Cell Death of Human Polymorphonuclear Neutrophils Induced by a Pseudomonas aeruginosa Cystic Fibrosis Isolate Requires a Functional Type III Secretion System

Abstract

With a coincubation model incorporating Pseudomonas aeruginosa and human polymorphonuclear neutrophils (PMNs), a cystic fibrosis (CF) P. aeruginosa isolate has been shown to resist the bactericidal action of PMNs and to induce their cellular death. An isogenic mutant of this CF isolate in which the type III secretion system was rendered nonfunctional was unable to induce cellular death of PMNs.

Pseudomonas aeruginosa is an opportunistic pathogen of humans which causes serious and often fatal infections, most notably in cystic fibrosis (CF) patients. Chronic respiratory infections, the associated host inflammatory responses, and the consequent pulmonary tissue destruction are the leading cause of morbidity and mortality in CF patients. Tissue destruction is in general attributed to the uncontrolled release of toxic mediators from necrotic polymorphonuclear neutrophils (PMNs) (7). Indeed, an excessive influx of PMNs at the site of infection is observed in CF patients (17, 18), but although activated, PMNs are not able to eliminate P. aeruginosa. A number of the P. aeruginosa virulence factors that contribute to pathogenesis during initial colonization and further chronic infection have been characterized. The mucoid exopolysaccharide (MEP) alginate, produced by strains that chronically infect CF patients, is considered to be a crucial factor in the persistence of P. aeruginosa in lungs (11). Recently, it has been shown that P. aeruginosa uses a specialized pathway, the type III secretion system, to secrete and translocate toxins into eukaryotic cells (28). Type III secretion systems are conserved in many gram-negative pathogens, and it is hypothesized that effectors inhibit the phagocytic response to infection and allow bacterial survival and multiplication (13, 16).

The aims of this study were to analyze the interaction between P. aeruginosa CF isolates and human PMNs and to evaluate the contribution of the type III secretion system to the resistance of P. aeruginosa to the bactericidal activity of PMNs.

To assess whether some P. aeruginosa strains modify the bactericidal function of PMNs, we first developed a cellular model of interactions between P. aeruginosa and human PMNs. The results presented here were obtained with two strains: PAO1, a nonmucoid standard strain, and CHA, a mucoid CF clinical isolate previously characterized in our laboratory as a producer of the MEP alginate when it was grown on Pseudomonas Isolation Agar (Difco) plates (4). The relevant properties of all strains and plasmids used in this study are summarized in Table 1. For interaction experiments, bacteria were cultivated overnight at 37°C and 300 rpm in Luria-Bertani (LB) liquid medium, diluted in LB medium, and grown to an optical density at 600 nm (OD₆₀₀) of 1. For infection, bacteria were collected by centrifugation, washed once with modified HEPES-buffered saline (mHBS) (15 mM HEPES, 8 mM glucose, 4 mM KCl, 140 mM NaCl, 0.5 mM MgCl₂, 0.9 mM CaCl₂), and opsonized for 5 min with pooled normal human serum (NHS). PMNs were collected from whole blood obtained by venipuncture from healthy volunteers (Etablissement de Transfusion Sanguine de l'Isère et de la Savoie) and purified by Percoll gradient centrifugation as described previously (6). The PMNs were washed twice with mHBS and resuspended to 10⁸/ml in mHBS. The viability of the PMNs, which was determined by acridine orange-ethidium bromide staining (20), was more than 95%.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Relevant genotype or phenotypea	Source or reference
Strains
E. coli DH5α		Gibco BRL
P. aeruginosa
PAO1	Wild type	B. Holloway
CHA	Mucoid CF isolate	25
CHA-D1	exsA::Gm mutant of CHA	This work
CHA-JC1	algD::Tc mutant of CHA	This work
CHA-JC1D1	exsA algD double mutant of CHA	This work
Plasmids
pNOT19	Apr	22
pMOB3	Cmr Kmr Apr	22
pUCPGm	Apr Gmr	23
pDD2	pUCP20-derived plasmid containing exsA gene	This work
pBR322	Apr Tcr	New England Biolabs
pEX100T	Apr	24
pUCP20	Apr	27

^aAp^r, ampicillin resistance; Tc^r, tetracycline resistance; Km^r, kanamycin resistance; Cm^r, chloramphenicol resistance; Gm^r, gentamicin resistance. 

