Role of LecA and LecB Lectins in Pseudomonas aeruginosa-Induced Lung Injury and Effect of Carbohydrate Ligands

Abstract

Pseudomonas aeruginosa is a frequently encountered pathogen that is involved in acute and chronic lung infections. Lectin-mediated bacterium-cell recognition and adhesion are critical steps in initiating P. aeruginosa pathogenesis. This study was designed to evaluate the contributions of LecA and LecB to the pathogenesis of P. aeruginosa-mediated acute lung injury. Using an in vitro model with A549 cells and an experimental in vivo murine model of acute lung injury, we compared the parental strain to lecA and lecB mutants. The effects of both LecA- and Lec B-specific lectin-inhibiting carbohydrates (α-methyl-galactoside and α-methyl-fucoside, respectively) were evaluated. In vitro, the parental strain was associated with increased cytotoxicity and adhesion on A549 cells compared to the lecA and lecB mutants. In vivo, the P. aeruginosa-induced increase in alveolar barrier permeability was reduced with both mutants. The bacterial burden and dissemination were decreased for both mutants compared with the parental strain. Coadministration of specific lectin inhibitors markedly reduced lung injury and mortality. Our results demonstrate that there is a relationship between lectins and the pathogenicity of P. aeruginosa. Inhibition of the lectins by specific carbohydrates may provide new therapeutic perspectives.

Pseudomonas aeruginosa is an opportunistic pathogen involved in acute infections, as well as chronic infections, especially in cystic fibrosis patients (7, 20). The therapeutic options for these infections remain limited because this pathogen exhibits increasing resistance to many antibiotics (26). Currently, antibiotic research and development are at an all-time low, and few new antipseudomonal compounds are in the pipeline. Therefore, there is a need for therapeutic approaches other than antibiotics.

For P. aeruginosa, as for other pathogenic microorganisms, the ability to adhere to host tissues is essential for initiating infection. Adhesion is often mediated by host cell surface glycoconjugates, which are a specific target for bacterial receptors (13). Such oligosaccharide-mediated bacterium-cell recognition and adhesion have been shown to be crucial in the early steps of P. aeruginosa pathogenesis. P. aeruginosa adhesion is mediated by a glycanic recognition pattern involving several adhesins, including lectins. Only a limited number of the carbohydrate-binding proteins of P. aeruginosa have been studied, and their role in recognition and adhesion is far from being elucidated. Two soluble lectins, LecA (PA-IL) and LecB (PA-IIL), specifically binding galactose and fucose, respectively, were initially identified and characterized in the cytoplasm of P. aeruginosa (9). However, large quantities of both of these lectins are present on the outer membrane of the bacteria, suggesting that lectins may play a role in adhesion (9, 32). These two lectins, which are produced by the bacteria, are also associated with virulence factors (8) and regulated by both quorum sensing and the alternative sigma factor RpoS (34), suggesting that they are also parts of the numerous systems involved in P. aeruginosa virulence.

Indeed, several studies suggested that both of these lectins may be determinants of virulence. The galactophilic molecule LecA has been shown to have a cytotoxic effect on respiratory epithelial cells by decreasing their growth rate, thus contributing to respiratory epithelial injury (2). In addition, it has been demonstrated that LecA induces a permeability defect in the intestinal epithelium, resulting in increased absorption of exotoxin A, an important extracellular virulence factor (19). Additionally, relationships between lectins and other virulence factors have been shown; for example, LecB was shown to be involved in pilus biogenesis and protease IV activity (29).

Although it is established that P. aeruginosa recognizes the glycoconjugates from epithelial and endothelial cells (14, 15), allowing initiation of colonization and infection, the role of glycoconjugate ligands, such as lectins, in P. aeruginosa pathogenesis is not clearly established. Therefore, the aims of our study were to evaluate the role of LecA and LecB in P. aeruginosa pathogenesis. To do this, we evaluated whether these lectins could contribute to the adhesion of P. aeruginosa to lung epithelial cells. We then evaluated whether these molecules were cytotoxic in vitro for lung epithelial cells and could contribute to epithelial damage. We next demonstrated their roles in acute lung injury in vivo. Following the previous studies using lectin mutant P. aeruginosa strains, we studied whether purified soluble lectins of P. aeruginosa could induce the cytotoxic effect of P. aeruginosa in vitro or pathogenic effects in vivo. Finally, we evaluated if addition of specific lectin-inhibiting carbohydrates (i.e., the methyl derivatives of galactose and fucose that mimic terminal sugars of eukaryotic cell surface glycoconjugates) could prevent or reduce P. aeruginosa-mediated lung injury.

MATERIALS AND METHODS

Animals.

Male BALB/c mice (20 to 25 g) purchased from Charles River Laboratories (Domaine des oncins, L'Arbresle, France) were housed in a pathogen-free unit of the Lille University Animal Care Facility and given food and water ad libitum. All experiments were performed with the approval of and following the guidelines of the Lille Institutional Animal Care and Use Committee.

Bacterial strains and growth conditions.

The strains and plasmids used in this study are listed in Table 1. The strains were cultured overnight in Luria-Bertani medium at 37°C and in the presence of 2 mM isopropyl-β-d-thiogalactopyranoside (IPTG) for complementation experiments. Parental strain P. aeruginosa PAO1 and insertional mutants PAO1lecA::Tc^r (hereinafter called PAO1::lecA) and PAO1lecB::Tc^r (hereinafter called PAO1::lecB) used in this study were obtained from the comprehensive transposon library generated in the PAO1 genetic background (11; http://www.genome.washington.edu). The locations of the transposon were mapped at nucleotide positions 48 and 204 for the lecA and lecB genes, respectively. The Escherichia coli TG1 and BL21 strains were used for standard genetic manipulations.

