Macrophage migration inhibitory factor enhances Pseudomonas aeruginosa biofilm formation, potentially contributing to cystic fibrosis pathogenesis

Abstract

Macrophage migration inhibitory factor (MIF) is a key proinflammatory mediator that we have previously shown to be associated with an aggressive clinical phenotype in cystic fibrosis. It possesses unique tautomerase enzymatic activity. However, to date, no human-derived substrate has been identified that has the capacity to interact with this cytokine’s unique tautomerase activity. This led us to hypothesize that MIF may have the capacity to interact with external substrates. We describe for the first time how Pseudomonas aeruginosa can utilize human recombinant MIF (rMIF) to significantly (P < 0.01) enhance its endogenous biofilm formation. Our in vivo studies demonstrate that utilizing a small-molecular-weight inhibitor targeting MIF’s tautomerase activity (SCD-19) significantly reduces the inflammatory response in a murine pulmonary chronic P. aeruginosa model. In addition, we show that in in vitro experiments, pretreatment of P. aeruginosa with rMIF is associated with reduced bacterial killing by tobramycin. Our novel findings support the concept of an anti-MIF strategy that targets this enzymatic activity as a potential future antibacterial therapeutic approach.—Tynan, A., Mawhinney, L., Armstrong, M. E., O’Reilly, C., Kennedy, S., Caraher, E., Jülicher, K., O’Dwyer, D., Maher, L., Schaffer, K., Fabre, A., McKone, E. F., Leng, L., Bucala, R., Bernhagen, J., Cooke, G., Donnelly, S. C. Macrophage migration inhibitory factor enhances Pseudomonas aeruginosa biofilm formation, potentially contributing to cystic fibrosis pathogenesis.

Macrophage migration inhibitory factor (MIF) is a key proinflammatory cytokine that has been implicated in the pathogenesis of a variety of inflammatory diseases, including autoimmunity, asthma, cystic fibrosis (CF), and pneumonia lethality (1–5). Historically, it has also been recognized to play a key role in sepsis, and in particular gram-negative sepsis. In seminal articles, researchers initially showed that administration of specific MIF mAbs in murine sepsis models significantly attenuated mortality both when given before and after septic insult (6, 7). It was subsequently shown that this cytokine has the ability to augment TLR4 cellular expression, maximize inflammatory cell recruitment via enhanced chemotaxis, and override the glucocorticoid antiinflammatory activity (7, 8), all of which contribute either directly or indirectly to a heightened inflammatory response to injury.

In the context of Pseudomonas aeruginosa infection, the original description of MIF-knockout mice showed that these mice had less pulmonary inflammation after P. aeruginosa infection (9). We have previously described a functional MIF polymorphism associated with enhanced MIF protein secretion (4). Further studies showed that patients with genetically enhanced MIF production have a more aggressive clinical phenotype in CF. In this study, we also found a significant association with early Pseudomonas spp. colonization in these patients (2). More recently, inhibiting MIF activity has been shown to ameliorate murine models of ocular P. aeruginosa infections (10, 11). Thus, there is an increasing body of evidence implicating MIF as a key driver of the hosts’ inflammatory response after exposure to gram-negative infection, and in particular P. aeruginosa infection.

In this study, we sought to investigate whether human-derived MIF has the ability to directly affect the biologic functions of P. aeruginosa. In particular, we wished to address our hypothesis that P. aeruginosa has the ability to hijack specific MIF functions for its biologic advantage. Here we present work revealing that this gram-negative organism can utilize human MIF to enhance biofilm formation.

MATERIALS AND METHODS

Bacterial strains and culture conditions

In total 3 different strains of P. aeruginosa were used. We used 2 well-characterized strains: PA01, purchased from American Tissue Culture Collection–LGC Standards (Middlesex, United Kingdom), and PA14 (a kind donation from G. O’Toole, Dartmouth School of Medicine, Hanover, NH, USA). For the confocal imaging, a green fluorescent protein (GFP)-tagged PA01 strain (a kind donation from G. O’Toole) was used. The clinical isolate of P. aeruginosa was obtained from K. Schaeffer (St. Vincent’s University Hospital). Pseudomonas selective agar was (Oxoid, Basingstoke, United Kingdom) used to confirm bacterial strains as Pseudomonas spp. All strains were routinely maintained on Luria-Bertani (LB) agar or LB broth (Sigma-Aldrich, St. Louis, MO, USA) at 37°C unless otherwise stated. Liquid cultures were incubated overnight at 37°C on a rotary shaker at 225 rpm.