The activities of PMNs against P. aeruginosa strains were evaluated by monitoring bacterial plate counts during coincubations. Samples (2.5 ml) containing 5 × 10⁶ PMNs/ml, 5 × 10⁷ CFU of P. aeruginosa (multiplicity of infection [MOI] of 10) per ml, and 10% NHS were incubated (250 rpm, 37°C) for 3 h. After the incubation periods indicated in Table 2, 50 μl of the coincubation sample was taken, serially diluted in LB medium, and pour plated on Pseudomonas Isolation Agar plates. CFUs were counted after 24 h of incubation at 37°C. In the absence of PMNs, the levels of growth of PAO1 and CHA bacteria were equal, even in the presence of 10% NHS (data not shown). In the presence of PMNs, during the first hour of coincubation, strain PAO1 was eliminated almost completely while the number of CHA bacteria remained stable. When coincubation continued, the number of living bacteria of the CHA strain started to increase rapidly and reached a fold increase of 2.8 in 3 h (Table 2). These results indicate that CHA is able to escape the bactericidal activity of PMNs.

TABLE 2.

Survival of P. aeruginosa strains in coincubation experiments with PMNs

Strain	Mean no. of CFU surviving ± SEa				Fold increaseb
	Initially	At 1 h	At 2 h	At 3 h
CHA	56 ± 15	44 ± 12	63 ± 15	160 ± 18	2.8
PAO1	60 ± 1	2 ± 1.5	2 ± 1	7 ± 3	0.1
CHA-D1	48 ± 0.8	42 ± 3.6	38.8 ± 2.4	62 ± 8	1.3
CHA-D1(pDD2)	45 ± 15	47 ± 15	52 ± 6	132 ± 30	2.9

^aValues are means ± standard errors of results of at least three experiments. The biggest standard errors, calculated for cytotoxic strains [CHA and CHA-D1(pDD2)], are probably a consequence of the varied functionalities of PMNs from one preparation to another. 

^bThe fold increases represent the ratios of the numbers of CFU at 3 h to the numbers of CFU initially. 

In order to establish a relationship between the resistance of the CHA strain and a possible alteration of PMN functions in the interactions, we assessed PMN viability during coincubation with bacteria by measuring the relative levels of release of the cytosolic enzyme lactate dehydrogenase (LDH) from PMNs into infection supernatants. Every 30 min, 100 μl of culture supernatant was harvested at 300 × g and analyzed photometrically for LDH activity with a cytotoxicity detection kit (Roche Diagnostics, Meylan, France). ODs for coincubation supernatants (OD_supernatants) or for total PMN extracts, obtained with 0.1% Triton X-100 (OD_totals) were measured in a 96-well plate reader (Labsystems, Eragny sur Oise, France). The percentage of cytotoxicity for each experiment was calculated with the following equation: (OD_supernatant/OD_total) × 100.

The viability of PMNs incubated without bacteria in mHBS or infected with PAO1 at an MOI of 10 was stable (at around 90%) over the 3-h period of incubation. In contrast, PMNs incubated with CHA started to die as soon as 30 min after infection and 80% of cell lysis was achieved 3 h after this (Fig. 1A). In order to establish whether CHA cells need to be phagocytosed in order to be able to induce cytotoxicity towards PMNs, cytochalasin D (Sigma), which inhibits actin polymerization and therefore phagocytosis, was added to the medium during interaction at a concentration of 5 μg/ml. The PMNs were preincubated in the presence of 5 μg of cytochalasin D per ml for 30 min before the addition of the bacteria. As can be seen in Fig. 1A, strain CHA was able to induce cell death with the same kinetics, even when cytochalasin D was present in the medium, indicating that the bacterium does not need to be internalized to exert cytotoxic activity. No more than the basal 10% level of cytotoxicity was detected in samples containing PMNs and cytochalasin D or PMNs, cytochalasin D, and PAO1 as controls. Furthermore, no cytotoxicity towards PMNs could be achieved with either 2×-concentrated CHA supernatants or lysed bacteria (data not shown).

FIG. 1.

Kinetics of mortality of PMNs in coincubation experiments with P. aeruginosa strains. The percentages of cytotoxicity were calculated according to the release of LDH activity. Data are the means of results of at least three experiments. (A) Comparison between CHA (●) and PAO1 (⧫) and the effect of cytochalasin D on the cytotoxicity of CHA (■). ▴, PMNs only. (B) Cytotoxicities of CHA-D1 (○) and of the complemented strain CHA-D1(pDD2) (▾). Results obtained with CHA () are shown.

