Bactericidal properties of group IIa secreted phospholipase A2 against Pseudomonas aeruginosa clinical isolates

Phospholipases A₂ (PLA₂) cleave fatty acids from the sn-2 position of phospholipids from bilayers or micelles, yielding free fatty acids and lysophospholipids. They have been classified into several groups based on their secretory or cytosolic nature and their molecular structure. Group IIa secreted PLA₂ (IIa sPLA₂) has been attracting interest for almost two decades, since its discovery in platelets (Murakami et al., 1997) and synovial fluids and its association with inflammatory disorders (Valdas et al., 1993). Much effort has been expended to determine the precise role of sPLA₂ in physiological mechanisms. Interestingly, Weinrauch et al. (1996) demonstrated that, in infectious contexts, inflammatory exudates could exhibit potent antibacterial properties against both Gram-positive and Gram-negative bacteria. Furthermore, it has been demonstrated that, during systemic bacterial challenge, sPLA₂ is fully mobilized, conferring on plasma a potent bactericidal activity against Escherichia coli and Staphylococcus aureus (Weinrauch et al., 1998). In non-pathological circumstances, physiological fluids such as tears are naturally enriched in sPLA₂: concentrations as high as 60 µg ml^-1 are frequently observed and are able to kill bacteria of the local flora such as Micrococcus species in vitro (Qu & Lehrer, 1998). More recently, studies involving transgenic mice expressing human group II sPLA₂ gene (PLA₂⁺ mice) and group II PLA₂-deficient mice (PLA₂^- mice) allowed an accurate assessment of the antimicrobial role of group II sPLA₂ in vivo. For this purpose, Laine et al. (1999) evaluated the response of PLA₂^- mice versus PLA₂⁺ mice when challenged by Staphylococcus aureus and noted an increased morbidity and mortality in the first group. In contrast, expression of sPLA₂ resulted in improved resistance of the animals by increased killing of bacteria, as indicated by the small number present in the tissues (Laine et al., 1999). Similar results were obtained when transgenic mice were challenged by E. coli (Laine et al., 2000). These results emphasize the importance of sPLA₂ in the inflammatory response related to bacterial invasion and its major role in host defence.

Biochemical studies have shown that the bactericidal properties of sPLA₂ rely on cell-wall phospholipid hydrolysis. While destruction of Gram-positive rods has been mostly attributed to direct group IIa sPLA₂ activity, Gram-negative bacteria are considered to require both sPLA₂ and co-factors. These co-factors [e.g. complement system or antimicrobial peptides secreted by polymorphonuclear neutrophils such as bactericidal/permeability-increasing protein (BPI)] are thought to be responsible for an acute disruption of the bacterial envelope (Elsbach et al., 1994). Nevertheless, the structurally related murine group II sPLA₂ secreted by Paneth cells of the intestine is able to display, on its own, effective bactericidal properties against Gram-negative pathogens such as Salmonella typhimurium and E. coli (Harwig et al., 1995). Therefore, group II sPLA₂ is one of the first intestinal host-defence mechanisms.

Among Gram-negative rods, Pseudomonas aeruginosa is an increasingly prevalent opportunistic pathogen and is the most prominent rod isolated from cystic fibrosis (CF) lungs. In airway secretions, it is associated with increased inflammation and a highly salt-enriched microenvironment known to impair innate immunity components such as defensins (Smith et al., 1996). Almost 80 % of CF patients beyond their teens are infected by P. aeruginosa and are more likely to die than patients with other kinds of pneumonia. Furthermore, P. aeruginosa is one of the few bacterial species that is naturally resistant to several ß-lactam antibiotics and frequently evolves to multiresistant antibiotypes. Therefore, in the past decade, a great deal of effort has been put into attempts to make the P. aeruginosa outer membrane permeable to antibiotics in order to promote bactericidal activity.

The aim of the present work was to assess whether IIa sPLA₂ could display direct bactericidal properties against P. aeruginosa clinical isolates and to study the impact of the enzyme on the P. aeruginosa cell wall.

Bacterial strains.
Staphylococcus aureus ATCC 25923, P. aeruginosa ATCC 27583 and PAO1 and Burkholderia cepacia ATCC 25416^T were used as reference strains. Clinical isolates of the three species were obtained from the collection of the Bacteriology department of Rangueil Hospital (Toulouse, France) and their origins are listed in Table 1. All strains were stored in glycerol at -80 °C and were plated on blood agar before use.

