer's instructions. The expression of ST3GAL4 BX or total ST3GAL4 expression was analysed by qPCR using the Mx3005p Quantitative System (Stratagene). The PCR assay chamber (25 μl) contained 12.5 μl of 2× Brilliant SYBR Green qPCR Mastermix (Thermo Fischer Scientific), 300 nM of primers and 4 μl of cDNA (1:40 dilution). DNA amplification was performed using the following primers: ST3GAL4 BX forward, 5′-gaaccgtgctgccccgcccc-3′; ST3GAL4 BX reverse, 5′-gggacttgctgaccatgttt-3′; ST3GAL4 forward, 5′-ataagaagcgggtgcgaaaggg-3′; and ST3GAL4 reverse, 5′-tccgtggctgttgcattggc-3′. The HPRT (hypoxanthine–guanine phosphoribosyltransferase) gene was used to normalize the expression of genes of interest using the following primers: forward, 5′-gccagactttgttggatttg-3′ and reverse, 5′-ctctcatcttaggctttgtattttg-3′. The thermal cycling profiles used were described previously . The analysis of amplification was performed using the Mx3005p software. All experiments were performed in triplicate using three different biological samples. The quantification was achieved by the method described by Pfaffl .
Effect of inhibitors on sLex expression
A549 and NCI-H292 cells were pre-incubated for 1 h in complete DMEM supplemented with DMSO (vehicle), U0126 and SB203580 in combination (20 μM), or SB747651A (20 μM) and then with or without TNF (40 ng/ml) for 24 h. Cells were collected by centrifugation (2000 g for 5 min at 4°C) and then washed three times in PBS. Cells were resuspended in lysis buffer [50 mM Tris/HCl (pH 7.4) containing 150 mM NaCl, 1 mM EDTA, 1 mM PMSF, 1% Triton X-100, 1% sodium deoxycholic acid and protease cocktail inhibitors (Roche)], vortex-mixed for 1 min and sonicated for 2 min.
After 1 h of pre-incubation in CMRL-1066, with or without 20 μM U0126/SB203580 or 20 μM SB747654A, bronchial mucosa explants were incubated with or without TNF (40 ng/ml) for 24 h. Pieces of mucosa were then collected and centrifuged at 10000 g for 5 min at 4°C and then rinsed twice in ice-cold PBS. Pieces of mucosa were then dissociated in lysis buffer using the gentleMACS dissociator (Miltenyi Biotec, Protein-01.01 programme).
Protein extracts were then centrifuged at 10000 g for 5 min at 4°C to get rid of non-homogenized tissue. The protein concentrations were determined with the Micro BCA Protein Assay Reagent kit (Bio-Rad Laboratories). A 20 μg aliquot of protein was boiled for 10 min in reducing Laemmli sample buffer and resolved by SDS/PAGE (6% mini-gels). After transfer on to a nitrocellulose membrane, blocking was performed using TBS/5% non-fat dried milk/0.05% Tween 20 for 1 h at room temperature. Incubations with anti-sLex and anti-β-tubulin antibodies were performed for 1 h at room temperature in TBS/5% non-fat dried milk/0.05% Tween 20. After three washing steps in TBS, membranes were incubated with HRP-conjugated goat anti-mouse or goat anti-rabbit antibodies diluted 1:5000 for 1 h at room temperature in TBS/0.05% Tween 20. Membranes were finally washed three times for 10 min in TBS/0.05% Tween 20 and detection was achieved using enhanced chemiluminescence (ECL+® advance Western blotting detection reagents).
The P. aeruginosa strains PAO1 and PAK, as well as their isogenic fliD mutant strains PAO1-D (PAO1 fliD::Gm) and PAK-D (PAK fliD::Gm) unable to synthesize the FliD flagellar cap protein, were kindly provided by Professor Reuben Ramphal (University of Florida, Gainsville, FL, U.S.A.). For adhesion assays, bacteria were grown 16 h in LB medium at 37°C under 150 rev./min agitation. Bacteria were then washed twice in DPBS (Dulbecco's PBS) and diluted in DPBS at working concentrations (5×106–1×107/ml).
