A Drug-Repositioning Screening Identifies Pentetic Acid as a Potential Therapeutic Agent for Suppressing the Elastase-Mediated Virulence of Pseudomonas aeruginosa

Abstract

Pseudomonas aeruginosa, a Gram-negative bacterium of clinical significance, produces elastase as a predominant exoprotease. Here, we screened a library of chemical compounds currently used for human medication and identified diethylene triamine penta-acetic acid (DTPA, pentetic acid) as an agent that suppresses the production of elastase. Elastase activity found in the prototype P. aeruginosa strain PAO1 was significantly decreased when grown with a concentration as low as 20 μM DTPA. Supplementation with Zn²⁺ or Mn²⁺ ions restored the suppressive effect of DTPA, suggesting that the DTPA-mediated decrease in elastase activity is associated with ion-chelating activity. In DTPA-treated PAO1 cells, transcription of the elastase-encoding lasB gene and levels of the Pseudomonas quinolone signal (PQS), a molecule that mediates P. aeruginosa quorum sensing (QS), were significantly downregulated, reflecting the potential involvement of the PQS QS system in DTPA-mediated elastase suppression. Biofilm formation was also decreased by DTPA treatment. When A549 alveolar type II-like adenocarcinoma cells were infected with PAO1 cells in the presence of DTPA, A549 cell viability was substantially increased. Furthermore, the intranasal delivery of DTPA to PAO1-infected mice alleviated the pathogenic effects of PAO1 cells in the animals. Together, our results revealed a novel function for a known molecule that may help treat P. aeruginosa airway infection.

INTRODUCTION

Persistent use of antibiotics has resulted in the unwanted emergence of drug-resistant bacterial strains, ushering in a new era in which alternative strategies are necessary for the treatment of bacterial infections. Attenuating bacterial virulence, rather than treating infections with bactericidal antibiotics, might be a more effective approach, as antibiotic treatment imposes a selection pressure that often results in the generation of escape mutants (1).

Pseudomonas aeruginosa, a highly adaptable bacterium that colonizes various environmental niches, is responsible for hospital-acquired infections (2) and airway infections in patients with cystic fibrosis (3). Importantly, the incidence of multidrug-resistant P. aeruginosa strains is increasing at a rate faster than the production of new antibiotics, rendering treatment of P. aeruginosa infections difficult (4).

Elastase, encoded by the lasB gene, is an important virulence determinant of P. aeruginosa. Expression of the lasB gene is controlled by interrelated P. aeruginosa quorum-sensing (QS) circuitry. The LasI/R (5), RhlI/R (6), and Pseudomonas quinolone signal (PQS) (7) QS systems all contribute to the uninterrupted production of elastase (8, 9). Elastase is the most abundant protein secreted into culture medium (8), and a wide range of protein substrates, which include elastin (10), collagen (11, 12), IgA (13), and IgG (14–16), are degraded by elastase. Therefore, compounds that inhibit elastase activity may have a potential to be developed as anti-infective agents to undermine P. aeruginosa virulence. However, identification of such inhibitors may not always result in clinical application due to concerns associated with safety for use in the human body.

In this study, we screened a library of FDA-approved drugs in search of compounds that inhibit elastase activity. The library comprises chemicals that are currently in use for a variety of treatments with known safety dosages and pharmacokinetic profiles. We found that diethylene triamine penta-acetic acid (DTPA, pentetic acid), which is clinically used as a contrast agent for diagnostic imaging, suppresses elastase production. Thus, we attempted to outline the mechanisms of DTPA-induced repression of elastase production and examined whether DTPA treatment was effective in reducing biofilm formation and in vivo virulence of P. aeruginosa. This report demonstrates how drug repositioning can be used to identify compounds of verified safety for application in antagonizing the virulence of a clinically important bacterial pathogen.

MATERIALS AND METHODS

Ethics statement.

Animal experiments were performed following the institutional guidelines for animal care and use of laboratory animals. The Yonsei University College of Medicine Committee on the Ethics of Animal Experiments approved the methods for animal experimentation with B6 mice (permit number 2011-0173-2).

Bacterial growth and chemical reagents.

Bacterial cultures were grown at 37°C in Luria-Bertani (LB) medium (10 g tryptone, 5 g yeast extract, and 10 g NaCl per liter). Chemical reagents, including Ca-DTPA (catalog no. 32325) and Gd-DTPA (catalog no. 381667), were purchased from Sigma-Aldrich (St. Louis, MO).

Drug screening for a P. aeruginosa elastase inhibitor.

The compound library used in this study was “The Spectrum Collection” purchased from MicroSource Discovery Inc. (Gaylordsville, CT) (17). The library, which is composed of 2,320 compounds in a 10 mM dimethyl sulfoxide (DMSO) solution, lists primarily FDA-approved human therapeutic drugs, drug-like compounds, and natural products. Each compound was diluted in DMSO to reach a concentration of 2.5 mM, which was used as the 100× compound stock. The PAO1 strain, which was determined to produce and secrete a large amount of elastase (8), was used for the experiments. The bacteria were grown in 1.0 ml LB broth with each compound (25 μM). The cultures were then centrifuged for 5 min, and culture supernatants were recovered for subsequent elastin-Congo red assays.