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Relevant characteristicsa	Source or reference
E. coli strains
TG1	supE Δ(lac-proAB) thi hsdRΔ5 (F′ traD36 rpoA+B+lacIqZΔM15)	Lab collection
P. aeruginosa strains
PAO1	Wild type	11
PAO1lecA::Tcr	PAO1 Tn5 mutant with mutation in the lecA gene	11
PAO1lecB::Tcr	PAO1 Tn5 mutant with mutation in the lecB gene	11
Plasmids
pMMB67-HE	Broad-host-range vector, IncQ ptac lacZα, Gmr	Lab collection
pDEST14	Destination vector for gateway technology, T7 promoter, Apr	Invitrogen
PMMBlecA	lecA gene cloned in pMMB67-HE, Cbr	This study
PMMBlecB	lecB gene cloned in pMMB67-HE, Cbr	This study
pRK2013	ColE1 ori tra+ mob+, Kmr	Lab collection
pDEST14-lecA	lecA gene cloned in pDEST14, Apr	This study
pDEST14-lecB	lecB gene cloned in pDEST14, Apr	This study

^aGm^r, gentamicin resistance; Ap^r, ampicillin resistance; Km^r, kanamycin resistance; Cb^r, carbenicillin resistance.

The lecA and lecB genes were obtained from a comprehensive P. aeruginosa gene collection (17), cloned into an entry vector of the Gateway system (Invitrogen, Cergy Pontoise, France), and then moved into a pDEST14 destination vector by using L and R lambda phage-specific recombination sites according to the manufacturer's instructions. After proper production of each lectin was checked, the lecA and lecB genes were further subcloned into the broad-host-range vector pMMB67HE at SmaI/SphI sites. Recombinant plasmids were introduced into P. aeruginosa using the conjugative properties of pRK2013 (lab collection). Transconjugants were selected on Pseudomonas isolation agar (Difco Laboratories) supplemented with appropriate antibiotics. The following antibiotic concentrations were used for E. coli: 50 μg/ml kanamycin and 50 μg/ml ampicillin.

P. aeruginosa lectins.

Recombinant LecB was purified from E. coli BL21(DE3) containing plasmid pET25pa2l as described previously (23).

The recombinant protein LecA was cloned using the following procedure. The lecA gene was amplified by PCR using genomic DNA from P. aeruginosa ATCC 33347 as the template with the following primers: 5′-CGG AGA TCA CAT ATG GCT TGG AAA GG-3′ and 5′-CCG AGA CAA GCT TTC AGG ACT CAT CC-3′ (NdeI and HindIII restriction sites are underlined). After digestion with NdeI and HindIII, the amplified fragment was introduced into the pET25(b+) vector (Novagen, Madison, WI), resulting in plasmid pET25pa1l.

E. coli BL21(DE3) cells harboring the pET25pa1l plasmid and E. coli BL21(DE3) cells harboring the pET25pa2l plasmid were grown in 1 liter of Luria broth at 37°C. When a culture reached an optical density (OD) at 600 nm of 0.5 to 0.6, IPTG was added to a final concentration of 0.5 mM. Cells were harvested after 3 h of incubation at 30°C, washed, and resuspended in 10 ml of loading buffer (20 mM Tris-HCl, 100 μM CaCl₂; pH 7.5). The cells were disrupted by sonication (Soniprep 150; Schoeller Instruments, Great Britain). After centrifugation at 10,000 × g for 1 h, the supernatant was further purified by affinity chromatography on Sepharose 4B (GE Healthcare) for LecA or on d-mannose agarose (Sigma-Aldrich, United States) for LecB. Lectins were allowed to bind to the immobilized saccharides in the loading buffer and then were eluted with the elution buffer (loading buffer containing 0.2 M d-galactose or 0.1 M d-mannose). All buffers used for LecB preparation contained an additional 100 mM NaCl. The purity of the recombinant proteins was over 98% as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The purified proteins were intensively dialyzed against distilled water for 7 days for sugar removal, lyophilized, and kept at −20°C.

The purified lectins were prepared in solution at a concentration of 100 μg/ml for in vitro assays and at a concentration of 1 mg/ml for in vivo assays.

Lectin inhibitors.

N-Acetyl-d-galactosamine (GalNAc) (Sigma-Aldrich), α-methyl-d-galactoside (Me-α-Gal) (a specific ligand of LecA), and α-methyl-l-fucoside (Me-α-Fuc) (Interchim) (a specific ligand of LecB) were used as inhibitors. d-Glucose (Glc) was used as an irrelevant control carbohydrate.

Preparation of bacterial inoculum for in vitro and in vivo experiments.

P. aeruginosa strains grown in Luria-Bertani medium at 37°C for 16 h with appropriate antibiotics if needed were centrifuged at 3,000 × g for 10 min. The bacterial pellets were washed two times in an isotonic saline solution and diluted in the isotonic saline solution to obtain an optical density of 0.63 to 0.65 as determined by spectrophotometry.

In vitro cytotoxicity of P. aeruginosa and its lectins.

The human lung epithelial cell line A549 was cultured to confluence in modified Eagle's medium with Earle's salts and l-glutamate (Invitrogen, Cergy Pontoise, France) supplemented with 10% heat-inactivated fetal bovine serum (Dustcher, Vilmorin, France) and 1% penicillin-streptomycin at 37°C with 5% CO₂. When the cells reached confluence, 2 × 10⁴ cells were transferred to 96-well tissue culture plates and incubated overnight. The following day, the cells were exposed for 4 and 6 h to 10 μl of each of the three different P. aeruginosa strains (PAO1, PAO1::lecA, and PAO1::lecB; 5 × 10⁸ CFU/ml). Cytotoxicity was quantified by determination of the release of lactate dehydrogenase (LDH) into the culture supernatants at 4 and 6 h (Cytotox 96; Promega Charbonniers, France). The maximal value (100%) represented the amount of LDH released from cells lysed by 0.8% Triton X-100 for 45 min. Control wells lacking bacteria were used to calculate the background level of LDH released (normalized to 0%). Then the level of cytotoxicity, expressed as a percentage, was calculated as follows: % cytotoxicity = [(OD of assay mixture − OD of cells)/(OD of Triton X-100 − OD of cells)] × 100.