Recombinant proteins and inhibitors

For our in vitro assays, we used recombinant human MIF (rMIF), a kind donation from R. Bucala (Yale School of Medicine, Yale University, New Haven, CT, USA). We have previously published our work on the development of an in-house novel small-molecular-weight inhibitor for MIF’s enzymatic activity: SCD-19 [3-(2′-methylphenyl)-isocoumarin] (12). ISO-1 [(S,R)-3-(4-hydroxyphenyl)4.5-dihydro-5-isoxazole acetic acid methyl ester] is a commercially available inhibitor (VWR, Dublin, Ireland).

Biofilm formation

Biofilm formation assays were performed on the basis of the methodology from previously published work (13). In brief, starting from an overnight broth culture, a dilution containing ∼10⁸ colony-forming units (CFU)/ml was made. Sixteen wells of a round-bottomed polypropylene 96-well microplate (Corning, Corning, NY, USA) were inoculated with 100 μl of this dilution. After 4 h of adhesion, the nonadhered cells were removed from each well and plates rinsed with PBS solution. Next, 100 μl of fresh medium was added to the wells, and the plate was then incubated for a further 24 h. The supernatants were again removed and the wells rinsed with PBS. Once the wells were washed, 100 μl of PBS was added to each well, along with 20 μl of a commercially available resazurin solution (CellTiter-Blue; Promega, Madison, WI, USA). The plate was incubated at 37°C for 1 h, and the fluorescence (λ_ex 560 and λ_em 590 nm) was measured.

Pellicle formation assay

Pellicle formation assays were performed as previously described (14). Briefly, 10⁸ CFU/ml of PA01 was cultured in glass test tubes with or without 100 ng/ml of rMIF in 5 ml of LB broth held stationary at 37°C. Pellicles were allowed to form at the air–liquid interface of the standing culture for 72 h. After this, the media was carefully removed and the remaining pellicle stained with the addition of 1% crystal violet (CV) solution (Sigma-Aldrich) into each tube. After staining for 20 min, the tubes were washed gently with water until the unbound CV solution was removed from the tubes. Images of the test tubes were taken at this time for visual analysis. Finally, bound CV was released from the pellicle by rinsing with 33% acetic acid (Sigma-Aldrich), and the absorbance of the liquid was measured at 590 nm using a spectrophotometer.

Antibiotic treatment of preformed biofilms

After 24 h of biofilm formation, the wells were treated with 50 μg/ml of tobramycin in LB broth or LB broth alone, and incubated for a further 24 h at 37°C. After incubation, the wells were washed with PBS 3 times to remove any planktonic cells, and 150 μl of PBS was added to each well. The plate was then sonicated in a water bath for 10 min. The sonicated bacteria were then serially diluted and plated on LB agar for bacterial counts (CFU/ml).

Confocal analysis of biofilm structure

This protocol was undertaken with the help and equipment of E.C.’s laboratory at the Institute of Technology Tallaght. The confocal analysis of biofilm structures was performed using the biofilm flow cell system, which has been previously described in detail (15). Briefly, LB medium alone or conditioned with 100 ng/ml MIF was pumped to the flow cell using a peristaltic pump (Lennox Pump and Process; Watson Marlow, Dublin, Ireland) at a flow rate of 0.5 rpm. Overnight bacterial cultures of GFP PA01 were adjusted to an optical density of 0.1 at 600 nm by diluting with sterile LB medium. The flow cells were inoculated with 300 µl of culture by injection ∼1 cm before the flow cell. The flow cells were inverted to allow the bacteria to adhere to the glass slide for 1 h. The flow cells were returned upright and the clamp removed; the pump was started at a flow rate of 0.5 rpm and the flow cells incubated at 37°C until they were ready to be imaged. The flow cells were analyzed at 24 and 48 h using confocal laser scanning microscopy (CLSM) to take z-stack images of the biofilms. Imaris software (Bitplane, Belfast, Ireland) was used to obtain volume per unit area (cubic micrometers per square micrometer); a ratio between total volume and total area covered by biofilm was calculated. The compactness of the biofilm was assessed as total fluorescence per volume of biofilm (16).