Elastin-Congo red assay and elastase Western blot analysis.

Elastin-Congo red assays were performed as previously described (18). For Western blot analyses, bacterial culture supernatants (20 μl) and cell extract fractions (20 μg) were loaded onto 12% SDS-PAGE gels. Subsequent Western blot analysis was performed as described elsewhere (8). Anti-elastase antibody was obtained as a gift from Efrat Kessler of Tel Aviv University in Israel.

qRT-PCR analysis, PQS assay, CFU measurement, antibiotic sensitivity, and biofilm assays.

Reverse transcription-quantitative PCR (qRT-PCR) analysis was performed as previously described (8, 19). Transcript levels of the rpoD gene were similar in cells grown in plain LB medium or in LB medium supplemented with DTPA or EDTA, which was used for normalization. Primers used in the qRT-PCR assay are listed in Table 1. The PQS assay was performed following procedures described elsewhere (20). CFU of PAO1 cells grown under various culture conditions were determined by enumerating the colony numbers of serially diluted bacterial cultures. For the antibiotic sensitivity assay, commercially purchased filter discs (BBL Sensi-Disc susceptibility test discs; Becton, Dickinson and Company, Franklin Lakes, NJ) were used. After overnight incubation of PAO1 cells on LB agar plates containing 0 μM or 50 μM DTPA, the zones of inhibition were measured. Biofilm assays were performed as described previously (19).

TABLE 1.

Primers used for quantitative RT-PCR

Gene and primer type	PCR primer sequence (5′ to 3′)
lasB
Forward	CCGCAAGACCGAGAATGACA
Reverse	CTTCCCACTGATCGAGCACT
lasI
Forward	TTCAAGGAGCGCAAAGGCTG
Reverse	GTTCTTCAGCATGTAGGGGC
lasR
Forward	TCTGGGAACCGTCCATCTAC
Reverse	GACCGACTCCATGAAACGGT
rhlI
Forward	CTTCATCGAGAAGCTGGGCT
Reverse	AGGTAGGCGAAGACGTCCTT
rhlR
Forward	TTACTACGCCTATGGCGTGC
Reverse	TCGCTCCAGACCACCATTTC
pqsA
Forward	CCACTCCGCTGGACGACAAC
Reverse	GCAGCATGTGCGAGGGAATC
lasA
Forward	GACGACCTGTTCCTCTACGG
Reverse	GCTCCAGGTATTCGCTCTTG
aprE
Forward	ATGTACATCGTGCCCAACAG
Reverse	GGTCTTGCTCTGGTTGAAGG
phzA2
Forward	AACCACTTCTGGGTCGAGTG
Reverse	TCGAGTTCGAAGGAATGGAT
rpoD
Forward	AGGCAGTGGCTCACGACCCAT
Reverse	ATGCGACTTGGTGGATCCGTCA

A549 cell viability assay.

The maintenance and growth of A549 alveolar type II-like adenocarcinoma cells were performed as described previously (8). For treatment, A549 cells were inoculated in each well of 96-well plates (1 × 10⁴ cells/well), and the plates were incubated overnight under normal culture conditions (37°C and 5% CO₂). Prior to treatment, A549 cells were incubated with serum-free medium for 1 h, and bacterial culture supernatants were diluted 10-fold with A549 cells. When necessary, serum-free medium containing DTPA with the indicated concentrations was used. For treatment with live bacterial cells, A549 cells were infected with PAO1 cells or the ΔlasB mutant resuspended in serum-free medium at a multiplicity of infection (MOI) of 100:1. After 6 h of treatment, A549 cell viability was measured by MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] assays (8).

In vivo mouse airway infection.

Forty-two B6 mice (male, 7 weeks old) purchased from Orient Bio Inc. (Sungnam, South Korea) were divided into three groups (14 mice per group). The first group was infected with the ΔlasB mutant (30 μl, 4 × 10⁷ CFU/mouse) and treated daily with saline buffer for 5 days. The other two groups were infected with equal doses of the PAO1 strain (30 μl, 4 × 10⁷ CFU/mouse), and each group was treated with saline (vehicle control) or Ca-DTPA, respectively. Ca-DTPA was purchased from Sigma-Aldrich in the form of pentetate calcium trisodium salt, and mice were treated daily with a dose of 100 mg/kg of body weight. Ca-DTPA, instead of DTPA, was used in this animal experiment, because it is readily dissolved in an aqueous solution. Bacterial infection and drug delivery were performed via the intranasal route. Prior to infection and drug treatment, the mice were anesthetized intraperitoneally with 30 mg/kg Zoletil (Virbac Korea Co., Seoul, South Korea) and 10 mg/kg of Rompun (Bayer Korea Ltd., Seoul, South Korea). The right lobe of the lung was harvested and fixed with formalin. The lung tissue embedded in paraffin was sectioned and stained with hematoxylin and eosin, as previously described (21).

Statistical analysis.

An unpaired t test was performed to analyze the data, and all P values of <0.05 were considered statistically significant. A log-rank analysis was performed to assess the statistical significance of the mouse survival assays shown in Fig. 8A.