Bacterial adhesion assays.

For the adhesion assays, A549 cells were seeded into 96-well microtiter plates at a density of 5 × 10⁴ cells per well and incubated overnight to obtain confluent monolayers.

Confluent monolayers in 96-well microtiter plates were washed five times with 150 μl phosphate-buffered saline (PBS) prewarmed to 37°C. Nonspecific binding was blocked by incubation for 1 h at 37°C with 0.5% (wt/vol) bovine serum albumin before the preparations were rinsed twice with prewarmed PBS. A 100-μl portion of the bacteria (10⁸ CFU/ml) was added to A549 cells and incubated for 1 h, 4 h, and 6 h at 37°C. Nonadherent bacteria were removed by rinsing the preparations five times with PBS. Cells were lysed by incubation for 30 min at 37°C with a 0.1% (vol/vol) Triton X-100 solution. Serial dilutions were prepared using PBS, and 100-μl aliquots were plated in duplicate on bromocresol purple plates and incubated at 37°C for 24 h.

Intratracheal instillation.

Mice were briefly anesthetized with inhaled sevoflurane (Sevorane; Abbot Laboratories, Queenborough, United Kingdom) and were placed in a supine position at an angle of approximately 30°. For each mouse, 50 μl of a bacterial inoculum calibrated to contain 5 × 10⁸ CFU/ml was instilled into the lungs through a gavage needle (24-gauge modified animal feeding needle; Popper & Sons, Inc., New Hyde Park, NY) inserted into trachea via the oropharynx. Proper insertion of the needle was confirmed by observing the movement of the solution inside the syringe during the animal's respiratory efforts.

In vivo quantification of acute lung injury: alveolar capillary barrier permeability.

Two different methods were used to assess alveolar capillary barrier permeability. The first method measures residual ¹²⁵I-albumin instilled intratracheally as an alveolar protein tracer in the lungs and its leakage and accumulation in the plasma. The second method measures ¹²⁵I-albumin injected as a vascular protein tracer (following reabsorption after intraperitoneal injection) and its leakage and accumulation in the extravascular spaces of the lungs.

The first method was used to assess capillary barrier injury at 6 h after infection as previously described (1). The intratracheal instillate was a mixture of 1 μCi of ¹²⁵I-labeled albumin (Seralb 125; CIS bio international, Gif-sur-Yvette, France) and 5% bovine albumin with an appropriate quantity of the specified P. aeruginosa or lectins. The total radioactivity in the instillate was measured. Fifty microliters of instillate was inoculated into the lungs of each anesthetized mouse. Six hours after instillation, mice were anesthetized with pentobarbital given intraperitoneally. The blood was collected by carotid arterial puncture, and a sternotomy was performed to harvest and measure the radioactivity in the lungs, trachea, and stomach. The quantity of ¹²⁵I-albumin that leaked into the circulation was calculated by multiplying the activity in a blood sample by the volume of blood.

The second method was used to evaluate the alveolar capillary barrier injury at 16 h after infection. We calculated the albumin flux across the barrier using a previously described index (12). Briefly, 2 h before the experiment, 0.5 ml of ¹²⁵I-albumin was injected intraperitoneally. After exsanguination, the lungs were removed, and the radioactivity in the blood and the hemoglobin (Hb) concentration were measured. Lung weight and radioactivity counts were determined before homogenization and centrifugation. The supernatant Hb content was also determined. Blood and lung homogenate samples were incubated at 40°C for 3 days to determine the ratio of lung wet weight to lung dry weight.

The permeability index (PI) was calculated as follows: PI = {[radioactivity count for lungs − (radioactivity count for intravascular blood per gram of blood × Q_B)]/(radioactivity count per gram of intravascular blood × weight of mouse)} ×100, where Q_B is the weight of intrapulmonary blood. Q_B was calculated as follows: Q_B = (weight of lung plus water × Hb concentration in supernatant × water ratio for homogenate × 1.039)/(Hb concentration in blood × water ratio for blood).

Measurement of the ratio of lung wet weight to lung dry weight.

The ratio of lung wet weight to lung dry weight was used to evaluate the amount of extravascular lung water in each group of animals. At the end of the experiment, the lungs were removed, and the wet weight was recorded. The lungs were then desiccated at 40°C for 3 days, after which the dry weight was recorded. For each pair of lungs, the ratio of lung wet weight to lung dry weight was calculated.

Quantitative blood culture and pulmonary bacterial load.

One hundred microliters of blood was plated on agar plates and incubated for 24 h at 37°C. At the end of the experiment, the lungs were removed and homogenized in 0.9 ml of sterile isotonic saline, and viable bacteria were counted by plating 0.1-ml portions of serial dilutions of the homogenates on agar plates and incubating them for 24 h at 37°C.

Experimental protocols. (i) In vitro studies.

The cytotoxicity of each of the P. aeruginosa strains and lectins studied was evaluated using A549 cells at 4 h and 6 h after exposure. The same experiments were performed with addition of specific inhibitors at a concentration of 15 mM.

The adhesion of each of the P. aeruginosa strains on A549 cells was also evaluated at 1 h, 4 h, and 6 h after exposure.

(ii) In vivo studies. (a) Evaluation of P. aeruginosa pathogenicity in mice.

Animals were randomly assigned to the following five groups with a minimum sample size of 10 animals per group for each series of experiments: PAO1 group, PAO1::lecA group, PAO1::lecB group, PAO1::lecA/pMMBlecA group, and PAO1::lecB/pMMBlecB group.