Chemotaxis migration assay

Primary neutrophils were isolated from 40 ml of whole blood obtained from healthy individuals using Polymorph prep (Biosciences, Dublin, Ireland) as previously described (17) and directly used in migration assays. Neutrophils (1 × 10⁵) were added to the apical compartment of 24-well polycarbonate 5 μm porous membrane Transwell inserts (Corning). rMIF (100 ng/ml) was preincubated with either 100 μM of SCD-19 or ISO-1 for 30 min and then applied to the basal compartments of the system. The cells were incubated for 16 h at 37°C, and cells that migrated to the basal side of the transwell insert were counted.

Chronic airway infection mouse model

Specific-pathogen-free female C57BL/6 mice aged 6 to 8 wk (∼20 g) were purchased from Charles River Laboratories (Wilmington, MA, USA). They were maintained in the Biomedical Facility, University College Dublin, under laminar airflow conditions. All animals had free access to standard laboratory food and water. All housing was temperature controlled and had a 12-h light–dark cycle. This work was approved by University College Dublin animal research committee, was carried out under license from the Department of Health, Ireland, and complied with the international best practice for the care and use of laboratory animals.

Chronic pulmonary infection with P. aeruginosa was induced in C57BL/6 mice using a previously described methodology (5, 18). Briefly, female specific-pathogen-free mice aged 8 to 10 wk were anesthetized and inoculated intratracheally with P. aeruginosa incorporated into agar beads or sterile agar beads. They were allowed to recover from anesthesia, placed into a negative pressure isolator, and kept in standard conditions as above. ISO-1 and SCD-19, dissolved in 5% DMSO, or controls were administered intraperitoneally daily at a concentration of 35 mg/kg throughout the course of the experiment starting on d 0.

Animals were monitored for the duration of the experiment. Total weight loss, differential cell counts, bacteria load, and histologic analysis were performed as previously described (19–23).

Statistical analysis

The Student’s t test (parametric data) or Mann-Whitney test (nonparametric data) were used using GraphPad InStat 3.00 (GraphPad Software, La Jolla, CA, USA). One-way analysis of variance with the Tukey-Kramer multiple comparisons post hoc test (parametric data) or the Kruskal-Wallis test with the Dunn multiple comparisons posttest (nonparametric data) were used to test for statistical significance of differences between more than 2 experimental groups. Statistical significance was recorded at P < 0.05.

RESULTS

Investigation of biofilm formation by PA01 and -14 lab strains, and clinical isolate CI58112 by rMIF using resazurin assay

The reference laboratory P. aeruginosa strains PA01 and PA14 were cultured for 24 h in the presence or absence of 100 ng/ml recombinant MIF (rMIF). Biofilm growth was assessed at 24 h using the resazurin biofilm assay. Here we demonstrate that when PA01 is grown in the presence of 100 ng/ml rMIF for 24 h, there is a significant increase in biofilm formation (195.60 ± 8.83 fluorescent units (FU), P < 0.001) compared to bacteria grown in LB medium alone (140.45 ± 4.18 FU) (Fig. 1A). In addition, the reference lab strain PA14 also demonstrated a significant increase in biofilm formation when grown in the presence of 100 ng/ml of rMIF for 24 h (82.769 ± 6.74 FU, P < 0.05) compared to bacteria grown in LB medium alone (58.82 ± 2.415 FU) (Fig. 1B). A clinical isolate CI58112 demonstrated a significant increase in biofilm formation when grown in 100 ng/ml rMIF for 24 h (125.5 ± 6.66 FU, P < 0.001) compared to biofilms grown in LB medium only (90.82 ± 7.29 FU) (Fig. 1C).

Figure 1.

Enhanced biofilm formation by rMIF. A) Biofilms of P.aeruginosa lab strain PA01 grown for 24 h, with or without 100 ng/ml rMIF, were measured by resazurin assay. Significant increase in biofilm formation was seen in biofilms formed in presence of rMIF (100 ng/ml). Data are presented from a series of n = 5 experiments. Wilcoxon matched pairs test was used to analyze data. ***P < 0.001. B) Biofilms of P. aeruginosa lab strain PA14 grown for 24 h, with or without 100 ng/ml rMIF, were measured by resazurin assay. Significant increase in biofilm formation was seen in PA14 biofilms formed in presence of rMIF (100 ng/ml). Data are presented as means ± sem fluorescence signals (FU) (n = 8). A paired Student’s t test was used to analyze data. *P < 0.05. C) Biofilms of clinical isolate CI58112 were grown for 24 h, with or without 100 ng/ml rMIF, and were measured by resazurin assay. Significant increase in biofilm formation was seen when biofilms were formed in presence of rMIF (100 ng/ml). Data are presented as means ± sem fluorescence signals (FU) (n = 8). A paired Student’s t test was used to analyze data. ***P < 0.001 (D, E). Air–liquid interface biofilms or pellicles were allowed to form with or without rMIF (100 ng/ml), stained with CV, washed thoroughly, and imaged. D) Representative images of CV-stained pellicles with or without 100 ng/ml rMIF. Bound CV was released and absorbance measured at 590 nm. E) Significant increase in biofilms grown with rMIF (100 ng/ml) was observed. Data are presented as means ± sem absorbance units (590 nm) (n = 8). Wilcoxon matched pairs test was used to analyze data. **P < 0.01.