Table 1. Characteristics of the strains tested Abbreviations: CF, cystic fibrosis; MR, multiresistant strain.

Group IIa sPLA₂.
Recombinant human group IIa sPLA₂ was produced in E. coli as described by Fourcade et al. (1995).

Assay for bactericidal activity.
Bacteria were grown overnight at 37 °C in trypticase soy broth (TSB) and then diluted 1 : 10 in fresh medium and subcultured to mid-exponential phase. After harvesting, the bacteria were sedimented by centrifugation at 14 000 g for 5 min, washed twice with 10 mM HEPES/NaOH (pH 7.6), 0.15 M NaCl, 10 mg BSA ml^-1 and 2.5 mM CaCl₂ and resuspended in the same solution at a concentration of 10⁸ c.f.u. ml^-1.

Bactericidal activity was assayed as follows. Bacterial suspensions (10⁸ c.f.u.) were incubated with appropriate dilutions (8, 16 and 32 µg ml^-1) of group IIa sPLA₂ in the saline buffer described previously containing either BSA or serum. Mixtures were incubated at 37 °C with shaking for 1 or 3 h and then filtered on 0.20 µm Millipore filters. Membranes were then plated onto trypticase soy agar (TSA) and grown at 37 °C for 18 h in an aerobic atmosphere.

The bactericidal effect of group IIa sPLA₂ was expressed as the residual number of c.f.u. with respect to the initial inoculum and the EC₅₀ (concentration eliciting 50 % effect) corresponding to the enzyme concentration able to kill 50 % of the initial inoculum. The results presented are means of at least three independent experiments.

Effects of salt concentration and proteins on bactericidal effects of sPLA₂.
The bactericidal properties of sPLA₂ were tested in the presence of various incubation media. These media were BSA in the range 0 to 10 mg ml^-1 and NaCl from 0 to 0.25 M. Results were expressed as means ± SD of triplicate determinations. Statistical analysis was then performed using a paired t-test.

Radiolabelling of bacterial lipids.
Bacteria were subcultured to mid-exponential phase in TSB supplemented with 2 µCi [1-³²P]oleic acid ml^-1 (Amersham Lifescience France) (1 Ci = 3.7 x 10⁷ Bq) and 0.1 % BSA (w/v). Bacteria were then harvested and cultured in the same medium without [1-³²P]oleic acid at 37 °C for 30 min in order to remove unesterified fatty acid precursor incorporated into ester positions. Finally, the bacteria were washed with saline buffer.

Lipid analysis.
Bacteria (10⁸ c.f.u.) were exposed to 5 or 10 µg sPLA₂ ml^-1 for up to 3 h. The composition of the lipids was then determined by extraction in CHCl₃/CH₃OH as described by Bligh & Dyer (1959) and analysis of the radiolabelled material recovered in the CHCl₃ phase (> 95 % of total) by TLC in the presence of standards. Results were expressed as the percentage hydrolysis of phospholipids compared with a control performed in the absence of the enzyme. Data are means ± SD of three independent experiments.

Assay for cytotoxicity to human pulmonary cells.
In order to test the possible cytotoxic effect of group IIa sPLA₂ on human cells, the viability of tracheal epithelial cells incubated with an appropriate dilution of the enzyme was evaluated for 24 h.

Immortalized and characterized tracheal epithelial cells from fetuses were a gift from URA CNRS 1283 (Paris, France) (Lemnaouar et al., 1993). Cells were grown to confluence at 37 °C under 5 % CO₂ in 1 : 1 DMEM/Ham's F12 medium supplemented with 2 % (w/v) Ultroser G, 100 IU penicillin ml^-1 and 100 µg streptomycin ml^-1.

Solutions containing 8, 16 or 32 µg sPLA₂ ml^-1 were mixed with DMEM while the same volume of saline buffer was used as a negative control. A 10 % (v/v) solution of Triton X-100 in DMEM was used as a positive control. At 0, 6, 12 and 24 h, the supernatant was carefully removed and centrifuged at 14 000 g for 5 min in order to remove cells. Lactate dehydrogenase (LDH) in the supernatant was evaluated and a ratio LDH_assay/LDH_{positive control} > 5 % was considered to indicate cytotoxicity. The cytotoxicity test was performed three times for each treatment.

