Pseudomonas aeruginosa OxyR Is Required for Full Virulence in Rodent and Insect Models of Infection and for Resistance to Human Neutrophils

The P. aeruginosa oxyR mutant is impaired in causing pneumonia and infecting burn wounds in mice.

P. aeruginosa wild-type strain PAO1, an isogenic oxyR mutant with and without a plasmid (pUCP19) control, and a complemented oxyR mutant (poxyR) were used in all experiments described herein. We first employed an acute pneumonia infection model of mice as described previously (11), with the exception that adult rather than infant mice were used. CD-1 mice were infected intranasally with ∼10⁷ bacteria each. Infected lungs were harvested at 16 h postinfection and homogenized, and serial dilutions were plated for the enumeration of CFU. When infected with wild-type P. aeruginosa, the mice were impaired in their ability to clear the organisms, and the CFU were enumerated at a log of ∼6.2 (Fig. 1). In contrast, the numbers of CFU of the oxyR mutants without and with the control plasmid pUCP19 were nearly ∼100-fold lower, at logs of 4.2 and 4.0, respectively (Fig. 1). Importantly, virulence was almost completely restored (∼5.6 log) in the oxyR mutant harboring poxyR.

• Open in new tab
• Download powerpoint

FIG. 1.

OxyR is critical for P. aeruginosa-mediated murine acute respiratory tract infection. CD-1 mice (groups of 10) were infected intranasally with ∼10⁷ CFU of the stationary-phase wild-type strain PAO1 or of the oxyR, oxyR (pUCP19), or oxyR (poxyR) strain. Infected mice were incubated for 16 h before their lungs were harvested and homogenized. Serial dilutions were plated onto LB agar-catalase plates, and CFU were enumerated. Attenuation is defined as the log₁₀ difference in numbers of CFU between wild-type and mutant bacteria recovered from lung tissues. The means ± standard errors of results from 10 mice are shown. *, P values were ≤5.3 × 10⁻⁵ (oxyR), 3.9 × 10⁻⁹ (pUCP19), and 0.0023 (poxyR) in Student t tests against PAO1.

We next tested the potential requirement for OxyR for the virulence of P. aeruginosa in a quantitative mouse burn infection model developed by Stevens et al. (10). In this model, a full-thickness burn wound is infected with P. aeruginosa by use of a slight modification of previous methods. First, a skin fold is elevated, and two preheated (92 to 95°C) brass blocks (2 by 2 cm) are applied for 5 s, thereby generating an injury with discrete margins that covers 5% of the total body surface area. With this model, the damaged epidermis and dermis undergo coagulative necrosis, but the underlying rectus abdominis (RA) muscles are not injured. Next, a P. aeruginosa suspension is injected intradermally into the midline crease of the burn eschar. The bacteria then proliferate in the burn wound tissue and invade the normal underlying RA muscles. The number of bacteria found in the RA muscles underlying and adjacent to the initial burn injury after 24 h serves as a sensitive and quantitative measure of local invasiveness. Some strains may also invade the vasculature and lymphatics, resulting in the widespread dissemination of bacteria throughout the infected animal. Hence, any of these organs are appropriate for documenting the systemic spread of P. aeruginosa from the dorsal burn wound. This progression of infection from a local soft-tissue infection caused by the low-dose bacterial inoculum to a systemic dissemination, leading to multisystem organ failure and eventual death, in many ways mimics the sequelae of lethal sepsis in burned or other severely immunocompromised patients.