FIG 8.

Effect of DTPA treatment on the survival rate of P. aeruginosa-infected mice. (A) A Kaplan-Meier survival plot of B6 mice where mouse infection was performed as described in Materials and Methods. Mice infected with the ΔlasB mutant were treated daily with saline for 5 days. Mice infected with PAO1 cells were divided into two groups (14 mice per group), and each group was treated with saline or Ca-DTPA every 24 h for 5 days. The P value between these two groups was 0.017. (B) Representative images of lung tissue of B6 mice challenged with the ΔlasB mutant, PAO1 with Ca-DTPA, or PAO1 with saline. Bacterial cells (4 × 10⁷) were used for infection, and mice were sacrificed after 24 h.

RESULTS

Identification of a compound that suppresses elastase production in P. aeruginosa.

Elastase is considered a major virulence determinant in P. aeruginosa (8). To identify chemical compounds that can downregulate the activity of P. aeruginosa elastase, we screened a compound library that consists of 2,320 FDA-approved drugs. Sixteen compounds were found to prevent PAO1 cells from producing elastase in our initial screening as assessed by elastin-Congo red assays. Elastase activity was decreased by >75% compared with that of untreated PAO1 cells. Among the 16 compounds, 15 were determined to exhibit antimicrobial or antifungal activity, and suppression of elastase activity was a consequence of bacterial cell killing. As shown in Table 2, CFU counts were significantly decreased when PAO1 cells were treated with each of these 15 compounds. Importantly, diethylene triamine pentetic acid (DTPA, pentetic acid) (Fig. 1A), which induced a substantial level of suppression of elastase secretion, exerted no cytotoxic effects on bacterial viability. The elastase activity of PAO1 cells grown in the presence of 25 μM DTPA was only 17.3% of that of untreated PAO1 cells, and the number of viable cells after DTPA treatment was slightly higher than that in the untreated control (Table 1).

TABLE 2.

Compounds that suppress P. aeruginosa elastase activity measured by elastin-Congo red assays

Chemical no.	Chemical name	Source	OD495 ratio (%)a	CFU ratio (%)b	Bioactivity
MS10 E8	Ciprofloxacin	Synthetic	0	2.00	Antibacterial, fungicide
MS10 E10	Azithromycin	Semisynthetic	1.60	1.00	Antibacterial
MS01-D9	Norfloxacin	Synthetic; MK-366	2.30	0	Antibacterial
MS03-A2	Difloxacin hydrochloride	Synthetic; Abbott-56619	2.90	0	Antibacterial, DNA gyrase inhibitor
MS03-H4	Enrofloxacin	Synthetic	3.20	0	Antibacterial
MS04-A8	Mitomycin	Streptomyces verticillatus; NSC-26980	3.30	0	Antineoplastic
MS03-E8	Rifaximin	Semisynthetic	3.60	65.70	Antibacterial, RNA synthesis inhibitor
MS03-F10	Gemifloxacin mesylate	Synthetic; SB-265805S, LB-20304a	3.60	0.00	Antibacterial
MS01-E5	Cefmenoxime hydrochloride	Synthetic; Abbott-50192	4.40	1.90	Antibacterial
MS03-H11	Acrisorcin	Synthetic	5.60	14.6	Antifungal
MS02-H4	Pentetic acid	Synthetic; DTPA	17.30	128.60	Chelating agent, diagnostic aid
MS05-A7	Chlortetracycline hydrochloride	Streptomyces aureofaciens	18.30	1	Antibacterial
MS07-B4	Polymyxin B sulfate	Bacillus polymyxa	20.90	0	Antibacterial
MS06-C6	Glucosamine hydrochloride	Polysaccharides in bacteria	22.30	0	Antiarthritic
MS06-C5	Gentian violet	Synthetic	22.60	80	Antibacterial, Anthelmintic
MS06-D4	Hexachlorophene	Synthetic	23.80	0	Anti-infective (topical)

^aRelative elastase activity (OD₄₉₅ ratio after treatment with compound versus vehicle control).

^bCFU ratio (CFU after treatment with compound versus vehicle control).

FIG 1.

Effects of DTPA treatment on elastase secretion and bacterial growth. (A) Chemical structure of DTPA. (B) Dose-dependent effect of DTPA and EDTA on elastase activity. PAO1 cells were grown for 16 h in LB medium containing the indicated amount of DTPA or EDTA, and culture supernatants were assayed for elastase activity as described in Materials and Methods. Three independent experiments were performed, and the mean ± standard deviation (SD) values are displayed in each bar. *, P < 0.05 versus elastase activity from untreated PAO1 cells. (C) Dose-dependent effect of DTPA and EDTA on bacterial growth. The number of viable cells was calculated from each culture after 16 h of growth. Three independent experiments were performed, and the mean ± SD values are displayed in each bar. (D) Growth curves of PAO1 cells grown aerobically with shaking in LB medium, LB medium with 100 μM DTPA, or LB medium with 100 μM EDTA. Each growth curve experiment was repeated in triplicate, and the mean ± SD values are displayed in each graph. (E) Dose-dependent effect of DTPA treatment on elastase activity. The PAO1 strain was grown in LB medium containing various amounts of DTPA for 16 h, and culture supernatants were harvested to measure elastase activity. Aliquots of PAO1 culture supernatants derived from 16 h of culture in LB medium were treated with various concentrations of DTPA for 2 h and assayed for elastase activity. Three independent experiments were performed, and the mean ± SD values are displayed in each bar. *, P < 0.05 versus elastase activity from untreated PAO1 cells or culture supernatants. (F) Dose-dependent effect of DTPA, Ca-DTPA, or Gd-DTPA on elastase activity. PAO1 cells were grown for 16 h in LB medium containing the indicated amounts of DTPA, Ca-DTPA, or Gd-DTPA, after which culture supernatants were assayed for elastase activity. Experimental conditions were identical to those described for Fig. 1B.