Group IIa sPLA₂ is able to kill P. aeruginosa clinical isolates in a concentration-dependent manner
In order to determine whether group IIa sPLA₂ could display antibacterial properties against P. aeruginosa clinical isolates, viable c.f.u. were evaluated after incubation of 10⁸ c.f.u. ml^-1 with various concentrations of sPLA₂ for 1 h. Since the enzyme was recombinant, we decided to control its activity against Staphylococcus aureus strains, which are known to be susceptible, in parallel. Under these experimental conditions, a marked reduction of the inoculum was observed for both species within 1 h for concentrations as low as 8 µg ml^-1 (Fig. 1). A progressive reduction of approximately 90 % of the initial inoculum was observed for all P. aeruginosa strains when concentrations were increased gradually up to 32 µg ml^-1 (data not shown). Interestingly, as listed in Table 2, comparable bactericidal properties were observed for both smooth and mucoid strains, which displayed similar EC₅₀ after 1 h of incubation. However, two CF strains isolated from adult patients required larger amounts of sPLA₂. In contrast, little or no sPLA₂ activity could be observed against the two B. cepacia strains tested. It is interesting to note that the range of concentrations necessary to observe a bactericidal effect was almost identical for the Staphylococcus aureus and P. aeruginosa strains tested. Furthermore, these results are the first evidence of direct activity of group IIa sPLA₂ against a Gram-negative species without the addition of any co-factor.

(28K): Fig. 1. Recombinant human sPLA2 is able to kill Staphylococcus aureus and P. aeruginosa strains in vitro. Samples were incubated for 60 min with 8, 16 and 32 µg sPLA2 () ml-1 or saline buffer () and then grown on TSA. Bactericidal effect was evaluated by counting c.f.u. after 24 h and is expressed as the percentage of inoculum remaining (c.f.u. %). Strong activity was observed for both Staphylococcus aureus strains ATCC 25923 (a) and 0412 (b) and P. aeruginosa strains ATCC 27583 (c) and 7844 (d) over the same range of sPLA2 concentrations, while B. cepacia strains ATCC 25416T (e) and 235 (f) were more resistant. Data shown are means ± SD (n = 3).

Table 2. sPLA2 displays time-dependent antimicrobial activity against Staphylococcus aureus and P. aeruginosa strains Bacterial strains were incubated with various concentrations (8, 16 or 32 µg ml-1) of sPLA2 for 1 or 3 h. The bactericidal effect of the enzyme was assessed by the enzyme concentration necessary to kill 50 % of the initial inoculum (EC50), which was deduced from curves obtained from three independent experiments.

Group IIa sPLA₂ bactericidal activity is time-dependent
In order to investigate whether bactericidal activity could be enhanced by longer incubation, all strains were subjected to sPLA₂ for up to 3 h. Table 2 lists the relative EC₅₀ values for bactericidal activity against all the strains tested. Recombinant sPLA₂ was thus more effective against both Staphylococcus aureus and P. aeruginosa strains when the incubation time was increased. The minimal concentrations required for EC₅₀ were mostly close to 28 µg ml^-1 when incubated for 3 h, compared with 512 µg ml^-1 after only 1 h. These results suggest that bactericidal activity is time-dependent and that increased contact between enzyme and bacterial strains, as would occur under physiological conditions, dramatically enhances the antibacterial properties of sPLA₂.

Group IIa sPLA₂ bactericidal properties are not affected by high concentrations of protein or NaCl
In order to test whether the abnormal composition of CF airway secretions could affect sPLA₂, we tested whether the enzyme bactericidal properties could be modified by protein and salt variations. As shown in Table 3, similar bactericidal effects were observed for concentrations as high as 32 µg sPLA₂ ml^-1, whatever the test conditions; no significant difference was observed either when the protein content was increased or when high salt concentrations were added to the incubation medium (P > 0.05). When human serum was used instead of BSA, no inhibitory effect could be observed (data not shown). These results demonstrate that, even if an unfavourable microenvironment is present, as observed in CF lungs, sPLA₂ is able to display potent bactericidal activity against both Staphylococcus aureus and P. aeruginosa strains.

Table 3. Effects of protein (BSA) and NaCl on bactericidal properties of sPLA2 against P. aeruginosa P. aeruginosa ATCC 29583 (108 c.f.u. ml-1) was incubated at 37 °C with 32 µg recombinant human sPLA2 ml-1 supplemented with various concentrations of BSA and NaCl (M) as described in Methods. After 60 min, bacterial viability was measured as described in Methods. Results are expressed as percentage viability with respect to bacteria incubated without sPLA2 and are means of three independent experiments.

