Enteric gram-negative bacilli, such as Escherichia coli are the most common cause of nosocomial pneumonia. In this study a wild-type extraintestinal pathogenic strain of E. coli (ExPEC)(CP9) and isogenic derivatives deficient in hemolysin (Hly) and cytotoxic necrotizing factor (CNF) were assessed in vitro and in a rat model of gram-negative pneumonia to test the hypothesis that these virulence factors induce neutrophil apoptosis and/or necrosis/lysis. As ascertained by in vitro caspase-3/7 and LDH activities and neutrophil morphology, Hly mediated neutrophil apoptosis at lower E. coli titers (1 × 105–6 cfu) and necrosis/lysis at higher titers (≥1 × 107 cfu). Data suggest that CNF promotes apoptosis but not necrosis or lysis. We also demonstrate that annexin V/7-amino-actinomycin D staining was an unreliable assessment of apoptosis using live E. coli. The use of caspase-3/7 and LDH activities and neutrophil morphology supported the notion that necrosis, not apoptosis, was the primary mechanism by which neutrophils were affected in our in vivo gram-negative pneumonia model using live E. coli. In addition, in vivo studies demonstrated that Hly mediates lung injury. Neutrophil necrosis was not observed when animals were challenged with purified lipopolysaccharide, demonstrating the importance of using live bacteria. These findings establish that Hly contributes to ExPEC virulence by mediating neutrophil toxicity, with necrosis/lysis being the dominant effect of Hly on neutrophils in vivo and by lung injury. Whether Hly-mediated lung injury is due to neutrophil necrosis, a direct effect of Hly, or both is unclear.
- Escherichia coli
- gram-negative bacilli
- cytotoxic necrotizing factor-1
nosocomial pneumonia accounts for 15% of hospital-acquired infections (2). Gram-negative enteric bacilli such as Escherichia coli, Pseudomonas aeruginosa, Klebsiella pneumonia, Enterobacter sp., Acinetobacter, and Stenotrophomonas are the most common cause of hospital-acquired pneumonia, being implicated in 55–85% of cases (8–10, 24). Severe disease usually occurs when these agents cause pneumonia, with associated crude mortality and estimated attributable mortality ranging from 24 to 76% and 20 to 50%, respectively, which translates into 36,000–80,000 deaths annually (2, 8, 24, 52). Over the last 10–15 years, there has been little improvement in the outcome from this infection. The estimated annual cost of nosocomial pneumonias alone due to gram-negative bacilli in this country is greater than one billion dollars (55).
A goal of our laboratory is to identify strategies to decrease the morbidity and mortality caused by gram-negative pneumonia. To accomplish this, we have been studying neutrophil-pathogen interactions in a rat model of bacterial pneumonia, using E. coli as a model pathogen (37, 38, 43). The importance of neutrophils in protecting against infection in the lung is as great as any other site in the body (46). Increased susceptibility and severity of pneumonia occur in the setting of neutropenia and with genetically inherited (e.g., chronic granulomatous disease) or drug-induced (e.g., steroids) disorders of neutrophil function (25). In animal models, enhanced recruitment and activation of neutrophils by TNF-α and CXC chemokines result in enhanced bacterial clearance and survival in gram-negative pneumonia (14, 28, 48, 49, 51). Likewise, inhibition of CXC receptor CXCR2 results in decreased pulmonary neutrophil recruitment, bacterial clearance, and survival (28, 49).
To combat this critical antimicrobial host defense system, extracellular bacterial pathogens possess virulence factors that not only protect against the bactericidal activity of neutrophils (7) but also have the capability to impede neutrophil pulmonary response. We have previously demonstrated that capsule and the O-antigen moiety of lipopolysaccharide (LPS) inhibit the recruitment of neutrophils into the pulmonary compartment after bacterial challenge (37). This is due, in part, to capsule and O-antigen decreasing pulmonary neutrophil chemotaxis mediated by both the formyl peptide and nonformyl peptide receptors (38).
Inducing neutrophil apoptosis or necrosis is another mechanism by which extracellular bacterial pathogens may subvert the host's pulmonary defense. Most neutrophils that migrate into the lungs will die there via apoptosis (11, 16). However, the time frame in which this occurs can be modified by both host and nonhost factors. As a general rule, proinflammatory mediators delay apoptosis, and, in contrast, downregulators (e.g., IL-10) and the process of phagocytosis accelerate apoptosis (22, 44). These findings can be interpreted as being beneficial to the host. If proinflammatory mediators increase functional neutrophil longevity, then their phagocytic potential will be maximized. Once phagocytosis is complete or downregulation of the inflammatory process has begun, neutrophils undergo elimination via apoptosis. The ability of E. coli to modulate this homeostatic process may enhance its pathogenic potential within the lung. The E. coli toxins alpha-hemolysin (Hly) and cytotoxic necrotizing factor (CNF)-1 are candidates for affecting neutrophil apoptosis or necrosis. Hly is a member of the repeats-in-toxin family of pore-forming toxins. E. coli Hly is toxic to a wide range of cells by inserting into the plasma membrane to generate a transmembrane pore. CNF is a member of a family of bacterial toxins that target the Rho family of small GTP-binding proteins in mammalian cells (6). Previous studies have demonstrated that these virulence factors can directly induce apoptosis or necrosis, pending concentration, in a variety of cell types (5, 12, 13, 15, 18, 30, 54).