DTPA suppresses elastase activity more efficiently than EDTA.

P. aeruginosa elastase is a metalloproteinase, and Zn²⁺ is involved in its catalytic mechanism (22, 23). As the role of DTPA as a suppressor of elastase secretion is likely mediated by its metal-chelating activity, we compared its elastase suppressive activity with that of EDTA, another prominent metal ion chelator. Although the elastase activity of PAO1 cells grown with increasing amounts of DTPA or EDTA gradually decreased, DTPA was more effective in repressing elastase activity. Upon growth with 20 μM DTPA, elastase activity was only ∼25% of that observed in untreated cells. In contrast, elastase activity was not reduced by treatment with 20 μM EDTA (Fig. 1B). When grown in the presence of 50 μM DTPA, elastase activity was completely abrogated in the elastin-Congo red assay. Nevertheless, about half of the original elastase activity was still detected when PAO1 cells were grown with an equal concentration of EDTA (50 μM) (Fig. 1B). These results demonstrated that DTPA suppresses elastase activity in a dose-dependent manner, and the ability of DTPA to antagonize elastase activity is considerably more efficient than that of EDTA.

Elastase production is controlled by QS, a known cell density-dependent gene regulatory mechanism in P. aeruginosa (6). To rule out the possibility that a decrease in elastase activity was a consequence of growth inhibition by DTPA, we monitored bacterial growth of PAO1 cells in the presence of DTPA. There was no difference in the numbers of viable cells after overnight growth with DTPA or EDTA (Fig. 1C). DTPA at levels of <100 μM did not exert any negative effects on bacterial growth. Likewise, PAO1 cells grown with up to 100 μM EDTA yielded similar numbers of viable cells. Growth curve experiments with 100 μM DTPA or EDTA indicated that the final optical density at 600 nm (OD₆₀₀) values were somewhat lower than those of the control. Bacterial growth, however, was not delayed by the presence of compounds, and no differences in growth phase development were detected (Fig. 1D).

Next, we set out to assess whether the DTPA-induced reduction of elastolytic activity is mediated by suppressed production of elastase or by direct inhibition of elastase activity. To address this issue, we performed an elastin-Congo red assay using the PAO1 culture supernatants, found to be rich in elastase (8). The culture supernatants were incubated with increasing concentrations of DTPA (0, 10, 50, and 100 μM) for 2 h before being assayed for the elastolytic activity. As shown in Fig. 1E, elastase activity was not affected at all when supernatants were treated with 10 or 50 μM DTPA. Even after treatment with 100 μM DTPA, nearly 45% of the original elastase activity was still detected. This finding suggests that DTPA-mediated suppression of elastase activity likely occurs at the production level, rather than upon posttranslational inhibition of enzyme activity. Thereafter, we tested whether elastase production might also be suppressed by either Ca-DTPA or Gd-DTPA, currently used for metal decontamination treatment (24) or as a contrast agent for magnetic resonance imaging (MRI) examination (25), respectively. As shown in Fig. 1F, Ca-DTPA was capable of suppressing elastase production, whereas elastase production was not affected in PAO1 cells in the presence of Gd-DTPA. The degree of inhibition by 100 μM Ca-DTPA was almost identical to that induced by DTPA, suggesting that the suppressive effect of DTPA may not be diminished when it is complexed with the Ca²⁺ ion.

Elastase activity is restored upon supplementation of Zn²⁺ and Mn²⁺ ions.

To better delineate the mode of inhibition by DTPA, we examined whether the suppressive effects of DTPA on elastase activity are alleviated by extraneously added metal ions; we tested ZnCl₂, CaCl₂, FeCl₃, MgCl₂, MnCl₂, and CuCl₂ (100 μM). In the absence of DTPA treatment, similar levels of elastase activity were observed in PAO1 cells grown without or with each of the metal ions (Fig. 2, left set). When grown with 50 μM DTPA and 100 μM ZnCl₂ or MnCl₂, PAO1 cells produced comparable levels of elastase, suggesting that the addition of Zn²⁺ or Mn²⁺ almost completely antagonized the suppressive effect of DTPA on elastase production (Fig. 2, right set). This result also suggests that Zn-DTPA and Mn-DTPA may not be potent suppressors of elastase production. The addition of Fe³⁺ or Mg²⁺ ions exerted no effects and addition of Ca²⁺ and Cu²⁺ ions exhibited intermediate effects (Fig. 2, right set) on DTPA-mediated suppression of elastolytic activity.