After infection with ∼10⁷ bacteria, the titer of wild-type bacteria in the burn eschars increased by 1.6 log after 24 h (Fig. 2). In contrast, there was little or no increase in the CFU counts of oxyR and oxyR (pUCP19) strains (Fig. 2). Importantly, the virulence of the oxyR mutant harboring poxyR was partially restored, with the viable bacterial load increasing by 1.2 log (Fig. 2). We hypothesized that the levels of H₂O₂ in blood (low micromolar levels) (6) or the levels of H₂O₂ generated by macrophages of the reticuloendothelial cell system in burned mice will kill the bacteria during their systemic spread. Thus, we examined the abilities of the oxyR mutant to spread systemically to various organs, including the spleen, kidney, lung, and liver, and, in the process, to expose the organisms to blood H₂O₂. As shown in Fig. 2, the bacterial titers of the oxyR mutant and vector control mutants in various mouse organs were consistently 2 to 3 log lower than the titers of bacteria recovered from identical organs infected with wild-type bacteria. Specifically, the numbers of oxyR mutant bacteria recovered from burn eschar, spleen, kidney, lung, and liver were, respectively, 1.1, 1.9, 2.0, 2.5, and 2.3 log less than the number of bacteria of the wild-type strain, PAO1. Similarly, the numbers of viable oxyR mutant bacteria harboring pUCP19 that were recovered from burn eschar, spleen, kidney, lung, and liver were, respectively, 1.6, 2.4, 2.3, 3.0, and 2.8 log lower than the number of the wild-type bacteria, suggesting that the inclusion of plasmid pUCP19 in trans does not restore the virulence of the oxyR mutant. Significantly, as in the lung infection studies, virulence was partially restored in the oxyR mutant harboring poxyR, and the loads of the viable bacteria in various organs were only 0.5 to 1 log lower than the load of viable wild-type bacteria (Fig. 2).

• Open in new tab
• Download powerpoint

FIG. 2.

OxyR is critical for P. aeruginosa-mediated murine burn sepsis infection. Burned CD-1 mice were intradermally injected with 10⁷ CFU of stationary-phase wild-type bacteria or of oxyR, oxyR (pUCP19), or oxyR (poxyR) bacteria in the midline creases of their burn eschars. Infected mice were then incubated for 24 h before the burn eschars and various organs were harvested, ground, and plated onto LB agar-catalase plates, and the CFU were enumerated. Attenuation is defined as the log₁₀ difference in numbers of wild-type- and mutant-bacterial CFU recovered from lung tissues. The means ± standard errors of results from 10 mice are shown. For oxyR and oxyR (pUCP19) strains compared to the wild-type strain, all P values were ≤0.01.

Neutrophil killing of the oxyR mutant is enhanced relative to that of wild-type bacteria

One of the primary cell-mediated defenses against P. aeruginosa is the human neutrophil, especially in the context of cystic fibrosis airway disease. Therefore, we examined whether oxyR mutant bacteria were more sensitive to neutrophil-mediated killing than wild-type bacteria. Each of the four P. aeruginosa strains was incubated for 0 to 3 h in Dulbecco's minimal essential medium in the presence or absence of human neutrophils that had been isolated from the venous blood of normal subjects. At defined times, viable bacteria were measured as numbers of CFU from serial dilutions of each suspension on LB agar containing catalase. As shown in Fig. 3, the oxyR mutant was found to be more sensitive (about 33%) than wild-type bacteria to in vitro killing by human neutrophils. Complementation of poxyR, but not the control plasmid, pUCP19, restored wild-type levels of resistance to neutrophils.

• Open in new tab
• Download powerpoint

FIG. 3.