NCI-H292 cells were seeded into 12-well plates and grown to confluence for 72 h in complete DMEM with or without TNF (40 ng/ml). Medium was renewed every 24 h. After two washes in DPBS, cells were incubated with 100 μl of bacteria (5×105–1×106 bacteria/well) for 2 h at 37°C. Unbound bacteria were removed by five washes with 1 ml of DPBS and cells were then lysed using deionized water containing 0.02% Triton X-100. Serial dilution of cell lysates in DPBS were then prepared and plated on to LB agar plates for 16 h. To analyse the role of sialic acid residues in bacterial adhesion, NCI-H292 cells were treated prior to the addition of bacteria with 0.5 unit/ml Clostridium perfringens sialidase (Sigma–Aldrich) for 45 min at 37°C.
ST3GAL4 expression data in A549 cells and bacterial adhesion data on NCI-H292 cells were tested by unpaired Student's t test with GraphPad Prism 5.0. ST3GAL4 expression data in explants were tested by ratio-paired Student's t test with GraphPad Prism 5.0.
Effect of the inhibition of TNF-activated signalling pathways on ST3GAL4 expression and sLex synthesis
In order to determine the signalling pathways involved in the TNF-induced up-regulation of ST3GAL4, we analysed the effect of different protein kinase inhibitors on the BX transcript expression in A549 cells. The binding of TNF to TNF receptor-1 leads to the activation of two main pathways: the canonical NF-κB (nuclear factor κB) pathway and the MAPK pathway. As shown in Figure 1, TNF treatment of A549 cells induced the nuclear translocation of the p65 NF-κB subunit (Figure 1A) as well as the phosphorylation of the ERK1/2, p38 and JNK (Figures 1B–1D). Furthermore, MG132 strongly inhibited p65 nuclear translocation (Figure 1A) and the specific protein kinase inhibitors U0126 , SB203580  and SP600125  impaired the phosphorylation of ERK1/2, p38 and JNK respectively (Figures 1B–1D). As described previously, TNF induced the up-regulation of ST3GAL4 BX transcript in A549 cells . Interestingly, neither MG132 nor specific MAPK inhibitors used alone was able to impair the TNF-induced up-regulation of ST3Gal IV (Figures 2A and 2B). However, the combination of the ERK1/2 inhibitor U0126 and of the p38 inhibitor SB203580 was able to impair BX transcript up-regulation, whereas other combinations of MAPK inhibitors had no effect (Figure 2C). Finally, the inhibition of the protein kinases MSK1/2, which are activated by both ERK1/2 and p38, with the specific inhibitor SB747651A , also inhibited the TNF-induced up-regulation of the ST3GAL4 BX transcript (Figure 2D). Altogether, these data suggest that the TNF-induced increased expression of ST3Gal IV in A549 cells results from the up-regulation of the BX transcript by protein kinases MSK1/2 through the ERK1/2 or p38 signalling pathways.
Immunoblot analysis of the signalling pathways activated in A549 cells by TNF treatment and the effect of inhibitors
qPCR analysis of the effect of pharmacological inhibitors on the TNF-induced up-regulation of the ST3GAL4 transcript BX in A549 cells
Similar experiments were performed in human bronchial mucosa explants. As shown in Figure 3, although inter-individual variations in the expression level of ST3GAL4 (Figure 3A) and the BX transcript (Figure 3B) were observed between the different samples, expression levels of ST3GAL4 and the BX transcript were increased by TNF and the treatment with the combination of the ERK1/2 and p38 inhibitors was also able to inhibit ST3GAL4 and BX transcript up-regulation, showing that ST3Gal IV is also regulated through the ERK1/2 and p38 signalling pathways in human bronchial mucosa.