FIG 2.

Effect of metal ion supplementation on DTPA-induced elastase suppression. PAO1 cells were grown in LB medium containing ZnCl₂, CaCl₂, FeCl₃, MgCl₂, MnCl₂, or CuCl₂ (100 μM) in the presence or absence of DTPA (50 μM). Experimental conditions were identical to those described for Fig. 1B. *, P < 0.05 versus elastase activity derived from PAO1 cells grown in LB medium supplemented with each corresponding ion.

Decrease in elastase activity is regulated at the transcriptional level in DTPA-treated PAO1 cells.

Our results suggested that decreased elastase activity by DTPA treatment is likely due to a decrease in elastase synthesis. To address this notion, we sought to quantitate elastase protein levels. Figure 3A shows the Western blot analysis of culture supernatants and cytoplasmic extracts of PAO1 cells treated with various concentrations of DTPA or EDTA. The secretion of elastase to the culture medium was not compromised when cells were treated with 10 μM DTPA or EDTA (Fig. 3A). No elastase was detected in the culture supernatants when cells were treated with 50 μM DTPA. In contrast, elastase was detected, albeit at reduced levels, in the culture supernatant of PAO1 cells treated with equal concentrations of EDTA. This result further supports our previous finding (shown in Fig. 1B) that DTPA represses elastase activity more effectively than EDTA. PAO1 cells treated with 100 μM DTPA or EDTA failed to produce elastase. Importantly, accumulation of elastase inside bacterial cells was not observed in PAO1 cells that did not secrete elastase. This result demonstrates that DTPA-induced decreases in elastase activity are due to reduced synthesis of elastase and not to the defective secretion.

FIG 3.

Effects of DTPA treatment on elastase expression at the protein and mRNA levels. (A) Western blot analysis of P. aeruginosa elastase. PAO1 cells were grown for 16 h in LB medium with the indicated concentrations of DTPA or EDTA. Culture supernatants and cell extracts from each culture were prepared for the analysis. Black arrowheads indicate elastase-specific bands. Ten microliters of each culture supernatant and 10 μg of protein of each cell extract were loaded onto SDS-PAGE gels. (B) Quantitative real-time PCR analysis showing lasB gene transcript levels. PAO1 cells were grown to the mid-logarithmic phase (for ∼5 h) in LB medium containing increasing amounts of DTPA or EDTA and subjected to RNA extraction. *, P < 0.05 versus lasB transcript levels of untreated PAO1 cells. Transcript levels from PAO1 cells grown without DTPA were used for normalization.

Next, we measured transcript levels of the lasB gene that encodes elastase. Upon treatment with increasing concentrations of DTPA, a gradual decrease in the lasB transcript was clearly observed. EDTA treatment, however, did not result in a substantial decrease in lasB gene transcription levels (Fig. 3B). These results suggest that DTPA-induced suppression of elastase activity is regulated at the transcriptional level.

The PQS quorum-sensing system is repressed in DTPA-treated PAO1 cells.

In P. aeruginosa, the LasI/R, RhlI/R, and PQS QS systems participate in a complex signaling cascade to activate lasB gene expression (8). To determine the relative contribution of each QS system to DTPA-induced suppression of elastase synthesis, we measured the transcript levels of the lasI, lasR, rhlI, rhlR, and pqsA genes in response to DTPA treatment. Transcript levels of these genes were normalized to that of the rpoD gene, which exhibited a similar degree of expression throughout the treatment. Of note, transcription of the lasI, lasR, rhlI, and rhlR genes was not altered significantly in response to DTPA treatment (Fig. 4A). In contrast, mRNA levels of the pqsA gene were noticeably decreased in DTPA-treated PAO1 cells. This finding suggests that DTPA treatment resulted in the specific suppression of the PQS QS system. Consistent with this notion, production of the PQS compound was gradually decreased in response to treatment with increasing amounts of DTPA based on thin-layer chromatography (TLC) analysis (Fig. 4B).

FIG 4.

Effects of DTPA treatment on expression of QS-related genes and on PQS production. (A) Quantitative real-time PCR analysis showing the transcript levels of genes indicated at the bottom. PAO1 cells were grown to the mid-logarithmic phase (for ∼5 h) in LB medium containing 0 μM, 10 μM, 50 μM, or 100 μM DTPA. Experimental conditions were identical to those described for Fig. 3B. *, P < 0.05 versus pqsA transcript levels of untreated PAO1 cells. The transcript levels of each gene from PAO1 cells grown with 0 μM DTPA were used for normalization. (B) PQS levels detected in the culture supernatants of PAO1 cells grown with increasing concentrations of DTPA. PQS was detected by TLC analysis. The black arrowhead indicates PQS. Synthetic PQS was used as a positive control (first lane, 20 nmol). Semiquantitative image analysis was performed to show the relative intensities of each spot. Spot intensities were normalized with that of a positive control.