Investigation of effect of rMIF treatment on PA01 pellicle formation using CV assay

After 72 h of culture in the presence or absence of rMIF (100 ng/ml), the pellicle was stained with CV. Images were taken of the stained test tubes before bound CV was released (Fig. 1D). This illustrates the formation of a larger pellicle when PA01 was grown in the presence of 100 ng/ml rMIF. Subsequently, the pellicle was washed to release the bound CV, and the absorbance was measured. This demonstrated that there was a significant increase in the pellicle formation of PA01 in the presence of rMIF (100 ng/ml) (0.42 ± 0.0097, absorbance units, P < 0.01) compared to the control growth (0.32 ± 0.0043, absorbance units) (Fig. 1E).

Characterization of effects of rMIF on P. aeruginosa biofilm formation using time-lapse CLSM after 48-h culture

In these studies, a GFP-tagged PA01 strain was used for experiments and biofilms were permitted to grow for 48 h. Images of biofilms were recorded and analyzed after 24 and 48 h of culture. After 24 h of culture, representative images captured by CLSM demonstrated that P. aeruginosa formed a biofilm consisting of a flat biofilm interspersed with classic mushroom-shaped biofilm structures (Fig. 2). A significant increase in mushroom-shaped structures was observed in representative images when bacteria were grown for 24 h in the presence of rMIF (Fig. 2B) compared to bacteria grown in LB control medium only (Fig. 2A). The images captured by CLSM were then subjected to quantitative analysis by Imaris software. This software applies an algorithm to each image that uses biofilm volume, compactness, thickness, and appearance as parameters to facilitate optimum quantification. The quantitative analysis demonstrated that there was a significant increase in the volume of biofilms formed after 24 h in the presence of 100 ng/ml MIF (130,553 ± 44,482 µm³, P < 0.001) compared to control (13,367 ± 9334 µm³) after 24 h of growth (Fig. 2C). After 48 h, quantitative analysis was performed on images, and it was shown that there was a significant increase in the volume of biofilms formed after 48 h in the presence of 100 ng/ml MIF (743,239 ± 191,892 µm³, P < 0.01) compared to control (103,512 ± 34,326 µm³) (Fig. 2D).

Figure 2.

Confocal microscopy of P. aeruginosa biofilms. GFP-tagged PA01 biofilms at 24 and 48 h in absence and presence of rMIF. Using biofilm flow cell system, GFP-tagged PA01 was allowed to form biofilms for 24 and 48 h at 37°C in absence or presence of 100 ng/ml rMIF. biofilms were then imaged using confocal microscope to collect z-stack compositions. Images were analyzed by Imaris software(Bitplane, Belfast, Ireland). A) Representative image of biofilms grown for 24 h in LB medium only. Scale bar, 50 μm. B) Representative image of biofilms grown in presence of 100 ng/ml rMIF for 24 h. Scale bars, 50 μm. C) Significant increase in volumes of biofilms grown for 24 h, in presence of 100 ng/ml rMIF, when calculated by Imaris software. Data are presented as means ± sem cubic micrometers (n = 10). A Mann-Whitney U test was used to analyze data. ***P < 0.001. D) At 48 h, significant increase in volume of PA01 biofilms grown in presence of 100 ng/ml rMIF was demonstrated, compared to LB alone, when calculated by Imaris software. Data are presented as means ± sem cubic micrometers (n = 10). A Mann-Whitney test was used to analyze data. ***P < 0.001.