A P. aeruginosa oxyR mutant is more susceptible to killing by human neutrophils than PAO1. PAO1 and the oxyR, oxyR (pUCP19), and oxyR (poxyR) mutants were grown overnight on LB agar (PAO1) or LB agar-catalase (oxyR strains) plates with or without carbenicillin (200 μg/ml). The organisms were scraped from the plate, washed, and suspended in Dulbecco's minimal essential medium to a final concentration of approximately 10⁷/ml by themselves or in the presence of 2 × 10⁶ human neutrophils per ml; the neutrophils had been separated from the blood of normal subjects as previously described (1). After incubation on a shaker rack at 37°C for defined periods (0 to 3 h), 20-μl aliquots of each suspension were transferred to 1 ml of sterile 0.9% saline containing catalase (100 U/ml). After 5 min to allow for neutrophil lysis, serial dilutions were plated on LB agar-catalase plates in triplicate. After overnight growth, CFU were enumerated. The results were found to be most reproducible after a 2-h incubation. Shown are the mean percentages of organisms that were killed for each strain after a 2-h incubation with neutrophils ± standard errors of the means relative to the mean of values for a paired control where neutrophils were omitted (seven experiments). The magnitude of killing was significantly greater for the oxyR strain than for PAO1 (P < 0.04 [*]). In contrast, results for the complemented oxyR (poxyR) strain were not significantly different from those for PAO1 (P > 0.05), and the results for the oxyR (pUCP19) strain were similar to those for the oxyR mutant (P > 0.05).

The P. aeruginosa oxyR mutant is impaired in its ability to kill the insect D. melanogaster

D. melanogaster has recently been shown to be a valuable alternative to animal models for studies of P. aeruginosa pathogenesis (7). We compared the abilities of P. aeruginosa PAO1 and oxyR mutant bacteria to infect D. melanogaster Oregon-R flies by monitoring fly survival over time. Flies were maintained on a standard yeast-agar-sucrose-corn meal medium at 25°C. Experiments were performed using 4- to 7-day-old adult male flies. The infections were initiated by pricking flies on the dorsal thorax with a 10-μm-diameter needle (Ernest Fullam, Inc., Latham, N.Y.) that had been dipped into a cell suspension containing 10⁷ CFU of PAO1 or oxyR mutant bacteria from an early-stationary-phase bacterial culture/ml (7). Infected flies were kept at 25°C. Flies that died within 12 h after infection (<5%) were not included in mortality determinations. As a control, we established that pricking flies with a sterile needle dipped in a 10 mM concentration of MgSO₄ caused no death and that pricked flies quickly recovered from such trauma (Fig. 4).

• Open in new tab
• Download powerpoint

FIG. 4.

The oxyR mutant of P. aeruginosa is delayed in the kinetics of D. melanogaster killing. D. melanogaster Oregon-R flies were infected with PAO1 or the oxyR, oxyR (pUCP19), or oxyR (poxyR) strain. Control flies were pricked with a 10-μm-diameter needle dipped in a sterile 10 mM concentration of MgSO₄. Fly survival was monitored for 48 h. All infected flies received ∼100 bacteria/fly. Approximately 100 flies were used for each experiment. Three independent experiments yielded similar results.

P. aeruginosa PAO1 is highly lethal to D. melanogaster (7). Survival studies have shown that PAO1-infected flies began to perish at 18 h postinfection and that the infection resulted in the deaths of all flies by 28 h (Fig. 4). In contrast, the mortality of oxyR and oxyR (pUCP19) mutant-infected flies did not occur until 28 h postinfection. Furthermore, the mortality kinetics were consistently delayed for ∼6 to 10 h in oxyR and oxyR (pUCP19) mutant-infected flies (Fig. 4). Importantly, the virulence of the oxyR mutant complemented with poxyR was equivalent to that of the wild-type strain. Thus, the oxidative killing of the oxyR mutant within Drosophila hemolymph may involve mechanisms similar to those utilized by mammalian hosts.

In summary, the data obtained from the three models of infection indicate that OxyR is important for P. aeruginosa in its establishment of pulmonary infection and burn sepsis, is critical to its systemic spread and survival in the bloodstream, and is required for its virulence in the alternative-model host D. melanogaster. The loss of OxyR-mediated antioxidative mechanisms significantly compromises the organism's ability to survive the onslaught of oxygen radicals produced in infected tissues, blood, and human neutrophils. Based upon the findings of this study, we are currently conducting experiments to identify nontoxic reducing agents that compromise OxyR function, which might eventually lead to a therapeutic agent that would enhance the H₂O₂-mediated killing of P. aeruginosa during infection.