To provide additional insight into the effect of DTPA treatment on the expression of other virulence-associated genes, we performed qRT-PCR assays on lasA, aprD, and phzA2 genes, which are involved in the production of LasA protease, alkaline protease, and phenazine, respectively. As shown in Fig. 5, expression of lasA was also suppressed in PAO1 cells grown in the presence of DTPA. DTPA-induced suppression of lasA transcription occurred more effectively than DTPA-induced suppression of lasB expression. Even with treatment at 10 μM DTPA, the transcript level of lasA was decreased to 20% of what was observed in untreated cells. Transcription of aprD and phzA2 genes was either unaffected or increased.

FIG 5.

Quantitative real-time PCR analysis showing the transcript levels of lasA, aprD, and phzA2 genes. PAO1 cells were grown to the mid-logarithmic phase (for ∼5 h) in LB medium containing increasing amounts of DTPA and subjected to RNA extraction. *, P < 0.05 versus transcript levels of lasA and phzA2 genes in untreated PAO1 cells. Transcript levels from PAO1 cells grown without DTPA were used for normalization.

Effects of DTPA cotreatment on antibiotic susceptibility.

Some classes of antibiotics, such as tetracyclines and quinolones, are known to have the ability to bind metal ions (26, 27), suggesting that DTPA treatment may exert a synergistic effect on antibiotic-induced bacterial killing. To address this notion, we tested the effect of DTPA cotreatment on antibiotic-induced bacterial killing. Among 18 antibiotics that are currently used to treat P. aeruginosa infection, minocycline, a broad-spectrum tetracycline antibiotic, exhibited a noticeably enhanced killing effect on PAO1 cells, when used together with DTPA. The diameter of the cleared zone increased to 17 cm versus 11 cm in the presence of 50 μM DTPA (Table 3). On the other hand, the susceptibility of PAO1 cells in response to the other antibiotics was not changed significantly by DTPA cotreatment.

TABLE 3.

Antibiotic sensitivities of PAO1 in response to the presence of DTPA

Antibiotic(s)	Diam of the cleared zone (cm)
	Without DTPA	With DTPA
Ampicillin	6	6
Cefepime	30	31
Meropenem	28	30
Piperacillin	28	29
Imipenem	25	26
Ampicillin-sulbactam	6	6
Cephalothin	6	6
Cefotaxime	20	22
Aztreonam	27	26
Ceftazidime	29	29
Cefoxitin	6	6
Minocycline	11	17
Gentamicin	19	18
Amikacin	23	25
Piperacillin-tazobactam	29	30
Levofloxacin	30	27
Sulfamethoxazole trimethoprim	12	9
Tobramycin	23	23

P. aeruginosa cytotoxicity in cultured epithelial cells was attenuated in the presence of DTPA.

The elastase-mediated cytotoxicity of P. aeruginosa in A549 cells is well characterized. A549 cells lost viability when treated with PAO1 culture supernatants (8). As shown in Fig. 6A, the normal cellular morphology was disrupted in A549 cells after treatment with PAO1 culture supernatants for 6 h. In contrast, A549 cells incubated with ΔlasB mutant culture supernatants remained viable, and their relative viability was nearly 80% of that of untreated A549 cells (Fig. 6B). These results indicate that elastase is the predominant virulence determinant for P. aeruginosa cytotoxicity toward A549 cells. Importantly, culture supernatants of PAO1 cells harvested after growth in 50 μM or 100 μM DTPA were not cytotoxic to A549 cells. A549 cells maintained their normal cell shape (Fig. 6A), and relative viabilities were comparable to those of untreated A549 cells (Fig. 6B). Next, we incubated A549 cells with live bacteria with or without DTPA treatment to examine whether DTPA-mediated virulence attenuation occurs during the bacterial interaction with the host cells. A549 cell viability was decreased when A549 cells were incubated with PAO1 cells in the culture medium lacking DTPA (Fig. 6D). When DTPA was added at a 50 μM or 100 μM concentration, the PAO1-induced cytotoxic effects were not observed. Intermediate levels of cytotoxicity were observed when 10 μM DTPA was used (Fig. 6D). Again, ΔlasB mutant cells did not elicit any cytotoxic effect on A549 cells (Fig. 6C and D). Together, these results suggest that DTPA treatment can downregulate the P. aeruginosa pathogenic potential during interactions with host airway tissue.

FIG 6.

DTPA-mediated attenuation of P. aeruginosa cytotoxicity toward A549 cells. (A, B) A549 cells were treated for 6 h with culture supernatants of PAO1 and its isogenic ΔlasB mutant. Bacterial strains were grown in LB medium with increasing concentrations of DTPA (0, 10, 50, or 100 μM) for 16 h prior to treatment. (A) Microscopic images of A549 cells pictured after 6 h of treatment with PAO1 and ΔlasB mutant culture supernatants. All images were captured at the same magnification. (B) MTT assays showing the percentage of viable cells in response to the treatment. The mean ± SD values are displayed in each bar (n = 6). *, P < 0.05 versus viability of A549 cells treated with LB medium only. The relative survival was calculated by normalizing the LB medium control to equal 100% survival. CS, culture supernatant. (C, D) A549 cells (1 × 10⁵ cells) were treated with live bacteria (1 × 10⁷ cells) for 6 h. Bacterial cells grown to the mid-logarithmic phase were resuspended in antibiotic-free Dulbecco's modified Eagle's medium (DMEM) containing the indicated concentrations of DTPA prior to treatment. (C) Microscopic images of A549 cells pictured after 6 h of treatment with bacterial cells. All images were captured at the same magnification. (D) MTT assays showing the percentage of viable cells in response to treatment. The mean ± SD values are displayed in each bar (n = 6). *, P < 0.05 versus viability of A549 cells treated with DMEM only. The relative survival was calculated by normalizing the DMEM control to equal 100% survival.