In the first series of experiments, mortality was assessed over a 7-day period following intratracheal instillation of 50 μl of a lethal inoculum (10⁹ CFU/ml) of P. aeruginosa.

Lung injury, pulmonary bacterial load, and blood cultures were studied at 6 and 16 h following intratracheal instillation of 50 μl of an inoculum calibrated to contain 10⁸ CFU/ml.

(b) Evaluation of the effect of P. aeruginosa lectins.

Two groups of animals were studied, the LecA group (n = 10) and the LecB group (n = 10). In these animals lung injury was evaluated at 6 h after intratracheal instillation of P. aeruginosa lectins.

(c) Evaluation of lectin inhibitors.

The following experimental groups were studied: PAO1 and GalNAc, PAO1 and Me-α-Gal, PAO1 and Me-α-Fuc, PAO1 and Glc, LecA and GalNAc, LecA and Glc, LecB and Me-α-Fuc, and LecB and Glc.

Statistical analysis.

The results are expressed below as means ± standard errors. Data were analyzed by the Kruskal-Wallis one-way analysis of variance test using Dunn's method to compare differences between groups. A P value of <0.05 was considered statistically significant. Survival was analyzed using a Kaplan-Meier algorithm (GraphPad Prism, v5.0), and cumulative survival rates were compared by using a log rank test. A level of 5% was considered statistically significant.

RESULTS

Characterization of lecA and lecB mutants.

The locations of the transposon insertion in the mutants designed were mapped previously using custom-designed primers, and the transposon was inserted at nucleotide positions 48 and 204 for the lecA and lecB genes, respectively. Alternatively, we designed external oligonucleotides lecAu (5′ CTCCTGCATGAATTGGTAGGC 3′) and lecAd (5′ GGGTCAGGAATCGATATTCCC 3′) and external oligonucleotides lecBu (5′ TAACAATCGAACGAGCCGGC 3′) and lecBd (5′ TCAACTGGACAGTCTGGGCG 3′) for the lecA and lecB genes, respectively, and PCR amplification of the corresponding genomic DNA region in the wild-type strain resulted in DNA fragments consisting of 736 and 836 nucleotides for lecA and lecB, respectively. Due to the PCR conditions and nature of the transposon, whose size exceeded 4.5 kbp, the absence of the corresponding bands for the lecA mutant (Fig. 1A, lane 2) and the lecB mutant (Fig. 1A, lane 5) and the presence of lecA (Fig. 1A, lane 1) and lecB (Fig. 1A, lane 4) in the parental isogenic strain, as well as in our reference PAO1 strain (Fig. 1A, lanes 3 and 6), confirmed that the transposon interrupted the lecA and lecB genes.

FIG. 1.

(A) PCR verification of lecA and lecB gene interruption using external lecA and lecB oligonucleotide pairs with which PCR amplification of the corresponding DNA region leads in the wild-type strain to 736- and 836-nucleotide DNA fragments for lecA and lecB, respectively. Lanes 1 and 4, parental PAO1 strain; lanes 2 and 5, PAO1lecA::Tc^r and PAO1lecB::Tc^r mutants, respectively; lanes 3 and 6, reference PAO1 strain. (B) Western blot analysis of LecA production in the parental strain P. aeruginosa PAO1 (lanes 1 to 3), PAO1lecA::Tc^r insertional mutants (lanes 4 to 5), and the PAO1::lecA/pMMBlecA strain (lane 6) with a polyclonal antibody directed against the LecA protein (lane 7). The numbers above the Western blot indicate the growth points at which the production was checked (1, 2, and 3 indicate aliquots obtained during the exponential, early stationary, and late stationary phases, respectively). (C) Growth curves for the parental strain P. aeruginosa PAO1 and insertional mutants PAO1lecA::Tc^r and PAO1lecB::Tc^r. nt, nucleotides.

To check whether lectins were produced in the different strains used in this study, production of LecA, for which we had a suitable polyclonal antibody that bound purified LecA protein, was assessed parallel to assessment of the bacterial growth curve. LecA production was detectable only in the late stationary phase in the parental wild-type strain (Fig. 1B, lanes 1 to 3). In the derived isogenic insertional lecA mutant, LecA production was undetectable in the late stationary phase (Fig. 1B, lanes 4 and 5), and it was fully restored to the parental level in the late stationary phase when the strain was transcomplemented with the lecA gene (Fig. 1B, lane 6). As a control, binding of the polyclonal antibody to purified LecA protein was checked (Fig. 1B, lane 7). No suitable antibody directed toward the LecB protein was available, and we were not able to assess LecB production in the different strains.

The growth curves obtained for the parental strain and the lecA and lecB mutants clearly showed that there was no major growth defect in the mutants compared to the parental strain (Fig. 1C).

In vitro cytotoxicity. (i) LecA and LecB lectins are determinants of P. aeruginosa cytotoxicity in A549 cells.

The cytotoxicities of the strains of P. aeruginosa with A549 cells were compared in vitro. Cytotoxicity was evaluated by measuring the release of LDH 4 and 6 h after infection. The amount of LDH released was statistically larger for PAO1-infected cells than for PAO1::lecA- and PAO1::lecB-infected cells at 4 and 6 h (Fig. 2A).

FIG. 2.