In this study a wild-type extraintestinal pathogenic strain of E. coli (CP9) and isogenic derivatives deficient in Hly and CNF were assessed in vitro and in a rat model of gram-negative pneumonia to test the hypothesis that these virulence factors mediate neutrophil apoptosis and/or necrosis/lysis.
MATERIALS AND METHODS
Bacterial strains and media.
The model pathogen CP9 is an E. coli blood isolate cultured from a patient with sepsis and has been previously described in detail (19, 39). CP9 possesses many of the characteristics of typical ExPEC strains (40) and is highly virulent in a urinary tract infection model (36), an intraperitoneal infection model (42), and a pneumonia model (35, 43). CP9hlyA is a TnphoA′1-generated derivative of CP9 in which the structural gene for Hly (hlyA) is disrupted. CP9hlyA was confirmed to be nonhemolytic by a lack of hemolysis after growth on blood agar plates and by an absence of red blood lysis in a quantitative hemolysis assay (47). CP9hlyA/pEK50 has restoration of its Hly phenotype via complementation with the hlyA-containing plasmid pEK50 (21). CP9/pEK50, by virtue of possessing the wild-type Hly operon and the multicopy plasmid pEK50, produces supraphysiological amounts of Hly. CP9cnf1 is an isogenic derivative of CP9, in which the gene for CNF has been partially deleted as described previously (a gift from Dr. Hank Lockman) (34). Stability of the genotype/phenotype of CP9hlyA was confirmed by Hly production and of CP9cnf1 by PCR amplification of cnf. All strains were maintained at −80°C in 50% Luria-Bertani (LB) broth and 50% glycerol. For in vitro assays bacterial strains were grown in LB broth (5 g yeast extract, 10 g tryptone, 10 g NaCl/l). Incubations were performed at 37°C unless otherwise described.
Pulmonary infection model.
Animal studies were reviewed and approved by the University at Buffalo and Veterans Administration Institutional Animal Care Committee. An established rat (Long-Evans) model for studying pulmonary damage was used as reported (35, 43). In brief, Long-Evans rats (250–300 g) were anesthetized with ∼3% halothane in 100% oxygen until unconscious and then maintained at 1.5% halothane. The trachea was exposed surgically, and a 4-in piece of 1-0 silk was slipped under the trachea to facilitate instillation of the inoculum. The animals were suspended in a supine position on a 60° incline board. Pulmonary instillation of normal saline, LPS, or bacteria prepared in normal saline was introduced intratracheally (1.2 ml/kg) via a 1-ml syringe and 26-gauge needle, and the incision was closed with surgical staples. At harvest, halothane anesthesia was induced, an FiO2 of 98% was administered, and a midline incision was made through the peritoneum and thoracic cavity. In lung injury studies an arterial blood gas was obtained from the aorta. We flushed the pulmonary vasculature of residual blood by injecting the right ventricle with 10 ml of 5 mM EDTA/normal saline via a 26-gauge needle. For RNA abundance studies, whole lungs were harvested and kept at 4°C throughout. The lungs were subsequently weighed, suspended in normal saline to a total weight of 10 g (assumed to equate to 10 ml), and homogenized, on ice, for 3× 3 s using a Polytron PT-2000 homogenizer (Brinkman Instruments, Westbury, NY). Aliquots of the homogenate were removed for E. coli titer. Prior studies had demonstrated a low incidence of bacteremia (2.8%, 1:35), which was observed 15 h after bacterial challenge (37); therefore, blood cultures were not done in this study. For neutrophil morphology, LDH and caspase assays, and apoptosis studies, bronchoalveolar lavage (BAL) was performed with 50 ml of normal saline (37°C) into the lungs by gravity via the tracheal cannula. For lung injury studies BAL was performed with 15 ml of normal saline. The recovered BAL fluid was kept at 4°C before its subsequent use.
mRNA abundance assays.