Investigation of effect of rMIF-induced biofilm formation on P. aeruginosa survival after antibiotic treatment

Bacterial counts from PA01 cultures pretreated with rMIF for 24 h and then cultured for a further 24 h in the presence of 50 μg/ml tobramycin (6.8 × 10⁶ ± 3.5 × 10⁵ CFU/ml, P < 0.01) were significantly higher than bacterial counts from cultures that were grown in LB broth only before 24 h culture in 50 μg/ml tobramycin (2.92 × 10⁶ ± 1.1 × 10⁵ CFU/ml) (Fig. 3A). We found that that pretreatment of the reference lab strain PA14 with 100 ng/ml rMIF for 24 h before administration of 50 μg/ml tobramycin for an additional 24 h significantly increased survival of bacteria (rMIF: 2.72 × 10⁷ ± 1.1 × 10⁶ CFU/ml, P < 0.05) compared to bacteria grown in LB medium only before antibiotic treatment (Ctrl: 8.67 × 10⁶ ± 1.8 × 10⁶ CFU/ml) (Fig. 3B).

Figure 3.

rMIF increases survival of (A) PA01 and (B) PA14 after treatment with tobramycin. Biofilms were grown for 24 h, with or without 100 ng/ml MIF, and then treated with 50 µg/ml of tobramycin for a further 24 h. Biofilms were sonicated, serially diluted, and plated on agar to analyze CFU per milliliter as measure of bacterial survival. A) PA01 demonstrates significant increase in bacterial survival when it is grown in presence of MIF 100 ng/ml before treatment with 50 µg/ml of tobramycin. Data are presented as means ± sem CFU per milliliter (N = 10). B) Biofilms of PA14 have significantly increased survival against tobramycin when pretreated with 100 ng/ml of rMIF compared to control grown in LB alone. Data are presented as means ± sem CFU per milliliter (n = 4). *P < 0.05, **P < 0.01; Kruskal-Wallis with Dunn post hoc test.

Assessment of effects of SCD-19 tautomerase inhibitor on human polymorph nuclear neutrophil migration using chemotaxis assay

We investigated the ability of SCD-19 to decrease the migration of neutrophils in response to MIF (100 ng/ml) treatment using human polymorphonuclear leukocytes from healthy volunteers. We demonstrated that in the presence of rMIF, there is a significant increase in the migration of neutrophils (1,927,666 ± 94,686, P < 0.001) compared to the control medium (487,875 ± 99,573) (Fig. 4). This increase is significantly reduced when the cells are incubated with either SCD-19 (653,167 ± 37,730, P < 0.001) or ISO-1 (889,833 ± 20,626, P < 0.05).

Figure 4.

SCD-19 inhibits migration of human neutrophils. Cell counts of lower compartment of Transwell plate revealed significant decrease in human neutrophil migration in response to 100 ng/ml rMIF for 16 h with either 100 μM SCD-19 or ISO-1. Data are presented as mean ± sem cells per milliliter, n = 6, representative of at least 2 experiments. **P < 0.01, ***P < 0.001.

Percentage weight change of mice on d 7 of murine model of P. aeruginosa pulmonary infection

The percentage weight loss of each of the mice on d 7 of the study is presented in Fig. 5A. At the d 7 time point, the uninfected control group gained weight (3.96 ± 1.12%) compared to the P. aeruginosa–infected group, which lost a significant amount of weight (−1.77 ± 1.01%, P < 0.01). The P. aeruginosa + DMSO group lost the most weight on d 7 of the study (−2.17 ± 0.92%, P < 0.01), and the group treated with P. aeruginosa + ISO-1 had a change in body weight of (−1.48 ± 1.7%, P < 0.05), with both groups experiencing significantly increased weight loss compared to the uninfected control group. Weight loss in the group treated with P. aeruginosa + SCD-19 (0.17 ± 0.6%) was not significantly different from the uninfected control group.

Figure 5.

Significant decrease in percentage weight change of mice on d 7 and presence of neutrophils after infection and treatment with SCD-19 in murine model of PA pulmonary infection. Animals were injected 30 min before intratracheal instillation and daily thereafter with 35 mg/kg SCD-19 (solubilized in 5% DMSO) or as indicated. Body weight was measured to monitor extent of infection. A) Day 7 weights are expressed as percentage of start weight. PA-infected, PA + DMSO, and PA + ISO-1 groups all lost significant amount of weight compared to uninfected control group; however, SCD-19 group had not lost significant amount of weight at end of study. Data are presented as means ± sem of percentage of body weight (n = 6 per group). A 1-way ANOVA with post hoc Tukey-Kramer test was used to test for statistical differences. *P < 0.05, **P < 0.01. B) Differential cell counts were performed on BAL fluid from right lung of mice. Significant decrease in percentage of neutrophils in mice treated with SCD-19 was found in comparison to PA-infected mice. Neutrophil numbers were assessed by differential cell count using Diff-Quick stain and are expressed as means ± sem of percentage total cell count. A Kruskal-Wallis test with Dunn’s multiple comparison post hoc test was used to test for statistical differences (n = 6 per group). ***P < 0.01.