Effect of ERK/p38 inhibition on the TNF-induced up-regulation of the ST3GAL4 transcript BX in human bronchial mucosa explants
Since ST3Gal IV is the main sialyltransferase involved in the biosynthesis of sLex in human bronchial mucosa , we analysed the effect of the inhibition of ERK1/2, p38 and MSK1/2 on the expression of sLex. As shown in Figure 4, TNF increased the expression of sLex on high-molecular-mass glycoproteins in A549 cells (Figure 4A) and this increased expression was also repressed by the combination of ERK1/2 and p38 and MSK1/2 inhibitors (Figure 4A). A similar result was obtained with the mucoepidermoid pulmonary carcinoma cell line NCI-H292 (Figure 4B). The NCI-H292 cell line was described previously as expressing sLex on different mucins, especially MUC5AC . Moreover, NCI-H292 cells also express the ST3GAL4 BX transcript in a TNF-dependent manner (results not shown). Western blot analysis showed that TNF also induced the expression of sLex on NCI-H292 cell glycoproteins and that ERK1/2 and p38 inhibitors impaired the TNF-induced expression of sLex (Figure 4B). Finally, we show that ERK1/2 and p38 inhibitors used in combination and a MSK1/2 inhibitor are able to repress sLex overexpression induced by TNF on high-molecular-mass proteins in human bronchial explants (Figures 4C and 4D). Altogether, these data indicate that the increased expression of ST3Gal IV induced by TNF through the ERK1/2 or p38 and MSK1/2 signalling pathways is responsible for the overexpression of sLex in human bronchial mucosa.
Immunoblot analysis of the sLex expression in A549 cells, NCI-H292 cells and human bronchial explants and effect of TNF and inhibitor treatment
TNF-induced sLex overexpression enhances P. aeruginosa adhesion
Finally, we analysed the effect of TNF on the adhesion of P. aeruginosa to the NCI-H292 cell line. We used the PAO1 and PAK P. aeruginosa strains, as well as their isogenic fliD mutant strains PAO1-D and PAK-D. PAK FliD and PAO1 FliD possess A- and B-type FliD flagellar cap proteins respectively. They share only 51% amino-acid sequence similarity (43% identity)  and display distinct glycosylation patterns [16,17]. The FliD protein is involved in the binding of the PAO1 strain to sLex  and has also been described as an adhesion factor of the PAK strain to bronchial mucins . However, the structural differences between PAO1 and PAK FliD proteins seem to lead to different ligand specificity . Indeed, sLex and 6-sulfo-sLex were clearly identified as PAO1 FliD receptors, whereas the identity of the PAK FliD receptor remains elusive . After 3 days of TNF treatment, adhesion assays showed a 70% and 30% increased adhesion for the PAO1 and PAK strains respectively (Figure 5). In parallel, the isogenic strains mutated for the FliD protein showed only a weak decrease in adhesion in the controls, but did not present an increased adhesion after TNF treatment. Furthermore, desialylation of the NCI-H292 cell surface showed that sialylated structures were required to enhance bacterial adhesion on TNF-treated cells (Figure 5). These results indicate that TNF-induced sialic-acid-containing structures are implicated in the increase in FliD-dependent adhesion of the PAO1 and PAK P. aeruginosa strains. As sLex is the best receptor for PAO1 FliD, its up-regulation by TNF could explain the increase in adhesion of PAO1 to TNF-treated NCI-H292 cells. In contrast, the enhanced adhesion mechanisms of the PAK strain remain unclear. The similarity of results between PAK and PAO1 adhesion could indicate that the FliD PAK strain is also able to interact with sLex. However, we cannot rule out the involvement of other sialylated structures specifically recognized by the PAK FliD and up-regulated by TNF.