Total RNA was extracted from whole lung homogenates by using a guanidinium isothiocyanate/chloroform-based technique followed by isopropanol precipitation. RNA concentration was determined by spectrophotometric measurement (A260/A280) and was checked for degradation by running on a formaldehyde/agarose RNA gel. 32P UTP-labeled mRNA antisense probes for rat apoptosis panels were synthesized using rat multiprobe template kits (Pharmingen, San Diego, CA). Samples (2 μg total RNA) were hybridized with the probe at 56°C for 12–16 h and then subjected to RNase (RNase A + T1) digestion (Pharmingen). Protected mRNA species were phenol-chloroform-isoamyl alcohol extracted and ethanol precipitated. The dried precipitate was dissolved in loading buffer and run on a 0.4-mm 5% acrylamide gel. The dried gel was loaded into a phosphorimaging cassette and exposed for 2–15 h, and the image was captured and analyzed using a Bio-Rad Molecular Imaging System (Hercules, CA). We identified specific mRNA species by comparing to migration distances of the unprotected probes, and the band densities were normalized to the housekeeping gene transcript L32. The mRNA species Fas antigen, bcl-x, Fas ligand, caspase 1–3, bax, and bcl-2 were quantitated. All reagents and equipment are carefully maintained as RNase-free (33).
“Naïve” neutrophils were purified from blood from healthy human volunteers using Polymorph prep (Axis-Shield; Poc, Oslo, Norway) per manufacturer instructions. After purification, neutrophil preparations consisted of 96–100% neutrophils and 0–4% lymphocytes.
Assessment of neutrophil morphology.
Human neutrophils, purified from blood, were used to assess cellular morphology after in vitro exposure to bacteria or normal saline. Rat neutrophils, recovered by BAL, were used to assess morphology after pulmonary instillation of normal saline, LPS, or bacteria. After 1 h of exposure at 37°C in vitro and 6 h in vivo, a 50-μl aliquot was diluted 1:200 in Isoton II solution (Beckman Coulter), and the leukocyte concentration was determined using a Multisizer 3 Coulter Counter (Beckman Coulter). We prepared a cytoslide by diluting cells to a final concentration of 5 × 104 leukocytes, using a Cytospin 3 cytocentrifuge (Shandon, Pittsburgh, PA), staining with Diff-Quik reagents (Baxter, Miami, FL), and examining under light microscopy (Nikon microscope ECLIPSE 80i; Nikon Instruments, Melville, NY). Photomicrographs were taken with a SPOT camera and SPOT INSIGHT software (version 4.0.4; Diagnostic Instruments, Sterling Heights, MI).
Caspase-3/7 activity was measured by conversion of the nonfluorescent caspase substrate Z-DEVD-R110 to the fluorescent rhodamine 110 as per the instructions (Apo-ONE homogeneous caspase-3/7 assay; Promega, Madison, WI). In vitro activity was measured in triplicate in a white, flat-bottom, 96-well Microplate (Thermo Labsystems, Franklin, MA) after 5 × 105 purified neutrophils were exposed to various E. coli strains at various titers or no bacteria (negative control) for 1 h at 37°C. In vivo activity was measured from cells obtained by BAL 6 h after pulmonary instillation of normal saline, LPS (670 μg), or E. coli [∼1 × 107 colony-forming units (cfu)]. Fluorescence was measured continuously by a SPECTRAmax microplate spectrofluorometer (Molecular Devices, Sunnyvale, CA) and data acquisition utilized SOFTmax PRO (version 3.1, Molecular Devices). Caspase activity was expressed as the Vmax (maximum slope on the linear portion of the curve) for a given bacterial or LPS concentration. Due to the fact that caspase-3/7 activity was compared between bacterial and nonbacterial groups (e.g., CP9 and LPS) and that caspase-3/7 activity may not have correlated directly with bacterial titers, caspase-3/7 activity was not normalized. Nonetheless, differences in cfus between E. coli groups were small. For in vitro studies, the mean differences in cfu between CP9 and its mutant derivatives CP9hlyA, CP9cnf1, and CP9hlyA/pEK50 were 0.26, 5.7, and 7.5%, respectively. For in vivo studies, the mean difference in cfu between CP9 and CP9hlyA was 1.4%. Results are the mean of two independent experiments.