Detection of P. aeruginosa in lung homogenates at d 7 in murine model of pulmonary infection

All mice were humanely killed on d 7 of the experiment. The left lung was subsequently excised and homogenized. To detect P. aeruginosa in the lung homogenates, genomic DNA was extracted from the samples and used in a TaqMan assay for P. aeruginosa detection. The results of the assay are shown in Table 1. P. aeruginosa was not detected in any of the uninfected control group samples as expected. Five of the 6 mice in the P. aeruginosa–infected group contained detectable levels of P. aeruginosa in their lungs at d 7. No samples from either the uninfected control group or the SCD-19 MIF enzymatic inhibitor treatment group contained detectable P. aeruginosa levels in their lungs.

TABLE 1.

Decreased detection of P. aeruginosa in lung homogenates of mice

Group	Pseudomonas detected [n (%)]
Uninfected control	0/5 (0)
P. aeruginosa infected	5/6 (83)
P. aeruginosa + DMSO	4/6 (67)
P. aeruginosa + ISO-1	1/6 (17)
P. aeruginosa + SCD-19	0/6 (0)

Decreased detection of P. aeruginosa in lung homogenates of infected mice treated with SCD-19. DNA was extracted from left lung homogenate (n = 6 per group) and analyzed as per PA01 TaqMan PCR kit instructions. Results are presented as number or mice testing positive for PA01. P. aeruginosa + SCD-19–treated group showed complete absence of detectable levels of PA01 after 7 d of treatment.

Differential cell counts in bronchoalveolar lavage fluid at d 7 in murine model of P. aeruginosa pulmonary infection

The right lung was cannulated and bronchoalveolar lavage (BAL) performed. BAL fluid was then centrifuged onto glass slides and stained using Diff-Quick. A total of 300 cells per slide were counted to determine the differential cell count. The percentage of neutrophils counted in each sample is presented in Fig. 5B. The normal range for neutrophils in a murine BAL sample is between 5 and 10% of the total cell count. Neutrophil counts in the uninfected control group is within this range (6.5 ± 2.1%), but both the P. aeruginosa–infected group (22.86 ± 1.3%, P < 0.001) and the P. aeruginosa + DMSO group (20.3 ± 1.26%, P < 0.01) were significantly elevated in comparison. The P. aeruginosa + ISO-1 group showed a small decrease in neutrophils (16.6 ± 2.58%) compared to the P. aeruginosa + DMSO group, but this was not statistically significant. The level of neutrophils in the SCD-19 group was significantly decreased (9.3 ± 1.22%, P < 0.01) compared to the DMSO group, which were also restored to the normal range.

Histologic analysis of lungs at d 7 in murine model of P. aeruginosa pulmonary infection

At d 7, mice were humanely killed; the lungs of 3 mice per experimental group were perfused and fixed in 4% paraformaldehyde for histologic analysis. The tissue was subsequently embedded in paraffin and sliced into 5-µm-thick sections, which were then stained with hematoxylin and eosin (H&E) to investigate inflammatory cell infiltration. H&E histology was completed using a minimum of 6 sections from 3 murine lungs. Histologic analysis of lung tissue revealed extensive inflammation, thickening of the alveolar walls, and inflammatory cell infiltration in the P. aeruginosa–infected lung (Fig. 6B) and the P. aeruginosa + DMSO lung (Fig. 6C). In contrast, the mice treated with P. aeruginosa + SCD-19 (Fig. 6E) had normal honeycomb alveolar structure compared to the uninfected control mice (Fig. 6A) and demonstrated minimal inflammatory cell infiltration. The P. aeruginosa + ISO-1 (Fig. 6D) mice demonstrated an intermediate phenotype with some alveolar wall thickening and a minimal amount of inflammatory cell infiltration.

Figure 6.