One of the main characteristics of type III bacterial secretion systems, as represented by the system of Yersinia spp., is a requirement for active contact between living bacteria and the host cells (3). The facts that CHA cytotoxicity requires live bacteria and that cytochalasin D does not have any effect suggest that the cytotoxic phenotype might be due to the action of effector proteins secreted by the P. aeruginosa type III secretion system. To test this hypothesis, we first looked for the ability of CHA to secrete type III proteins from its secretion system. As in Yersinia spp., the secretion from the type III system of P. aeruginosa can be achieved in vitro when bacteria are cultured in a calcium-depleted medium (29). For analysis of extracellular proteins, P. aeruginosa strains were grown for 5 h in LB medium supplemented with 5 mM EGTA and 20 mM MgCl₂ and 40-μl samples of the culture supernatants were analyzed by 0.1% sodium dodecyl sulfate–12% polyacrylamide gel electrophoresis (PAGE) (Fig. 2). Several proteins were induced when CHA was cultivated in the calcium-depleted medium. Four of these proteins, analyzed by matrix-assisted laser desorption ionization–time of flight mass spectrometry (14), corresponded to previously identified type III secreted effectors (ExoS and ExoT) and members of the type III translocation apparatus (PopB and PopD) (9). Induction of the type III system by calcium depletion led to a much smaller amount of protein secretion with PAO1 than with CHA (Fig. 2). This result indicated that, although functional, the type III system in the PAO1 strain is less active than in the CHA strain.

FIG. 2.

Extracellular bacterial protein profiles. Silver nitrate-stained 0.1% sodium dodecyl sulfate–12% PAGE of supernatants of CHA (lanes 1 and 2), CHA-D1 (lanes 3 and 4), CHA-D1(pDD2) (lanes 5 and 6), and PAO1 (lanes 7 and 8) cultured with (lanes 2, 4, 6, and 8) or without (lanes 1, 3, 5, and 7) induction of the type III secretion system by calcium depletion. Molecular weight markers (MW) are shown, with weights being reported in thousands. The positions of ExoT, ExoS, PopB, and PopD are shown.

In order to test whether this secretion system is involved in the cytotoxicity of CHA, a mutant unable to synthesize type III secreted effectors or type III secretion machinery proteins was constructed. To do this, the exsA gene, which encodes the ExsA transcriptional factor necessary for type III system synthesis (9), was inactivated. The exsA gene from CHA was isolated by PCR with the primers 5′TTTGGGCCCATTCTACTCAT and 5′CGATCCGTGCTCATGGCT, based on published sequences (10). PCR-amplified exsA was cloned in a blunt-ended PstI site of pNOT19 (22) and then inactivated by insertion of the gentamicin cassette from pUCPGm (23) into the PstI site located in exsA. Allelic exchange of exsA was conducted as described previously (22). One double recombinant, CHA-D1, was selected and further characterized. Southern blot analysis confirmed that the correct recombination event occurs at the exsA locus (data not shown). For complementation experiments, the exsA gene was cloned into the pUCP20-derived plasmid pX12 downstream from a strong promoter isolated from CHA genomic DNA (1) and introduced into CHA-D1 by electroporation. The sequence analysis revealed that this promoter is located upstream from the sicA gene (31) and corresponds to nucleotides 3110 to 3514 of the M74132 sequence (EMBL data bank). Analysis of the CHA-D1 culture supernatants by PAGE showed that the mutant is unable to synthesize proteins secreted by the type III system under inducing conditions but that the CHA-D1(pDD2)-complemented strain secretes proteins in the same manner as the parental strain CHA (Fig. 2).

In interaction experiments with PMNs at an MOI of 10, the cytotoxic activity of the exsA mutant was completely abolished, resulting in the same phenotype as that of the noncytotoxic strain PAO1 (Fig. 1B), which indicates that the cytotoxicity of CHA towards PMNs requires a functional type III secretion system. In contrast, the exsA mutant complemented in trans with the wild-type exsA gene was able to induce 80% cytotoxicity in the first 30 min of incubation, which represents twice as much activity as that obtained with the parental strain, CHA. One hundred percent PMN cell lysis had already occurred after 2 h of coincubation (Fig. 1B). The more rapid cell death induced with CHA-D1(pDD2), when compared with that of CHA, may be explained by the presence of wild-type exsA alleles on a multicopy plasmid under the control of a promoter which may be more transcriptionally efficient than the native exsA promoter.