LDH activity was measured via a coupled enzymatic assay, which results in the conversion of a tetrazolium salt into a red formazan product, which was detected colorimetrically (CytoTox96, Promega). In vitro activity was measured in triplicate in a 96-well microtest tissue culture plate (Becton Dickinson, Franklin Lakes, NJ) after 5 × 105 purified neutrophils were exposed to various E. coli strains at various titers for 1 h at 37°C. Neutrophils in plasma not exposed to bacteria, plasma alone, or plasma plus lysis buffer established baseline LDH activity. Neutrophils plus lysis buffer or LDH served as positive controls. In vivo activity was measured from BAL obtained 6 h after pulmonary instillation of normal saline, LPS (645 μg), or E. coli (CP9, CP9hlyA, ∼1 × 107 cfu). Absorbance at 490 nm was measured continuously by a SPECTRAmax microplate spectrofluorometer (Molecular Devices, Sunnyvale, CA) and data acquisition utilized SOFTmax PRO (version 3.1, Molecular Devices). LDH activity was expressed as the Vmax. Due to the fact that LDH activity was compared between bacterial and nonbacterial groups (e.g., CP9 and LPS) and that LDH activity may not have correlated directly with bacterial titers, LDH activity was not normalized. Nonetheless, differences in bacterial titers between E. coli groups were small. For in vitro studies, the mean differences in cfu between CP9 and its mutant derivatives CP9hlyA, CP9cnf1, and CP9hlyA/pEK50 were 0.58, 4.19, and 3.6%, respectively. For in vivo studies, the mean difference in cfu between CP9 and CP9hlyA was 1.4%. In vitro results were the mean of two independent experiments and in vivo results were the mean of 11 or 12 animals per group (four experiments with two or three animals per group).
Neutrophil lysis assay.
Neutrophil lysis was measured by a flow cytometry-based assay. Purified neutrophils (5 × 105) were exposed to various E. coli strains at various titers or no bacteria (negative control). After incubation for 1 h at 37°C, 1 × 105 sulforhodamine-impregnated polystyrene beads (6 μm; Polysciences, Warrington, PA) (FL-1) were added to the neutrophil suspension. Neutrophil lysis was determined by assessing the number of R-phycoerythrin (R-PE)-CD15-labeled neutrophils (FL-2) in the suspension using a FACSCalibur flow cytometer (Becton Dickinson Biosciences Immunocytometry Systems, San Jose, CA). Data acquisition for each sample was terminated after 2,000 bead events had been detected. This ensured that an equal volume of cell suspension was analyzed for each sample since each sample contained the same bead concentration. A gate in the forward-scatter vs. side-scatter plot discriminated neutrophils and beads from “debris” and a subsequent back-gating of the side-scatter vs. FL-2 discriminated the neutrophils from the beads and enabled neutrophil events to be counted.
Flow cytometry assessment of neutrophil apoptosis and necrosis/lysis in vitro.
Purified neutrophils (5.0 × 105) were resuspended in 100 μl of buffer (10 mM HEPES, 140 mM NaCl, 2.5 mM CaCl2; pH 7.4) and incubated with various strains of E. coli at various titers for 1 h at 37°C. Five microliters of annexin V-FITC (BD Biosciences, San Diego, CA) were added to the suspension and incubated at room temperature for 10 min, followed by the addition of 5 μl of 50 μg/ml 7-amino-actinomycin D (7-AAD, BD Biosciences) and incubation at room temperature for 10 min. The neutrophil population was identified by R-PE-conjugated CD15 (BD Biosciences) (FL-2). Annexin V-positive/7-AAD-negative cells were designated as early apoptotic, and annexin V- and 7-AAD-positive cells were designated as late apoptotic cells/early necrotic.
Lung injury studies.
For lung injury studies, rat lungs were instilled with either normal saline, LPS (670 μg), or ∼1 × 107 cfu of the E. coli strains CP9hly (produces no Hly), CP9 (produces physiologic levels of Hly), or CP9/pEK50 (produces supraphysiological levels of Hly). Six animals were present in each group. Harvest was performed 2 h postchallenge so that the potential confounding effect of differential growth or clearance of these strains would be minimized. The total bacterial titer, PaO2/FiO2, and BAL albumin were measured as described (20, 35).
Normally distributed data were analyzed by a two-tailed unpaired t-test. Data are presented as means ± SE, and P values <0.05 were considered statistically significant.
A wild-type, extraintestinal clinical isolate of E. coli (CP9) and derivatives deficient in their ability to produce Hly (CP9hlyA) and CNF (CP9cnf1) were used in a variety of in vitro and in vivo assays to determine the effects of the E. coli virulence factors Hly and CNF on neutrophil apoptosis and necrosis/lysis.
Effect of Hly, CNF, capsule, and O-antigen on total lung mRNA abundance of apoptotic genes in the rat pneumonia model.