Effect of TNF on P. aeruginosa adhesion on NCI-H292 cells
In the present study, we have investigated the mechanism by which TNF up-regulates ST3GAL4 and induces sLex overexpression in bronchial mucosa. We have already shown that ST3Gal IV is the main sialyltransferase involved in the biosynthesis of sLex and that TNF up-regulates the ST3GAL4 BX transcript in bronchial explants and the A549 cell line . As already described previously, TNF is a potent activator of NF-κB and MAPK . Indeed, the levels of nuclear p65 and phosphorylated p38, JNK and ERK1/2 were increased in TNF-treated A549 cells (Figure 1), and the effect of TNF on the expression of the ST3GAL4 BX transcript was only repressed by the combination of ERK and p38 phosphorylation inhibitors  in both A549 cells and bronchial explants (Figures 2 and 3). The redundancy of the signalling pathways, involving either ERK or p38, leads us to hypothesize that ERK and p38 could activate the same target, the mitogen- and stress-activated protein kinase MSK1/2. Using a recently developed MSK1/2 inhibitor , we were able to show that the inhibition of MSK1/2 almost completely antagonized the effect of TNF (Figure 2D). Interestingly, MSK1/2 are able to regulate gene transcription at different levels. MSK1/2 can directly target transcription factors, such as CREB (cAMP-response-element-binding protein) and NF-κB, increasing their transcriptional activity. They can also induce histone H3 phosphorylation, leading to chromatin relaxation and the increased binding of regulatory proteins (for a review see ). Only a few studies have reported MSK activation by TNF. For example, TNF up-regulates cerebrovascular endothelin-1, a protein that contributes to a decrease in cerebral blood flow in acute brain injury via MSK2 activation . MSK1/2 are also required for the TNF-induced phosphorylation of CREB and ATF1 (activating transcription factor 1) in MEFs (mouse embryonic fibroblasts), and in the regulation of the nuclear orphan gene Nur77 [also known as NR4A1 (nuclear receptor subfamily 4, group A, member 1)] transcription . In the present study, we show for the first time that the TNF-induced up-regulation of ST3GAL4 requires MSK1/2 activation.
In agreement with our previous work showing the crucial role of ST3Gal IV in sLex biosynthesis in bronchial mucosa , the repression of TNF-induced ST3GAL4 overexpression by either a combination of ERK1/2 and p38 inhibitors (U0126 and SB203580) or the MSK inhibitor SB747651A was able to inhibit the overexpression of sLex on high-molecular-mass glycoproteins exhibiting mucin features in both A549 and NCI-H292 cells (Figures 4A and 4B) and bronchial explants (Figures 4C and 4D). We have shown previously that MUC4 is the main mucin carrying sLex in human bronchial explants, and Ishibashi et al.  have shown that TNF-induced sLex expression on MUC5AC in NCI-H292 cells. However, several other sLex-carrying glycoproteins need to be identified. In CF patients, it has been shown that MUC5B, and most dramatically MUC5AC, are reduced compared with normal airway mucus [25,26]. In parallel, MUC2 is expressed in CF bronchial explants and the transcriptional activation of the MUC2 gene by P. aeruginosa lipopolysaccharide has also been demonstrated . MUC2 could be therefore one of the sLex-carrying mucins, as well as MUC1, that has been shown to mediate P. aeruginosa adhesion in human airway epithelial cells .
In addition to providing a favourable environment for P. aeruginosa adhesion, TNF-induced sLex overexpression of bronchial mucins could also modulate the immune response in the lungs of CF patients. The interaction between mucins carrying sLex and selectins expressed by leucocytes or endothelial cells could modulate the inflammatory response, by competing with the relevant selectin ligands , and promote leucocyte migration and extravasation toward the airways. Moreover, mucins carrying sLex can also interact with Siglec9 (sialic acid-binding immunoglobulin-like lectin 9) expressed at the surface of neutrophils . Siglec9 contains a sialic acid-binding domain with high affinity for sLex/6-sulfo-sLex, and is involved in apoptotic or non-apoptotic cell death pathways . The over-expression of sLex on CF patients’ bronchial mucins could therefore modify the interactions of immune cells with their natural ligands in the respiratory tract and impair mucosal immunity.
Although the transcription factors involved in ST3GAL4 BX transcript regulation by TNF remain to be determined, it is well-known that CREB- and NF-κB-related transcription factors play major roles downstream of MSK1/2. Bioinformatics analysis and luciferase assays allowed us to identify a 2 kb genomic sequence surrounding the BX exon that contains a promoter region regulated by TNF . Site-directed mutagenesis experiments are underway in our laboratory to identify the binding site(s) of transcription factors involved in TNF-induced BX transcription.