To test whether type III-dependent cytotoxicity is sufficient for the resistance of CHA to the bactericidal activity of PMNs, the bactericidal assay was performed as described above with the CHA-D1 mutant and the complemented strain CHA-D1(pDD2). Although it is noncytotoxic, like PAO1, CHA-D1 is not completely eliminated during interactions with PMNs (Table 2). The number of CHA-D1 bacteria remained stable during 3 h of interaction, with a fold increase of 1.3 at 3 h. The complemented exsA mutant was able to escape the bactericidal activity of PMNs even more efficiently than the parental strain, CHA. One of the main differences between the two noncytotoxic strains, PAO1, which is very sensitive to the bactericidal action of PMNs, and CHA-D1, which is more resistant, is the MEP alginate production ability of the latter. In order to test the role played by the MEP alginate in resistance to the bactericidal activity of PMNs, we constructed isogenic mutants in which the algD gene involved in the synthesis of the MEP alginate had been deleted. The algD gene from CHA was isolated by PCR with the primers 5′CGCTACCAGCAGATGCCCTCGGCC and 5′CGCGATGCCTATCGATAGTTATGG, according to the published sequence (5). The PCR-amplified algD gene was cloned in a SmaI site of pEX100T (24) and subsequently inactivated by insertion of the AvaI-EcoRI tetracycline cassette from pBR322. The final construction was introduced into the chromosomes of CHA and CHA-D1 as described previously (24). The double recombinants CHA-JC1 (an algD mutant) and CHA-JC1D1 (an exsA algD double mutant) were selected and further characterized by Southern blot analysis (data not shown). As expected, in the 3-h coincubation experiments, the CHA-JC1 strain displayed a fold increase of 2.3, showing a capacity to resist PMNs that was intermediate between that of the mucoid noncytotoxic CHA-D1 strain (fold increase, 1.3) and that of the mucoid cytotoxic CHA strain (fold increase, 2.8). The nonmucoid and noncytotoxic CHA-JC1D1 mutant was less resistant than the exsA mutant, with a fold increase of 0.7. Thus, we can define four classes of P. aeruginosa strains based on their capacities to resist the bactericidal activity of PMNs (as measured by the fold increase in 3 h of coincubation with PMNs): noncytotoxic and nonmucoid strains (very low resistance), noncytotoxic but mucoid strains (low resistance), cytotoxic but nonmucoid strains (mild resistance), and cytotoxic and mucoid strains (complete resistance).

Taken together, these results show that the cytotoxicity of the CF clinical isolate CHA towards human PMNs is dependent on the functional type III secretion system and that effectors secreted and translocated into PMNs may be responsible for PMN cell death.

Four P. aeruginosa cytotoxins secreted by the type III secretion system, including two isoforms of exoenzyme S (ExoS and ExoT [28]), have been identified to date. ExoS and ExoT possess ADP-ribosylating activities towards low-molecular-weight GTP-binding proteins of the Ras family (2). The expression of ExoS correlates with cytotoxicity towards Detroit 532 fibroblasts (19) and CHO cells (21), with significant loss of the cell viability after only 24 h of infection. ExoY, a recently discovered adenylate cyclase, was shown to be responsible for the rounding up of CHO cells (30). These three type III system-secreted effectors have pronounced effects on cell morphology (26). However, no induction of rapid cell death, as observed in our cellular model, was reported. Another powerful type III system-secreted cytotoxin, ExoU (PepA), was identified with cocultures of P. aeruginosa PA103 and Madin-Darby canin kidney (MDCK) epithelial cells (8, 12). The expression of ExoU was directly correlated with the acute cytotoxicity measured 3 h after infection (8, 12).

To check whether the rapid cell death of PMNs induced by strain CHA is due to the activity of ExoU, the supernatant of the CHA cultures grown under inducing conditions (addition of EGTA) was analyzed. No ExoU-specific amino acid sequences were found by matrix-assisted laser desorption ionization–time of flight analyses of proteins migrating between 60 and 70 kDa. In addition, Southern blot analysis with an exoU-specific probe showed that the exoU gene is not present in the CHA genome (data not shown). Subsequent to the submission of this article, Coburn and Frank (1a) reported the cell death of macrophages induced with the ExoU-deficient P. aeruginosa strain 388. Our preliminary experiments performed on the J774 macrophage cell line infected with the CHA strain showed kinetics of cell killing similar to those with PMNs. It is possible that the mechanism of phagocyte killing, which is ExoU independent, is the same in our cytotoxic CF clinical isolate as in strain 388.

We reported in this work that in addition to type III cytotoxicity, other factors, such as MEP alginate synthesis, are involved in resistance to the bactericidal activity of PMNs, although the interaction experiments described in this paper were carried out under conditions (liquid culture and early exponential phase of growth) in which the synthesis of the MEP alginate measured by the carbazole method (15) was low (<10 μg/ml).

The induction of PMN cell death by the CF clinical isolate CHA under ex vivo conditions may explain the increased inflammatory response in CF lungs by the promotion of the release of toxic mediators from PMNs.