To assess the effect of E. coli and its virulence factors Hly, CNF, capsule, and the O-specific antigen moiety of the LPS on the mRNA abundance of apoptosis genes (see materials and methods for list of genes assessed), we challenged Long-Evans rats intratracheally with ∼1 × 107 cfu of CP9 (wt), CP9hlyA (Hly-negative), CP9cnf1 (CNF-negative), CP9.137 (capsule-negative), and CP921 (O-antigen-negative). Six hours postchallenge, the absence of bacterial capsule or O-antigen had no effect on apoptosis genes at the level of mRNA abundance (Fig. 1). In contrast, the presence of CNF and to a lesser degree Hly diminished the abundance of Fas and bcl-x mRNA (CP9 compared with CP9cnf1 and CP9hlyA in Fig. 1). bcl-x is an oncogene that inhibits apoptosis (4) and has been shown to protect against apoptosis in both neutrophils and macrophages (31, 45, 53). Therefore, the repression of bcl-x by CNF1 and Hly is consistent with a proapoptotic role for these virulence traits. Although controversial, repression of Fas antigen by CNF1 and Hly is most consistent with an antiapoptotic effect (1). Therefore, our next study focused on the role of Hly and CNF on modulating apoptosis. Furthermore, since mRNA abundance studies utilized whole lung mRNA derived from multiple cells, an apoptotic effect of live CP9, CP9cnf1, and CP9hlyA on purified neutrophils in vitro was subsequently assessed.
Effect of Hly on neutrophil morphology in vitro.
We assessed the effect of Hly on neutrophil morphology in vitro by exposing purified neutrophils to 1 × 104–7 cfu of CP9 (wt) and CP9hlyA (Hly-negative) for 1 h and subsequent examination of Diff-Quik-stained neutrophils on a cytospin slide (Fig. 2). Neutrophils exposed to 104 cfu of CP9 demonstrated the same normal cellular morphology as neutrophils not exposed to bacteria (Fig. 2, A and B). Neutrophils exposed to 105 cfu of CP9 demonstrated early apoptotic changes, manifested by a ruffled membrane and clearly apoptotic cells with pyknotic nuclei (Fig. 2C). Neutrophils exposed to 106 cfu of CP9 demonstrated occasional apoptotic changes; however, the majority were undergoing various stages of necrosis (Fig. 2D). Greater than 90% of neutrophils exposed to 107 cfu of CP9 demonstrated advanced necrosis (Fig. 2E). In contrast, neutrophils exposed to 104 and 105 cfu of CP9hlyA demonstrated the same normal cellular morphology as neutrophils not exposed to bacteria (Fig. 2, F and G, respectively). Neutrophils exposed to 106 cfu of CP9hlyA demonstrated some cells with apoptotic changes (Fig. 2H). Neutrophils exposed to 107 cfu of CP9hlyA demonstrated a mixed morphology with the majority consisting of normal cells, but occasional apoptotic cells and cells in early stages of necrosis were also observed (Fig. 2I). These findings demonstrate that the effect of Hly on neutrophils is dependent on the bacterial concentration. At 104 cfu, no effect was observed, at 105 apoptotic changes were observed, and at 106 and 107 necrosis became the dominant feature. These data also demonstrate that although Hly was required for these changes, other bacterial factors are also most likely contributory as evidenced by the abnormal neutrophil morphology observed in some cells after exposure to 107 cfu of CP9hlyA (Fig. 2I). To further assess and quantitate the effects of Hly and CNF on neutrophil apoptosis and necrosis/lysis at various concentrations (given the subjectivity of denoting neutrophils as early apoptotic or early necrotic based on morphology), alternative assays were utilized.
Effect of Hly and CNF on neutrophil 3/7 caspase activity in vitro.
The effect of Hly and CNF on neutrophil 3/7 caspase activity, an alternative indicator of apoptosis, was assessed in vitro (Fig. 3). Purified neutrophils (5 × 105) were exposed to various titers of CP9 (wt), CP9cnf1 (CNF-negative), CP9hlyA (Hly-negative), and CP9hlyA/pEK50 (Hly-positive via complementation) or no bacteria for 1 h at 37°C. Caspase-3/7 activity was measured by the conversion of the nonfluorescent caspase substrate Z-DEVD-R110 to the fluorescent rhodamine 110. Neutrophils not exposed to E. coli demonstrated less caspase activity (P < 0.0001, unpaired t-test) compared with all bacterial titers assessed (1 × 104–7 cfu). Neutrophils exposed to 106 cfu of the Hly-producing strains CP9 and CP9hlyA/pEK50 demonstrated more caspase-3/7 activity compared with its Hly-negative derivative (CP9hlyA) at the same titer (P < 0.003, Fig. 3). Interestingly, neutrophils exposed to 106 of CP9 also demonstrated significantly more caspase-3/7 activity compared with its CNF-negative derivative (CP9cnf1) at the same titer (P < 0.005, Fig. 3). These data support the hypothesis that Hly and perhaps CNF are proapoptotic factors. However, the effects of Hly and CNF appear to be dose dependent, with maximal activity observed when neutrophils were exposed to 106 cfu. The decrease in caspase-3/7 activity correlated with our prior observation (see Effect of Hly on neutrophil morphology in vitro) that neutrophils exposed to 107 cfu of wild-type E. coli were in various stages of necrosis (not apoptosis). Therefore, experiments were performed to assess the effect of Hly and CNF on neutrophil necrosis/lysis.