Group IIa sPLA₂ bactericidal activity parallels phospholipid hydrolysis
In order to evaluate whether the antimicrobial properties of group IIa sPLA₂ against P. aeruginosa paralleled significant hydrolysis of the bacterial phospholipids, as demonstrated for Staphylococcus aureus strains, the ability of the enzyme to liberate oleic acid from the bacterial membrane was also studied.

As shown in Fig. 2, incubation of radiolabelled bacteria with 5 µg sPLA₂ ml^-1 led to the degradation of 60 % of the major phospholipids within 3 h. Furthermore, increased degradation of phospholipids was observed for higher concentrations (10 µg ml^-1). These results indicate that the ability of IIa sPLA₂ to hydrolyse phospholipids of the bacterial envelope correlates with the concentration- and time-dependent bactericidal activity.

(25K): Fig. 2. Recombinant type IIa sPLA2 bactericidal activity against P. aeruginosa strains parallels bacterial phospholipid envelope hydrolysis. P. aeruginosa ATCC 27583 was metabolically labelled in TSB supplemented with [1-32P]oleic acid and 0.1 % (w/v) BSA, chased with non-radioactive TSB and then incubated with 5 (filled bars) or 10 (open bars) µg recombinant sPLA2 ml-1 or without any enzyme (hatched bars). Bacterial phospholipid hydrolysis was determined by TLC and liquid scintillation counting as described in Methods. Results correspond to the percentage of phospholipid hydrolysis with respect to an untreated control. Data are means ± SD of three experiments.

Group IIa sPLA₂ is not cytotoxic to human cells
Human tracheal epithelial cells were incubated with saline buffer or sPLA₂ concentrations identical to those used to display bactericidal activity in order to investigate possible cytotoxic effects. As shown in Table 4, for all samples, the ratio LDH_assay/LDH_{positive control} was <5 %, even after 24 h incubation and was considered not to indicate cytotoxicity. These results show that sPLA₂ concentrations able to display bactericidal activity are totally innocuous to human pulmonary cells.

Table 4. Viability of human tracheal epithelial cells incubated with sPLA2 Cells were grown in DMEM supplemented with saline buffer (negative control), IIa sPLA2 or Triton X-100 (positive control). Supernatants were removed carefully and lysis was assessed by the ratio LDHassay/LDHpositive control, expressed below as a percentage. A ratio >5 % was considered to indicate cytotoxicity.

P. aeruginosa is an increasingly prevalent opportunistic pathogen and is frequently isolated from CF chronic lung infections, which are responsible for considerable morbidity in over 80 % of patients (Cystic Fibrosis Foundation, 2001). P. aeruginosa chronic infection is essentially characterized by the emergence of mucoid strains that produce an exopolysaccharide coat, which confers resistance to phagocytosis and antibiotics and is responsible for major obstruction of the airways (Doggett et al., 1966). Furthermore, P. aeruginosa infection is accompanied by an acute inflammatory response and altered antimicrobial peptide activity due to the highly salt-enriched lung secretions (Smith et al., 1996). The ineffectiveness of innate host-defence mechanisms is clearly a major clinical problem for CF patients. Furthermore, this bacterium is intrinsically resistant to many antibiotics. Part of this resistance can be attributed to the relatively low permeability of the P. aeruginosa outer membrane to a variety of antibiotics (Hancock & Wong, 1984; Yoshimura & Nikaido, 1982). Another part of the resistance appears to be caused by multidrug efflux systems (Poole et al., 1996a, b). Finally, in some cases, enzymes that specifically inactivate ß-lactam antibiotics, e.g. inducible cephalosporinase or imipenemase of P. aeruginosa (Giwercman et al., 1990; Livermore, 1987; Richmond & Sykes, 1973), can be secreted. Hence, over the past decade, a great deal of effort has gone into attempts to make the Gram-negative outer membrane permeable to antibiotics and to deliver drugs as close as possible to the target, e.g. by using tobramycin nebulization.

In this report, we have examined the effects of recombinant IIa sPLA₂ on the viability of P. aeruginosa clinical isolates. Several investigators had shown previously that IIa sPLA₂, initially described as part of the inflammatory response, also displays bactericidal activity against Gram-positive bacteria such as Staphylococcus aureus, while activity against E. coli strains requires co-factors. The latter include the complement system or BPI and p15s, produced by neutrophils (Weinrauch et al., 1995), which lead to synergistic disruption of the bacterial envelope. Nevertheless, the closely related murine type II sPLA₂ produced by Paneth cells is able to exert a direct antimicrobial effect against Gram-negative rods (Harwig et al., 1995).