In parallel, we investigated whether TNF could modulate P. aeruginosa adhesion through sLex overexpression. sLex has been described as the best ligand for PAO1 FliD flagellar cap protein , whereas the nature of the PAK FliD ligand remains unclear. Moreover, it has been shown that the FliD protein is involved in adhesion to mucins , and has an impact on the structure of the PAO1 biofilm in mucin-coated adhesion assays . In the present study, we show that TNF treatment enhances the adhesion of the P. aeruginosa PAO1 strain and, to a lesser extent, of the PAK strain on NCI-H292 lung epithelial cells (Figure 5). This enhanced adhesion is dependent on a functional FliD protein and sialic-acid-containing glycan expression. These results suggest that the TNF-induced expression of sLex, the best FliD ligand, effectively enhances adhesion of the PAO1 strain. For the PAK strain, the lesser increase adhesion on TNF-treated cells seems to indicate that different ligands are involved, according to previous work failing to implicate sLex in PAK FliD recognition . Another striking feature of the PAK strain adhesion assay is the effect of desialylation. Unlike the PAO1 strain, the PAK strain shows a significant decrease in adhesion after sialidase treatment (Figure 5). These results seem to indicate the involvement of another sialic-acid-binding lectin in adhesion of the PAK strain. A study of an ocular P. aeruginosa strain suggested previously a preferential interaction with α2,6-sialic-acid-containing glycans ; however, the clear identification of the lectin involved and the precise structure remain to be established.
Altogether, the results of the present study highlight the mechanism by which inflammation could modify bronchial mucin glycosylation and how these changes could influence host–pathogen interactions. Most CF patients experience an early, chronic and severe lung inflammation  associated with chronic infection, especially by P. aeruginosa. This lung infection and inflammation is correlated with glycosylation changes on bronchial mucins, predominantly sLex epitope expression , that have been described as preferential P. aeruginosa receptors. The present study shows that the TNF-induced sLex overexpression effectively increases P. aeruginosa adhesion on lung epithelial cells expressing high-molecular-mass glycoproteins and possibly in bronchial mucosa. In the context of CF pathology, with viscous and thick mucus clogging the lung, an increased expression of P. aeruginosa receptors could promote chronic colonization and biofilm formation, leading to an unresolved infection. Furthermore, we describe the molecular pathway involving ERK/p38 and MSK1/2 by which TNF drives ST3GAL4 expression and sLex biosynthesis. Interestingly, ERK, JNK and p38 MAPK pathway activation is excessive and prolonged in CF patient cells . These modifications affect the expression of the IL-8 (interleukin 8) pro-inflammatory cytokine which is overexpressed in CF patients, but it is probable that many other genes are also overexpressed in CF patient's lungs as a consequence of the dysregulation of these signalling pathways. ERK and p38, and more especially the downstream kinase MSK1/2, should be considered as potential targets to hamper inflammation, bronchial mucins glycosylation changes and P. aeruginosa colonization in CF patient's lungs.
Dulbecco’s modified Eagle’s medium
c-Jun N-terminal kinase
mitogen-activated protein kinase
mitogen- and stress-activated kinase 1/2
nuclear factor κB
sialic acid-binding immunoglobulin-like lectin 9
β-galactoside α-2,3-sialyltransferase 4
tumour necrosis factor
Florent Colomb, Philippe Delannoy, Sophie Groux-Degroote, Marie-Ange Krzewinski-Recchi and Olivier Vidal conceived and designed the experiments. Marie Bobowski, Florent Colomb, Sophie Groux-Degroote, Marie-Ange Krzewinski-Recchi and Olivier Vidal performed the experiments. Florent Colomb, Philippe Delannoy, Sophie Groux-Degroote and Olivier Vidal analysed the data. Anne Harduin-Lepers, Eric Mensier, Sophie Jaillard and Jean-Jacques Lafitte contributed reagents, materials and analysis tools. Florent Colomb, Philippe Delannoy and Sophie Groux-Degroote wrote the paper.
This work was supported by a Vaincre la mucoviscidose grant [grant number IC0908] and a Ph.D. fellowship (to F.C.)