Effect of Hly on neutrophil LDH activity in vitro.
Neutrophil morphology studies described above demonstrated that neutrophils underwent changes consistent with necrosis after exposure to 1 × 106–7 cfu of CP9 (wt) and perhaps after exposure to 107 cfu of CP9hlyA (Hly-negative). To both confirm and quantitate this observation, purified neutrophils were exposed to various titers of CP9 (wt), CP9hlyA (Hly-negative), CP9cnf1 (CNF-negative), and CP9hlyA/pEK50 for 1 h at 37°C, and LDH activity in the supernatant was measured (Fig. 4). Compared with neutrophils not exposed to bacteria and neutrophils exposed to CP9hlyA, a significant increase (P < 0.001, unpaired t-test) in LDH activity was detected after neutrophils were exposed to E. coli strains that express Hly (CP9, CP9cnf1) at titers of ≥107 cfu and at titers ≥106 cfu for CP9hlyA/pEK50, presumably due to the multicopy number of the complementing plasmid. These data demonstrate neutrophils exposed to titers of ≥107 cfu of CP9 strains that produce Hly at normal expression levels primarily undergo necrosis/lysis.
Effect of Hly on neutrophil lysis in vitro.
To confirm the effect of Hly on neutrophil necrosis/lysis, a flow cytometry-based assay was developed that measured neutrophil lysis (see materials and methods for details). Purified neutrophils were exposed to 1 × 107 cfu of CP9 (wt), CP9hlyA (Hly-negative), CP9cnf1 (CNF-negative), and CP9hlyA/pEK50 for 1 h at 37°C. Subsequently, after data acquisition for each sample was terminated following detection of 2,000 sulforhodamine-impregnated polystyrene beads (6 μm) events, neutrophil-gated events were measured (Fig. 5). Compared with neutrophils not exposed to bacteria and neutrophils exposed to CP9hlyA, a significant decrease (P < 0.0001, unpaired t-test) in neutrophil events was detected after neutrophils were exposed to E. coli strains that express Hly (CP9, CP9cnf1, CP9hlyA/pEK50). These data further demonstrate that at titers of ≥107 cfu of CP9 strains that produce Hly at normal expression levels neutrophils primarily undergo necrosis/lysis.
Annexin V/7-AAD staining was an unreliable assessment of apoptosis using live E. coli.
Annexin V and 7-AAD staining, combined with analysis via flow cytometry, were initially used to examine how Hly and CNF modified neutrophil apoptosis in vitro. However, these data demonstrated that annexin V/7-AAD staining was an unreliable assessment of apoptosis under these experimental conditions using live E. coli. Neutrophils identified as early apoptotic after exposure to 107 cfu of CP9 were in fact necrotic as was subsequently established by cellular morphology, caspase-3/7, and LDH activity (data not shown).
Effect of Hly on neutrophil caspase-3/7 and LDH activity and morphology in vivo.
Long-Evans rats were challenged intratracheally with purified LPS (670 μg) or a titer of CP9 (wt) or CP9hlyA (Hly-negative) that caused pneumonia (1–2 × 107 cfu). Six hours after challenge, cell-free LDH and cellular caspase-3/7 activity was assessed. The majority of cells (>95%) recovered from BAL were neutrophils. The cells recovered from CP9 and CP9hlyA-challenged animals had similar low levels of caspase activity (Fig. 6A, P = 0.5). There was a trend for even less caspase-3/7 activity after challenge with LPS compared with E. coli (P = 0.06, P = 0.2 for CP9 and CP9hlyA, respectively; Fig. 6A). In vivo caspase-3/7 activity (per 5 × 105 cells) was significantly less than the maximal in vitro caspase-3/7 activity (per 5 × 105 cells) detected after exposure to 1 × 106–7 cfu of CP9 (P < 0.005, Fig. 3).
In contrast, in vivo cell-free LDH activity was similar to the maximal in vitro LDH activity detected after neutrophils were exposed to ≥1 × 107 cfu of CP9 (Figs. 4 and 6B). Furthermore, significantly less LDH activity was measured in cell-free BAL from animals exposed to CP9hlyA compared with CP9 (P = 0.02).