The first step in our study was to determine recombinant IIa sPLA₂ activity against P. aeruginosa in parallel with Staphylococcus aureus strains in order to validate the efficacy of the recombinant enzyme. We found that IIa sPLA₂ could be fully bactericidal against the P. aeruginosa and Staphylococcus aureus clinical isolates in a concentration-dependent fashion and at a similar range of concentrations for both species without the addition of any co-factors. To our knowledge, these results provide the first demonstration that IIa sPLA₂ can be intrinsically microbicidal to a Gram-negative species. It is interesting to note that sPLA₂ displayed comparable properties against both smooth and mucoid strains, indicating that bacterial alginate does not necessarily interfere with the bactericidal properties of the enzyme. Furthermore, most of the P. aeruginosa strains tested expressed antibiotic-multiresistant phenotypes, which did not interfere with their susceptibility to sPLA₂.

The abnormal composition of CF airway secretions is known to affect the action of antimicrobial peptides, since antimicrobial function is often dependent on ion strength and salt sensitivity is crucial for cationic peptides (Smith et al., 1996). We therefore tested the effects of protein and salt variations as well as acidic pH (data not shown) in order to mimic the CF lung microenvironment. No difference in activity was found among the various groups. This is of particular interest because it shows that, unlike many host-defence antimicrobial peptides. which are totally ineffective in CF because of airway secretion hyperosmolarity, sPLA₂ remains active.

We finally decided to evaluate the time-course of bactericidal properties and demonstrated that incubation for up to 3 h significantly enhanced the reduction of the initial inoculum. Taken together, these results show that IIa sPLA₂ can be an effective microbicide, the direct antibacterial potency of which against P. aeruginosa approximates to that against Staphylococcus aureus even under unfavourable conditions close to those observed in CF lungs.

We also observed that the enzyme was responsible for phospholipid hydrolysis and concomitant liberation of free fatty acids, as already demonstrated for Gram-positive bacteria. Indeed, incubation of radiolabelled bacteria with 5 µg sPLA₂ ml^-1 led to degradation of 60 % of the major phospholipids within 3 h. This observation is consistent with the amount expected (e.g. 50 %) when only the phospholipids of the extracellular leaflet of the bilayered membrane (which is accessible to sPLA₂) are hydrolysed (Koduri et al., 2002), and is in agreement with observations on Micrococcus luteus strains (Buckland et al., 2000). Moreover, the degree and time-course of phospholipid hydrolysis correlated well with bacterial viability, even for mucoid and antibiotic-multiresistant strains.

These findings suggest that IIa sPLA₂ is intrinsically able to cause disruption of bacterial envelope integrity, thus leading to impairment of bacterial viability. One could then imagine that the enzyme could potentially synergize the effects of antibiotics such as ß-lactam in combination with natural antibacterial peptides such as BPI. Because no cytotoxicity could be observed for concentrations comparable to those necessary to display bactericidal activity against P. aeruginosa strains, IIa sPLA₂ may be considered as a potential therapeutic agent synergizing antibiotics. However, one of the main problems when administering antibiotics to CF patients is how to achieve targeted drug delivery to the diseased tissue in order to obtain microbicidal concentrations as high as possible while limiting side effects. Therefore, new administration routes such as tobramycin nebulization have been proposed in recent years (Doring et al., 2000). In this context, the dual potency of IIa sPLA₂ to hydrolyse not only bacterial phospholipids but also micelles of phospholipid bilayers should be of particular interest. Hence, new targeting systems such as liposomal drug-carrier systems are being developed. The action of IIa sPLA₂ on liposomes can be used to release their molecular content (e.g. drugs) directly where needed. This method has in fact been demonstrated to be fully efficient for anti-cancer drug delivery, while tumorous tissues are known to hyperexpress IIa sPLA₂ (Davidsen et al., 2003). One could easily imagine that this method could be used in CF therapy.

Additional work is obviously needed to clarify the various possible strategies. Nevertheless, these findings may prove interesting with regards to new therapeutic strategies, particularly since more drug-resistant bacterial strains are emerging.

The authors want to thank Mr L. Aguéda for his technical assistance and URA CNRS 1283 for providing the tracheal epithelial cells. This work was supported by grant no. 00.014/DR12/UPR9027 from Ministère de l'Education Nationale, de la Recherche et de la Technologie (PRFMMIP, 2000).