(A) Cytotoxicity of P. aeruginosa strains with lung epithelial cells. A549 cells (2 × 10⁴ cells) were cocultured with each strain at a multiplicity of infection of 250. (B and C) Effects of carbohydrates on the cytotoxicity of P. aeruginosa (PA) (B) and its lectins (C). A549 cells (2 × 10⁴ cells) were cocultured with P. aeruginosa PAO1 (parental strain) at a multiplicity of infection of 250 or with each lectin in the presence of Glc, GalNAc, Me-α-Fuc, or Me-α-Gal. The carbohydrates were added at a concentration of 15 mM. The cytotoxicity with lung epithelial cells was evaluated by measuring the release of LDH at 4 and 6 h. The values are the averages of three assays for the means (bars) and standard deviations (error bars). *, P < 0.05 for a comparison with PAO1; **, P < 0.01 for a comparison with PAO1; ***, P < 0.001 for a comparison with PAO1; +, P < 0.05 for a comparison with the corresponding lectin.

(ii) Carbohydrate lectin inhibitors decrease in vitro cytotoxicity for A549 cells.

The effects of the specific lectin-inhibiting carbohydrates (GalNAc and Me-α-Gal for LecA and Me-α-Fuc for LecB) on the cytotoxicity of P. aeruginosa and its lectins were compared to the effect of a nonspecific lectin inhibitor (Glc) used as a control.

Addition of specific inhibitors (GalNAc and Me-α-Gal for LecA and Me-α-Fuc for LecB) induced a significant reduction in PAO1-induced cytotoxicity that was observed only after 6 h of incubation compared to the results obtained with Glc (Fig. 2B). Similarly, the cell cytotoxicity obtained with the purified LecA and LecB lectins was significantly decreased when the specific inhibitors were added (Fig. 2C). Significant reductions were observed for the LecA inhibitors GalNAc and Me-α-Gal and for the LecB inhibitor Me-α-Fuc after 4 h and 6 h of incubation, respectively (Fig. 2C).

LecA and LecB mediate P. aeruginosa adherence to A549 cells.

The adherence of each strain of P. aeruginosa to A549 cells was assessed at 1, 4, and 6 h during incubation with A549 cells. The number of adherent bacteria was approximately twofold greater for PAO1-infected cells than for PAO1::lecA- and PAO1::lecB-infected cells (Fig. 3).

FIG. 3.

Relative attachment of P. aeruginosa strains to A549 cells: numbers of P. aeruginosa PAO1 (parental strain), PAO1::lecA (lecA mutant), and PAO1::lecB (lecB mutant) bacteria adhering to A549 cells. An inoculum containing 10⁸ CFU/ml was used. Bacteria were allowed to attach for 1, 4, and 6 h. The values are the means (bars) and standard deviations (error bars) of five independent experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (for a comparison with the PAO1 group).

Analysis of survival. (i)

LecA or Lec B mutation does not affect mortality. Survival studies were performed with the PAO1 wild-type strain and the insertional mutants PAO1::lecA and PAO1::lecB. Instillation of the PAO1 strain was associated with a mortality rate of 100% at 4 days, and most of the animals died within the first 48 h. Interestingly, although instillation of the PAO1::lecA mutant led to decreased 24-h mortality and a trend toward increased 4-day survival (12%) (Fig. 4A), the difference was not statistically significant. Instillation of the PAO1::lecB mutant also resulted in mortality rates similar to those obtained with the parental strain (Fig. 4B). The use of complemented strains resulted in mortality rates similar to those obtained with the parental strain.

FIG. 4.

Survival study. An inoculum containing 5 × 10⁷ CFU was instilled into the lungs, and then survival was monitored for 7 days for 20 mice per group. (A) Survival of mice after instillation of PAO1::lecA compared with the survival after instillation of PAO1 (parental strain) or PAO1::lecA/pMMBlecA (complemented lecA mutant). (B) Survival of mice after instillation of PAO1::lecB (lecB mutant) compared with the survival after instillation of PAO1 (parental strain) or PAO1::lecB/pMMBlecB (complemented lecB mutant).

(ii) Lectin-inhibiting carbohydrates improve survival.

Coinstillation of specific LecA inhibitors with P. aeruginosa decreased mortality compared to that observed for both the PAO1 group and the control group which received Glc (the irrelevant carbohydrate control) (Fig. 5). Furthermore, for the lectin-inhibiting carbohydrates, the rate of survival was statistically highest for mice which received a LecA inhibitor, either GalNAc or Me-α-Gal. No difference was observed between the group which received the LecB inhibitor Me-α-Fuc and the PAO1 group.

FIG. 5.

Effect of carbohydrates on mortality in mice. Mice were inoculated intratracheally with a suspension containing 5 × 10⁷ CFU of P. aeruginosa PAO1 (parental strain) in the presence and absence of Glc, GalNAc, Me-α-Gal, or Me-α-Fuc, and then survival was monitored for 7 days. The carbohydrates were used at a concentration of 15 mM (20 mice per group). *, P < 0.05 for a comparison with the PAO1 group; **, P < 0.001 for a comparison with the PAO1 group. PA, P. aeruginosa.

In vivo measurements. (i) LecA and LecB increase alveolar capillary barrier injury.

Lung injury was evaluated by measuring the efflux of the protein tracer ¹²⁵I-albumin either from the lungs into the blood at 6 h after intratracheal instillation or from the blood into the lungs 16 h after intraperitoneal injection. The efflux of the protein tracer was statistically greater for the PAO1 group than for the PAO1::lecA and PAO1::lecB groups at both 6 and 16 h (Fig. 6). The efflux of the protein tracer for the PAO1::lecA and PAO1::lecB groups was restored to the level of the PAO1 group through transcomplementation of the lecA gene (pMMBlecA) and of the lecB gene (pMMBlecB) in the corresponding mutants at 16 h (Fig. 6B). Likewise, purified lectins induced an increase in permeability at 6 h (Fig. 6C).

FIG. 6.