Greater than 95% of BAL cells were neutrophils; therefore, BAL caspase-3/7, which was measured from BAL cells, reflected neutrophil activity. However, LDH activity was measured from BAL supernatant and may reflect necrosis/lysis of any cell or combination of cell types within the pulmonary compartment. To assess the effect of Hly on pulmonary neutrophils after E. coli challenge in vivo, examination of Diff-Quik-stained BAL neutrophils on a cytospin slide was performed. No morphologic changes were observed in neutrophils after challenge with normal saline (control), and neutrophils were readily distinguished from macrophages (Fig. 7A). Membrane ruffling but no apoptotic cells were observed after challenge with LPS (Fig. 7B). After challenge with CP9 (Hly-positive), a fraction of neutrophils demonstrated cellular changes consistent with necrosis (Fig. 7, C and D). After challenge with CP9hlyA (Hly-negative), occasional neutrophils demonstrated cellular changes consistent with an early stage of necrosis, albeit not as pronounced as what was observed after challenge with CP9 (wt) (Fig. 7, E and F). Many, but not all, of the necrotic neutrophils had phagocytosed bacteria (Fig. 7, C–F), thereby establishing their identity. Apoptotic changes were not observed in neutrophils after challenge with either CP9 or CP9hlyA. Together, these findings support that necrosis, not apoptosis, is the primary mechanism by which Hly affects neutrophils in our in vivo gram-negative pneumonia model.
Effect of Hly on lung injury in vivo.
The finding that Hly mediated neutrophil necrosis within the pulmonary compartment prompted an assessment of lung injury. Long-Evans rat lungs were instilled intratracheally with normal saline, purified LPS (670 μg), or ∼1 × 107 cfu of the E. coli strains CP9hly (produces no Hly), CP9 (produces physiological levels of Hly), or CP9/pEK50 (produces supraphysiological levels of Hly). Lung injury was assessed by measurement of oxygenation (PaO2/FiO2) and leakage of albumin into the pulmonary compartment (BAL albumin). Compared with the control animals (normal saline), BAL albumin was increased in animals after bacterial challenge. Furthermore, BAL albumin was significantly greater in animals challenged with the Hly-producing strains CP9 and CP9/pEK50 compared with the Hly-deficient strain CP9hly (Table 1). Likewise, oxygenation was significantly decreased in animals challenged with CP9 and CP9/pEK50 compared with CP9hly. The diminished oxygenation and increased albumin leakage in the presence of hemolysin were not due to an increased clearance of CP9hly compared with CP9 and CP9/pEK50 or differences in total BAL neutrophils (Table 1). Together, these findings demonstrate that Hly either directly or indirectly mediates lung injury.
ExPEC are a group of successful, highly evolved pathogens capable of causing serious infections in both healthy and immunocompromised hosts (41). Because ExPEC are extracellular pathogens, the ability to evade professional phagocytes, such as neutrophils, would greatly enhance their pathogenic capability. In this paper, we tested the hypothesis that the ExPEC toxins Hly and CNF1 modulate neutrophil apoptosis and necrosis/lysis. Our data demonstrate that in vitro Hly induces apoptosis at bacterial titers <106 cfu, but at higher titers, Hly causes necrosis/lysis (Figs. 2–5). When the bioactivity of Hly was assessed in an in vivo rat pneumonia model, Hly primarily caused neutrophil necrosis/lysis (Figs. 6 and 7). Likewise, lung injury was increased in animals challenged with E. coli strains that produced Hly compared with challenge with an isogenic Hly-deficient derivative (Table 1). Whether Hly-mediated lung injury is due to neutrophil necrosis, a direct effect of Hly, or both is unclear.
Our in vitro caspase-3/7 data also suggest that CNF-1 may induce neutrophil apoptosis (Fig. 3, Table 1); however, it is clear that CNF-1 does not mediate neutrophil necrosis/lysis (Figs. 4 and 5). The RNA abundance in vivo data from whole lungs is consistent with both CNF and Hly modulating apoptosis within the lung (Fig. 1). However, experiments on BAL (>95% neutrophils) did not disclose apoptosis in vivo, thereby suggesting that Hly and CNF may modulate apoptosis in nonneutrophil pulmonary cells. Other investigators, employing a rat Streptococcus sanguis pneumonia model, noted epithelial cells, but not infiltrating neutrophils, appeared to be undergoing apoptosis (26). In addition, results from our experiments suggest that ExPEC possesses additional bacterial virulence factors that cause neutrophil necrosis, since even in the absence of Hly, neutrophils were observed to undergo necrotic changes (Fig. 2I). Furthermore, results from in vivo lung injury experiments demonstrated that lung injury still occurs in the absence of Hly (Table 1). Whether this injury is mediated by unidentified bacterial factors or host factors remains undetermined.