Effects of DTPA treatment on P. aeruginosa biofilm formation and in vivo virulence.

Biofilm formation is another important virulence determinant of P. aeruginosa (3, 19). Because biofilm formation is also controlled by QS (28), we next examined the effect of DTPA treatment on P. aeruginosa biofilm formation. When green fluorescent protein (GFP)-labeled PAO1 cells were statically grown in plain LB medium for 24 h, a robust biofilm formed with a depth as high as 60 μm (Fig. 7A). In contrast, when cells were grown in LB medium amended with 50 μM DTPA, biofilm robustness was substantially decreased, and biofilm depth was about 33 μm (Fig. 7B and C). In addition, bacterial cell density was relatively low in several regions of the biofilm when cells were grown in the presence of DTPA, suggesting that DTPA treatment can reduce the biofilm-mediated virulence potential of P. aeruginosa.

FIG 7.

Effect of DTPA treatment on PAO1 biofilm formation. A PAO1 strain harboring a plasmid encoding green fluorescent protein (GFP) was grown statically for 24 h in LB medium containing 0 μM (A) or 50 μM (B) DTPA. Three-dimensional (top) and side (bottom) images of each biofilm are shown. Dimensions in each axis (x, y, and z) are 317.6 × 317.6 × 60 μm. The 60-μm biofilm depth in panel A was the highest depth processed. As can be seen in the images, the actual depth of untreated biofilm is clearly higher than 60 μm. (C) The green fluorescent intensities in each of the 117 sliced focal planes taken at 0.513-μm intervals for a total of 60 μm are plotted as a function of biofilm height and compared between two biofilms.

Finally, we examined whether DTPA treatment might alleviate pathogenic symptoms in P. aeruginosa-infected mice. Of note, mice infected with the ΔlasB mutant remained viable for 5 days (Fig. 8A). Again, this result strongly demonstrates that elastase is a crucial virulence determinant of P. aeruginosa in vivo. When PAO1-infected mice were treated with control solution, all 14 mice perished within 5 days. DTPA-treated mice, however, exhibited a significantly enhanced survival rate in response to PAO1 infection, as 6 out of 14 mice remained viable (Fig. 8A). The lung sections of mice challenged with ΔlasB mutant bacteria exhibited normal lung architecture when stained with hematoxylin and eosin (Fig. 8B). In contrast, a high degree of inflammatory cell infiltration was observed in the lung sections of mice infected with PAO1 (Fig. 8B, bottom panel). Importantly, the inflammatory response in PAO1-infected mice was less with DTPA treatment (Fig. 8B, middle panel).

DISCUSSION

As an opportunistic pathogen, P. aeruginosa is capable of establishing chronic infections in the abnormally altered airways of patients with bronchiectasis or cystic fibrosis (29). While antibiotic inhalation has been widely used to treat P. aeruginosa infections (30, 31), the long-term use of inhaled antibiotics was determined to be a risk factor for the selection of multiple drug-resistant P. aeruginosa strains (32). In this study, we screened a library of FDA-approved drugs and identified DTPA as a compound that can repress the production of elastase, one of the most critical virulence determinants of P. aeruginosa. The elastolytic and proteolytic activities were almost completely suppressed in the lasB mutant. In contrast, those activities were still detected in a lasA mutant (33). Alkaline protease was only attributed to the residual activity found in the lasB mutant (34). Together, these results demonstrated that lasB-encoded elastase is predominantly responsible for the elastolytic activity of P. aeruginosa. Consistent with this notion, previous results from our own and other groups indicated that elastase is the major protein secreted into culture medium (8, 35). Moreover, the survival rate of mice infected with the lasB mutant was not affected at all, while a significant virulence was observed in PAO1-infected mice. This result verified that the in vivo infectivity of P. aeruginosa is also dependent on lasB-encoded elastase. For these reasons, we postulated that it would be beneficial to identify a compound that can specifically inhibit elastase activity.

Application of DTPA as an anti-infective treatment is of potential interest for the following reasons. First, DTPA treatment does not induce growth inhibition, and, thus, bacterial cells may not experience a high degree of selective pressure that often leads to the generation of escape mutants. Second, since DTPA is already an FDA-approved drug (36), subsequent endeavors to apply DTPA as an anti-Pseudomonas drug might be facilitated without extensive clinical trials. DTPA in a complex with the gadolinium (Gd) ion was administered intravenously for MRI examination at a concentration of 0.1 mmol/kg (37). This amount is equivalent to 1.3 mM for an individual with a body weight of 60 kg. Our results suggested that 100 μM DTPA did not induce any cytotoxicity to cultured epithelial cells. Moreover, B6 mice did not show any abnormal side effects when intranasally treated with 0.2 mmol/kg Ca-DTPA (data not shown). Together, these findings clearly demonstrate that DTPA would exert no cytotoxic effects on humans.