Alveolar capillary barrier permeability after intratracheal instillation of P. aeruginosa PAO1 (parental strain), PAO1::lecA (lecA mutant), PAO1::lecB (lecB mutant), and P. aeruginosa lectins (LecA and LecB). (A and C) Efflux of the protein tracer ¹²⁵I-albumin from the lungs into the blood 6 h after infection. (B) Efflux of the protein tracer ¹²⁵I-albumin from the blood into the lungs 16 h after intraperitoneal injection. The data are means (bars) and standards errors (error bars) for 10 mice per group. °, P < 0.05 for a comparison with the control; °°, P < 0.01 for a comparison with the control; °°°, P < 0.001 for a comparison with the control; *, P < 0.05 for a comparison with PAO1; **, P < 0.01 for a comparison with PAO1; ***, P < 0.001 for a comparison with PAO1; +, P < 0.05 for a comparison with the lecB mutant or lecA mutant; +++, P < 0.001 for a comparison with the lecB mutant or lecA mutant. CTR, control (no bacteria or no lectin).

(ii) LecA and LecB increase the amount of extravascular lung water.

The ratios of lung wet weight to lung dry weight were significantly greater for the PAO1 group than for both the PAO1::lecA and PAO1::lecB groups at 6 h (Table 2). At 16 h, the values for the extravascular lung water were similar to values obtained at 6 h, and they remained significantly higher for the PAO1 group than for the PAO1::lecA and PAO1::lecB groups.

TABLE 2.

Ratio of wet lung weight to dry lung weight^a

Strain	Ratio of wet lung weight to dry lung weight
	6 h after instillation	16 h after instillation
Control	3.61 ± 0.11	3.98 ± 0.04
PAO1	4.86 ± 0.04	4.79 ± 0.16
PAO1::lecA	4.04 ± 0.04 A	4.09 ± 0.07 B
PAO1::lecB	4.00 ± 0.28 A	4.08 ± 0.04 A

^aThe ratio of wet lung weight to dry lung weight was evaluated 6 h and 16 h after intratracheal instillation of P. aeruginosa PAO1 (parental strain), PAO1::lecA (lecA mutant), and PAO1::lecB (lecB mutant). The data are means ± standard errors (10 mice per group). A, P < 0.05 for a comparison with PAO1; B, P < 0.01 for a comparison with PAO1. The control contained no bacteria.

(iii) Coinstillation of lectin-inhibiting carbohydrates decreases lung injury.

To evaluate the effects of specific lectin inhibitors for preventing P. aeruginosa-induced lung injury, we coinstilled these molecules with P. aeruginosa. The effect of the specific lectin-inhibiting carbohydrates (GalNAc and Me-α-Gal for LecA and Me-α-Fuc for LecB) was compared to the effect of an irrelevant carbohydrate (Glc) used as control.

All of the specific lectin inhibitors decreased injury, as assessed by alveolar barrier permeability. Me-α-Gal was more potent than GalNAc and Me-α-Fuc at 6 and 16 h. At a concentration of 15 mM, the LecA inhibitor Me-α-Gal was the only inhibitor which significantly reduced the PAO1-induced lung permeability disorder (Fig. 7A). An improvement in permeability was also observed with GalNAc or Me-α-Fuc at a concentration of 50 mM but not at a concentration of 15 mM compared to the results obtained with Glc 6 h after infection (Fig. 7A and 7B). At 16 h, lung injury was significantly decreased with GalNAc or Me-α-Gal (Fig. 7B).

FIG. 7.

Effect of carbohydrates on lung injury. Lung injury was evaluated by examining alveolar capillary barrier permeability after intratracheal instillation of P. aeruginosa PAO1 (parental strain) and its lectins alone or in combination with Glc, GalNAc, Me-α-Gal, or Me-α-Fuc in mice. (A and C) Efflux of the protein tracer ¹²⁵I-albumin from the lungs into the blood 6 h after infection. (B) Efflux of the protein tracer ¹²⁵I-albumin from the blood into the lungs 16 h after intraperitoneal injection. The data are means (bars) and standards errors (error bars). The carbohydrates were used at concentrations of 15 and 50 mM, and there were 10 mice per group. *, P < 0.05 for a comparison with PAO1; **, P < 0.01 for a comparison with PAO1; ***, P < 0.001 for a comparison with PAO1. CTR, control (no bacteria or no lectin); PA, P. aeruginosa.

The combination of purified lectin LecA with GalNAc or purified LecB with Me-α-Fuc was associated with a significant reduction in lung injury at 6 h (Fig. 7C).

LecA and LecB mutation or inhibition increases lung bacterial clearance.

We assessed the ability of animals to clear bacteria from the lungs. Six hours after instillation of the pathogen, the bacterial loads were not significantly different for the PAO1, PAO1::lecA, and PAO1::lecB groups (Fig. 8A). However, after 16 h, the animals infected with PAO1 showed a significantly higher lung bacterial load than the animals in the PAO1::lecA or PAO1::lecB groups (Fig. 8A).

FIG. 8.

Lung bacterial clearance. Mice were infected with a suspension containing 5 × 10⁶ CFU of P. aeruginosa PAO1 (parental strain), PAO1::lecA (lecA mutant), or PAO1::lecB (lecB mutant). The number of viable bacteria remaining in the infected lungs was counted 6 h (H6) and 16 h (H16) after instillation (A). The effect of carbohydrates on lung bacterial clearance was evaluated after intratracheal instillation of 5 × 10⁷ CFU of P. aeruginosa PAO1 mixed with Glc, GalNAc, Me-α-Gal, or Me-α-Fuc at 6 h (B) and 16 h (C). Carbohydrates were used at concentrations of 15 and 50 mM, and there were 10 mice per group. The data are means (bars) and standard errors (error bars). *, P < 0.05 for a comparison with PAO1; **, P < 0.01 for a comparison with PAO1; ***, P < 0.001 for a comparison with PAO1. PA, P. aeruginosa.