An important aspect of the experimental design of these studies was the use of a clinically relevant extraintestinal pathogenic strain of E. coli (CP9) (not a commensal or intestinal pathogenic strain of E. coli) that possesses the full repertoire of virulence factors needed to be a successful extraintestinal pathogen and isogenic derivatives deficient in Hly (CP9hlyA) or CNF1 (CP9cnf1) (34, 39–41). Although a polar effect of the TnphoA′1 insertion in hlyA has not been excluded, complementation studies with pEK50 established Hly was responsible for the observed effects. Furthermore, live organisms were used for in vitro assays and in a relevant in vivo rat model of gram-negative pneumonia (35). The use of live E. coli guaranteed the appropriate level of expression for virulence factors, despite unpredictable and changing environmental signals in vivo. Moreover, the manner in which microbial virulence factors are presented to the host can differ when a purified product is used compared with a whole organism. Additionally, direct bacteria-host cell interactions may further modify host response, which would not be detected with the use of purified bacterial components. The importance of this concept was clearly observed when the effects of LPS compared with live E. coli on neutrophil death in vivo were assessed (Figs. 6 and 7). Another important technical note was the lack of reliability of the flow cytometry-based apoptosis assay. This assay was unable to reliably differentiate early apoptosis from necrosis. Whether this was due to the fact that live bacteria were being used is not known but underscores the importance of confirming results with alternative assays.
The findings from these experiments have important implications. They demonstrate that one of the mechanisms by which Hly contributes to ExPEC virulence is through increasing neutrophil death. Our data demonstrated that in vitro the effects of Hly are dose dependent, inducing apoptosis at bacterial titers <106 cfu, but at higher titers, causing necrosis/lysis. Our in vivo data suggest that, in at least the rat pneumonia model, which appears to be a reasonable approximation of the clinical scenario in patients, necrosis/lysis is the dominant effect of Hly on neutrophils. First, the challenge inoculum for in vivo experiments was ∼1 × 107 cfu. This is the minimal inoculum needed to reproducibly cause pneumonia (37) and is similar to that used in other pulmonary infection models (27, 32, 50). At the time of harvest, mean BAL E. coli titers were 1.5 × 106 ± 4.1 × 105 cfu/ml and total lung titers were 1.1 × 108 ± 4.9 × 107 cfu. In humans a quantitative BAL culture of ≥104–106/ml supports the diagnosis of ventilator-associated pneumonia (29, 56). In in vitro studies, neutrophils started to demonstrate necrotic changes with a bacterial titer 2 × 106/ml (Fig. 4); however, given the surface area of the lung, it is unclear how the measured in vivo titers equate to cfu per milliliter. In addition, we appreciate that it may be difficult to visualize apoptotic neutrophils ex vivo due to their rapid clearance by pulmonary macrophages. However, BAL cellular caspase-3/7 activity (>95% of BAL cells were neutrophils) was significantly less than the activity observed in in vitro studies that utilized the same number of cells (Figs. 3 and 6A). It is also possible that neutrophils are protected from apoptosis by the ensuing inflammatory milieu within the lung (3, 17). Together, these data support necrosis/lysis as the primary action of Hly on neutrophils in vivo. Whether Hly-mediated apoptosis occurs in other cells, at other sites of infection, or at lower titers in vivo awaits further study.
The biologic ramifications of Hly's mediating neutrophil necrosis/lysis are less clear, but the pathogenesis of pneumonia can be affected on at least two levels. First, Hly's mediating neutrophil necrosis/lysis would be predicted to affect phagocytosis and bactericidal activity. In support of that supposition are data from a previous study from our laboratory, which demonstrated that neutrophils contributed to the clearance of E. coli in the rat pneumonia model (43). In that study, neutrophil-depleted and neutrophil-replete animals were compared. Second, the fate of pulmonary neutrophils is not only important for host defense, but also for lung injury. Under normal physiological conditions, pulmonary neutrophils undergo apoptosis and clearance by macrophages (11, 16). However, a significant number of neutrophils undergoing necrosis/lysis may contribute to pulmonary injury via release of histotoxic granule contents (23). In addition, necrosis, as opposed to apoptosis, is proinflammatory. Data from this study demonstrate that Hly mediated lung injury (Table 1). However, whether this damage is direct or indirect and whether neutrophil necrosis is contributory are unclear. We have previously established that pulmonary damage correlated with bacterial titer but not pulmonary myeloperoxidase (43). This finding suggests, but does not establish, that acute lung injury in E. coli pneumonia is mediated by either bacterial or nonneutrophil host factors. A more complete understanding of the consequences of neutrophil necrosis due to Hly may lead to a decrease in acute lung injury and improved outcomes from gram-negative pneumonia.
Financial support was provided by a Department of Veterans Affairs Merit Review (T. A. Russo); National Institutes of Health Grants HL-69763 (T. A. Russo), HL-48889 and AI-46534 (P. R Knight); and The John R. Oishei Foundation (T. A. Russo).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 the American Physiological Society