Staining of lung sections. H&E staining was performed on murine lung sections at d 7 after infection from the following groups: uninfected control (A), PA infected (B), PA + DMSO (C), PA + ISO-1, (D) and PA + SCD-19 (E). PA-infected mice treated with SCD-19 tautomerase inhibitor (E) displayed reduced level of inflammation compared to DMSO control mice (C). Sections from ISO-1-treated mice (D) also demonstrated decrease in inflammation compared to DMSO-treated mice (C). Images are representative of minimum of 6 sections from 3 murine lungs. Scale bars, 100 µm. PA, P.s aeruginosa.

DISCUSSION

P. aeruginosa is a common pathogen found globally in watery environments. Historically, it has shown a highly effective ability to adapt to changing hostile environments, ultimately for its enhanced survival. It is a major cause of hospital-acquired infections, particularly in the lungs, and chronic infection with mucoid strains in CF is associated with accelerated decline in lung function and increased morbidity and mortality (24). One of the most characterized mechanisms by which this pathogen evades host defenses is via enhanced biofilm formation (25).

We here present evidence for the ability of this bacteria to hijack a human cytokine, namely MIF, for its biologic advantage. MIF possesses unique tautomerase enzymatic activity, and to date, no human-derived substrate has been identified that has the capacity to interact with this cytokine’s unique tautomerase activity. We thus decided to investigate whether MIF, the tautomerase enzyme, may have the capacity to interact with external bacterial substrates. This led to our novel finding that MIF could enhance biofilm formation of P. aeruginosa by means of this enzymatic activity and that inhibitors of this enzymatic activity may potentially act as an effective adjunct therapy to antibiotics to treat human P. aeruginosa infections in the future.

A unique functional characteristic of MIF is that it possesses tautomerase enzymatic activity and was originally described as having the ability to tautomerase the nonphysiologic substrates d-dopachrome and l-dopachrome methyl ester into their indole derivatives (26). ISO-1 represents the most characterized commercially available inhibitor of MIF’s tautomerase activity (27). In the context of gram-negative sepsis, ISO-1 has been shown to attenuate mouse mortality in a classic LPS murine sepsis model (28). To date, no definitive physiologic substrate for MIF’s enzymatic activity has been identified in humans. This led us to hypothesize that human-derived MIF may interact with external bacterial derived substrates. In support of this, it is well recognized that enzymatic MIF has structural homology with bacterial tautomerase enzymes, including 4-oxalocrotonate tautomerase and 5-carboxymethyl-2-hydroxymuconate isomerase (29, 30). The original seminal work describing the phenotype of MIF-knockout mice showed that mice lacking global MIF activity had less pulmonary inflammation in a murine model of pulmonary infection with P. aeruginosa (9). In our original work showing that the MIF CATT polymorphism was associated with more aggressive disease in CF, we also found a significant association with earlier pulmonary P. aeruginosa colonization in these patients (2). More recently we published work that utilized a tautomerase-null MIF gene knock-in mouse in a chronic pulmonary murine Pseudomonas infection model and found that these mice had significantly less infection, bacteria load, and pulmonary injury (5). Investigators have previously shown that blocking MIF activity in a murine ocular Pseudomonas infection model significantly reduced bacterial burden and ocular inflammation (11). Further work has demonstrated that mice deficient for CD74, the putative MIF receptor, developed milder P. aeruginosa–induced disease. Additionally, topical inhibition of MIF in this model further promoted recovery from disease (10). This suggests an additional CD74-independant mechanism for the role of MIF in Pseudomonas infection. Thus, there exists supporting evidence linking MIF, its enzymatic activity, and infection with P. aeruginosa.

This led us to hypothesize that P. aeruginosa has the ability to utilize host-derived MIF, and specifically its tautomerase enzymatic activity, for survival advantage. Our results showed that P. aeruginosa biofilm formation could be enhanced with the addition of rMIF and that this enhancement could be demonstrated in multiple ways. We noted that the rMIF-enhanced biofilm formation in our studies did not occur in a concentration-dependent manner but was consistently observed when rMIF concentrations >50 ng/ml were used (data not shown). Concentrations of MIF in the lung vary greatly, with previous studies having shown that in plasma and BAL fluid from septic lungs, concentrations of between 60 and 100 ng/ml have been detected (31, 32). The finding that rMIF can enhance biofilm formation raises the possibility that P. aeruginosa has the ability to sense the concentration of MIF as a marker of the local inflammatory environment and can accelerate endogenous biofilm formation when the local inflammatory defenses reach a certain threshold. Although this is an intriguing possibility, much additional work will be required to validate this proposition.