Consistent with these results, coadministration of specific lectin inhibitors with P. aeruginosa increased lung bacterial clearance. At 6 h, the lung bacterial load was significantly decreased in the group of mice that received Me-α-Fuc (Fig. 8B), and the greatest reduction was observed at a concentration of 50 mM. Me-α-Gal at a concentration of 15 mM or GalNAc at a concentration of 50 mM was also associated with a significant decrease in the lung bacterial load 6 h after infection compared to the results obtained with the irrelevant carbohydrate Glc (Fig. 8B). At 16 h, the presence of all inhibitors (GalNAc and Me-α-Gal for LecA and Me-α-Fuc for LecB) led to a significant reduction in the lung bacterial load (Fig. 8C).

LecA and LecB mutation or inhibition decreases bacterial dissemination.

Blood culture analyses were performed at 6 and 16 h after instillation of P. aeruginosa. The number of positive blood cultures was statistically lower for mice instilled with strains PAO1::lecA and PAO1::lecB than for mice instilled with the parental PAO1 strain at 6 and 16 h (Table 3).

TABLE 3.

Pulmonary translocation^a

Strain	No. of positive blood cultures/total no. of cultures
	6 h after instillation	16 h after instillation
PAO1	9/10	10/10
PAO1::lecA	2/10 A	2/10 A
PAO1::lecB	3/10 B	2/10 A

^aThe results were determined 6 h and 16 h after intratracheal instillation of P. aeruginosa PAO1 (parental strain), PAO1::lecA (lecA mutant), and PAO1::lecB (lecB mutant) (10 mice per group). A, P < 0.01 for a comparison with PAO1; B, P < 0.001 for a comparison with PAO1.

Coinstillation of the PAO1 strain with the LecA inhibitor Me-α-Gal or with the LecB inhibitor Me-α-Fuc significantly decreased bacterial dissemination at 6 h compared to the results for the groups without any inhibitor or with the irrelevant carbohydrate (Glc). At 16 h, the bacterial dissemination remained significantly decreased (Table 4). No dissemination was observed with Me-α-Fuc and Me-α-Gal at that time point (Table 4). Addition of both inhibitors did not further improve the results.

TABLE 4.

Effect of carbohydrates on the bacterial dissemination of P. aeruginosa^a

Group	No. of positive blood cultures/total no. of cultures
	6 h after instillation	16 h after instillation
PAO1 control	9/10	10/10
PAO1 + Glc	8/10	9/10
PAO1 + GalNAc	9/10	1/10 A
PAO1 + Me-α-Fuc	0/10 B	0/10 B
PAO1 + Me-α-Gal	4/10 A	0/10 B
PAO1 + GalNAc + Me-α-Fuc	5/10	4/10 A
PAO1 + Me-α-Gal + Me-α-Fuc	5/10	3/10 A

^aThe results were determined 6 h and 16 h after intratracheal instillation of P. aeruginosa PAO1 (parental strain) alone or in combination with Glc, GalNAc, Me-α-Gal, or Me-α-Fuc in mice. The carbohydrates were used at a concentration of 15 mM (10 mice per group). A, P < 0.01 for a comparison with PAO1; B, P < 0.001 for a comparison with PAO1.

DISCUSSION

This study was designed to determine the contribution of the lectins LecA and LecB to P. aeruginosa pathogenicity. Using a lung epithelial cell line and an experimental murine model of lung injury, we compared three strains of P. aeruginosa: PAO1 (the parental strain which produces LecA and LecB) and isogenic mutants in which the lecA and lecB gene were inactivated. We then evaluated the roles of specific lectin inhibitors. Our results demonstrate that there is a relationship between the lectins and P. aeruginosa pathogenicity.

The participation of lectins, particularly LecA, in the pathogenicity of P. aeruginosa has been previously observed in vitro. Bajolet-Laudinat et al. showed that LecA had a cytotoxic effect on human epithelial cells in primary culture (2). The exposure of these cells to concentrations of LecA of >10 μg/ml inhibited their growth and decreased ciliary activity. With higher concentrations of LecA (100 μg/ml), these authors observed cellular lesions. Consistent with these findings, LecA was also shown to have a cytotoxic effect on the digestive epithelium (36). Using A549 cells, we demonstrated that the parental PAO1 strain has a greater cytotoxic effect than both mutant strains with the lectin genes inactivated. Even if P. aeruginosa-induced cytotoxicity is multifactorial, our observation that purified lectins have a cytotoxic effect further shows that P. aeruginosa cytotoxicity is at least in part a result of the lectins themselves rather than a complex result of cross-regulation or activation of genes due to the inactivation of LecA and LecB. Therefore, lectins may represent virulence factors which can lead to severe lung epithelial injury, contributing to lung epithelial damage.

We further assessed the in vivo relevance of LecA and LecB as virulence factors by measuring the mortality rate in a murine model of P. aeruginosa-induced lung injury. Survival was followed over a 7-day period. In our model we did not find a statistical difference between survival with the parental strain and survival with the two mutants with the lectin genes inactivated. In a different setting, the contribution of LecA to virulence was evaluated in an intestinal injury model. In this murine model, the combination of LecA and exotoxin A led to a high rate of mortality at 48 h after inoculation; however, no mortality was observed when these two virulence factors were inoculated separately (19). These results suggest that LecA acts in synergy with other virulence factors. The results of the second part of our study suggest that LecA and LecB are not major or direct determinants of mortality but could act as cofactors in virulence nonetheless.

The impact of P. aeruginosa lectins on alveolar capillary barrier permeability was analyzed further. This parameter is very sensitive and can quantitatively assess lung injury as well as the functional consequences of epithelial injury. For this assessment, the size of the inoculum was reduced to 10⁸ CFU/ml to avoid a mortality bias at 6 and 16 h postinfection. All the animals could therefore be studied. The leakage of the radioactive tracer decreased when mice were instilled with strains which did not produce either LecA or LecB, and this was associated with reduced extravascular lung water compared with that in the parental strain. Inoculation of the complemented strains led to permeability disorders similar to those observed with the parental strain at 16 h. However, we did not observe any difference between the strains with lectin genes inactivated and the lectin-complemented strains at 6 h; this was found to be related to a slower growth of the complemented strains (data not shown). We thus demonstrated that LecA and LecB are involved in P. aeruginosa-induced alveolar capillary barrier injury.