DTPA possesses a high affinity for divalent metal ions (38). Our data show that DTPA complexed with Zn²⁺, Mn²⁺, or Gd³⁺ failed to suppress elastase activity, while Ca-DTPA was still capable of repressing elastase activity. Consistently, an FDA manual indicates that Ca-DTPA can form a stable complex with metals by replacing Ca²⁺ with a metal ion of higher binding affinity (http://www.remm.nlm.gov/dtpa.htm). On the other hand, gadolinium ion (Gd³⁺) was reported to form a very stable complex with DTPA with a greater binding affinity (39). Together, these results suggest that the suppressive effect of DTPA is strongly influenced by metal ions in the cellular environment, and this should be considered in the further development of DTPA as an agent to treat P. aeruginosa infection. It is of note that DTPA itself would have limited application, possibly due to the solubility issue. In our work, the DTPA stock solution was made either with 1 N HCl or DMSO. In contrast, Ca-DTPA is readily solubilized in water, rendering it more suitable for therapeutic application. Importantly, the suppressive effect of DTPA on elastase production was stronger than that of the well-known ion chelator EDTA, and this likely stems from the octadentate binding ability of DTPA, compared to the tetra- or heptadentate binding ability of EDTA (40). Consistent with these findings, the formation constants of DTPA with metal ions were higher than those of EDTA (41). Furthermore, the water solubility of DTPA is higher than that of EDTA (41), suggesting that DTPA has greater applicability for delivery to the conjunctiva or the airway.

Brumlik and Storey showed via lasB-lacZ translational fusion data that β-galactosidase activity was increased when the reporter strain was grown in a defined medium containing high levels of Zn²⁺ (42). It was therefore suggested that Zn²⁺ regulates the expression of elastase at the translational level. Based on our results, however, lasB transcription was noticeably decreased in DTPA-treated PAO1 cells, suggesting that its transcription was also affected by the depletion of metal ions. Of note, the level of PQS that contributes to lasB transcription (8) was considerably decreased in DTPA-treated PAO1 cells, further supporting the notion that DTPA treatment affects lasB expression at the transcriptional level. Given that expression of the lasB gene is controlled by P. aeruginosa QS systems, this may indicate that DTPA can globally influence QS regulation in P. aeruginosa. Consistent with this notion, biofilm formation, an event controlled by QS (43), and lasA gene transcription were also decreased by DTPA treatment.

P. aeruginosa elastase is a strong exotoxin that induces tissue damage; therefore, significant efforts have been made to develop chemical compounds that suppress elastase activity. Recently, the new synthetic Zn²⁺ chelators, tropolones and N-mercaptoacetyl-Phe-Tyr-amide, were identified as inhibiting elastase activity in PAO1 cells (44, 45). The in vivo efficacy of these two compounds, however, has not been examined. Our results demonstrated that 50 μM DTPA, a safe dosage for clinical use, attenuated elastase-mediated cytotoxicity in an in vitro tissue culture model. Furthermore, intranasal administration of Ca-DTPA considerably reduced PAO1-mediated pathogenic effects in a murine in vivo infection model.

Elastase production and biofilm formation were both suppressed by DTPA treatment. This suggests that DTPA can exert pleiotropic inhibitory effects on P. aeruginosa virulence. In our acute mouse infection model, a P. aeruginosa lasB mutant did not exhibit any cytotoxicity for the first 5 days. Furthermore, mice were protected from infection by the wild-type PAO1 strain, when administered a daily dose of Ca-DTPA. Therefore, DTPA-mediated suppression of elastase production may be of use in attenuating P. aeruginosa virulence, especially during the initial stages of infection. On the other hand, DTPA-induced biofilm inhibition is likely more important for virulence attenuation during the later stage. P. aeruginosa can establish persistent infection inside the airway mucus. Its proliferation, however, is deemed to occur at a very slow rate. Airway mucus is not nutritionally favorable and also contains a range of antimicrobial agents, including lysozyme and lactoferrin (46). Therefore, the rate of biofilm formation inside a patient's airway should be significantly lower than that of laboratory-grown in vitro biofilm. Based on this notion, the biofilm mode of infection is considered more important during the chronic stage. Together, our results indicate that DTPA may play differential roles, depending on the infection stage, and additive effects are anticipated to contribute to decreases in the pathogenic potential of P. aeruginosa.

In conclusion, we identified DTPA as an agent that diminishes P. aeruginosa virulence and evaluated its efficacy in vitro and in vivo. P. aeruginosa airway infection has been a source of global concern due to its chronic nature and resistance to antibiotic-mediated treatments. We expect that the use of DTPA will be appealing as a therapy for respiratory tract infection, because it can be easily delivered to the airway in its nebulized form and it prevents the production of elastase by transcriptional inhibition. We believe that the data presented in the current study offer hope that DTPA, possibly in combination with antibiotics, can be used to establish an efficient treatment strategy for curing P. aeruginosa airway infection.