To investigate whether this novel observation of enhanced biofilm formation confers a survival advantage to the bacteria, we pretreated P. aeruginosa with rMIF and assessed bacterial killing after exposure to tobramycin. We found pretreated bacteria to have enhanced survival after antibiotic exposure in our in vitro system. From a clinical perspective, this raises the tantalizing possibility that targeting MIF’s tautomerase enzymatic activity represents a valid therapeutic strategy against this gram-negative organism.

We have recently developed a specific in-house set of inhibitors for this enzymatic activity, and we have shown that our lead compound, SCD-19, has significant anticancer activity (20). To address the role of this unique enzymatic activity in P. aeruginosa function, and specifically MIF-induced biofilm formation, we used an animal model of chronic P. aeruginosa infection in which we used our own lead inhibitor of MIF enzymatic activity and a known inhibitor, ISO-1. We found that SCD-19 could significantly inhibit MIF-induced primary human neutrophil chemotaxis both in vitro and in vivo. This is important because neutrophils are a source of MIF and also have been shown to contribute to biofilm formation in the lung (33). We also found that SCD-19 treatment in our murine pulmonary P. aeruginosa infection model reduces pulmonary inflammation and bacterial burden in treated mice. These data support the idea of the importance of this tautomerase activity in accelerating MIF-induced biofilm formation.

Enhanced MIF activity has been shown to be a key driver of an injurious inflamma4tory response, and previously published work, including our own, has implicated this cytokine in the pathogenesis of a variety of chronic inflammatory diseases (4, 10, 11). In the context of gram-negative sepsis, it has been shown to contribute to a sustained inflammatory response by directly promoting proinflammatory cytokine release in infected organs and indirectly via its ability to first augment cellular TLR4 expression, a key pattern recognition receptor for LPS (8); second to down-regulate p53, leading to a longer life span for these activated inflammatory cells; and third to potentially override the body’s own antiinflammatory glucocorticoid activity (31). This thus leads to a sustained antibacterial inflammatory response. In response, P. aeruginosa hijacks a human proinflammatory cytokine, which results in enhanced survival of this gram-negative organism and which confers a significant biologic advantage to these bacteria, compared to other organisms, in this local inflammatory firestorm.

MIF has been shown to promote leukocyte recruitment during inflammation, which can be partly explained by its chemokine properties (32). In in vivo arthritis models, direct injection of MIF induces a significant neutrophil influx into affected joints (34). Conversely, these models, on a MIF-knockout murine background, show a significant reduction of neutrophils within inflamed joints (35). More recently, a series of elegant experiments definitively showed MIF to be a potent neutrophil chemokine in murine inflammatory arthritis (36). Here we present data showing that SCD-19 significantly inhibits neutrophil chemotaxis both in in vitro systems utilizing primary human neutrophils and in our in vivo murine model of pulmonary P. aeruginosa infection. The search for novel methods to deal with antibiotic resistance represents a global imperative and a significant challenge. This work supports the targeting of MIF’s tautomerase activity as a valid adjunctive therapy in combating bacterial infections.

Glossary

BAL

bronchoalveolar lavage

CF

cystic fibrosis

CFU

colony-forming unit

CLSM

confocal laser scanning microscopy

CV

crystal violet

FU

fluorescent units

GFP

green fluorescent protein

H&E

hematoxylin and eosin

ISO-1

(S,R)-3-(4-hydroxyphenyl)4.5-dihydro-5-isoxazole acetic acid methyl ester

LB

Luria-Bertani

MIF

macrophage migration inhibitory factor

rMIF

recombinant human macrophage migration inhibitory factor

SCD-19

3-(2′-methylphenyl)-isocoumarin

AUTHOR CONTRIBUTIONS

A. Tynan, M. E. Armstrong, G. Cooke, and S. C. Donnelly designed research; A. Tynan, L. Mawhinney, M. E. Armstrong, C. O’Reilly, K. Jülicher, L. Maher, and G. Cooke performed research; S. Kennedy, E. Caraher, K. Schaffer, A. Fabre, E. F. McKone, L. Leng, R. Bucala, and J. Bernhagen contributed new reagents or analytical tools; A. Tynan, L. Mawhinney, C. O’Reilly, K. Jülicher, D. O’Dwyer, A. Fabre, G. Cooke, and S. C. Donnelly analyzed data; and A. Tynan, G. Cooke, and S. C. Donnelly contributed to writing the article.