Oligosaccharide-mediated recognition and adhesion are key points in the early steps of P. aeruginosa pathogenesis, and lectins probably have a major role in adhesion (14, 15). Therefore, we also aimed at evaluating the role of lectins in clearance of P. aeruginosa from the lungs. The significant decreases in the lung bacterial loads in the groups instilled with the lecA and lecB mutant strains compared to that in the PAO1 group support our hypothesis that lectins act as virulence factors mainly in an early phase. Two recent studies have shown that the LecA and LecB lectins were involved in biofilm formation. Tielker et al. showed that LecB could bind to oligosaccharide patterns present on the cellular surface and that lecB-deficient mutants were impaired in biofilm formation (32). Similar results were obtained with LecA (6). Thus, these data suggest that modulation of the expression of LecA and LecB could potentially inhibit the initial adhesion of P. aeruginosa and consequently biofilm formation at a later phase of infection. Our data are consistent with data obtained previously; the increased bacterial clearance observed with the mutants can probably be related, at least partially, to decreased adhesion. To confirm this hypothesis, we compared the adhesion of the lectin mutant strains to that of the parental strain. We observed significant decreases in the adhesion of the lecA and lecB mutants, supporting the hypothesis that a deficiency in lectin-mediated adhesion allows increased clearance of the mutant bacterial strains.

We then focused on the bacterial dissemination into the bloodstream, which we found to be decreased with lecA and lecB mutant strains. Bacterial dissemination depends on several factors, among which are the severity and extent of alveolar capillary barrier damage, virulence factors (16), and bacterial load (27). At 6 h, the bacterial loads were equivalent for the three groups. The permeability of the alveolar capillary barrier was, however, significantly increased in the group instilled with the parental strain compared to the groups instilled with the two mutants. This result is consistent with the data of Kurahashi et al. (16), who found a correlation between the increase in permeability and bacteremia. Plotkowski et al. (25) hypothesized that besides causing permeability disorders, P. aeruginosa could cross the endothelial barrier and adhere to the endothelial cells to invade intravascular space through the involvement of glycanic structures on the surface of the endothelial cells. Such glycanic structures represent a potential target for lectins, as recently demonstrated ex vivo (14, 15). The decrease in P. aeruginosa dissemination could therefore be related not only to decreased permeability disorders but also to alterations in cellular glycan-lectin interactions.

P. aeruginosa produces two lectins with well-described carbohydrate-binding capacities (10). The glycoconjugates from epithelial and endothelial cell surfaces serve as binding sites for P. aeruginosa (15). LecA is a d-galactose-binding lectin with high affinity for terminal glycans presenting an α-galactose residue at the nonreducing end (4, 18). The LecB lectin has a high affinity for l-fucose and its derivatives (24, 35). Several studies have suggested that blockade or inhibition of these lectins by specific carbohydrates may be useful for prevention and treatment of P. aeruginosa infections (28, 31). In neonatal mice, aerosolized dextran significantly reduced P. aeruginosa pneumonia sequellae (3). Similarly, inhaled galactose and fucose successfully reduced P. aeruginosa airway infection (33).

In our study we investigated the effects of specific lectin-inhibiting carbohydrates on P. aeruginosa lung infection in vitro and in vivo. The addition of these inhibitors was associated in vitro with reduced cytotoxicity of P. aeruginosa or its purified lectins. In vivo, we observed an improvement in survival and reductions in lung injury, lung bacterial load, and bacterial dissemination.

Of the carbohydrates evaluated, only GalNAc and Me-α-Gal, which are specific LecA inhibitors, were associated with an improvement in survival. The beneficial effect of GalNAc on survival has previously been demonstrated in a model of P. aeruginosa induced-intestinal injury (36). The specific inhibitor of LecA, Me-α-Gal, appeared to be more effective than GalNAc, resulting in a significant improvement in lung injury. Similar results were obtained with Me-α-Fuc, a specific inhibitor of LecB. We also demonstrated that these two carbohydrates were most effective at a concentration of 50 mM. However, a combination of the two inhibitors did not further improve permeability.

In our study, we observed a significant decrease in the lung bacterial load and bacterial dissemination when specific lectin-inhibiting carbohydrates were coadministered with the wild-type strain. The mechanism probably involves inhibition of the carbohydrate-lectin interaction. Again, the effect was dose dependent with GalNAc and Me-α-Fuc. Coadministration of specific LecA inhibitors or LecB inhibitors at a concentration of 15 mM with the wild-type strain significantly decreased the lung bacterial load at 6 and 16 h after infection. Again, the combination of inhibitors was not additive or synergistic. These results support the hypothesis that LecA and LecB participate in the adhesion of P. aeruginosa to endothelial cells (25) and that there is a direct interaction between lectins and endothelial cells (14, 15).

Although the pathogenesis of P. aeruginosa is multifactorial, our results suggest that the lectins are key virulence factors that play a major role in P. aeruginosa-induced lung injury. Our results show that there is a significant correlation between the lectins and the severity of lung injury, lung bacterial load, and dissemination of the pathogen, influencing survival. Administration of specific lectin inhibitors was remarkably effective. The elucidation of the crystal structures of P. aeruginosa lectins complexed with carbohydrate ligands (5, 23), together with a synthetic chemistry effort to produce high-affinity carbohydrate-based ligands (21, 22), opens new possibilities for lectins as novel therapeutic targets in P. aeruginosa infections.