Predatory Bacteria Attenuate Klebsiella pneumoniae Burden in Rat Lungs

ABSTRACT Bdellovibrio bacteriovorus and Micavibrio aeruginosavorus are predatory bacteria that naturally—and obligately—prey on other Gram-negative bacteria, and their use has been proposed as a potential new approach to control microbial infection. The ability of predatory bacteria to prey on Gram-negative human pathogens in vitro is well documented; however, the in vivo safety and efficacy of predatory bacteria have yet to be fully assessed. In this study, we examined whether predatory bacteria can reduce bacterial burden in the lungs in an in vivo mammalian system. Initial safety studies were performed by intranasal inoculation of rats with predatory bacteria. No adverse effects or lung pathology were observed in rats exposed to high concentrations of predatory bacteria at up to 10 days postinoculation. Enzyme-linked immunosorbent assay (ELISA) of the immune response revealed a slight increase in inflammatory cytokine levels at 1 h postinoculation that was not sustained by 48 h. Additionally, dissemination experiments showed that predators were efficiently cleared from the host by 10 days postinoculation. To measure the ability of predatory bacteria to reduce microbial burden in vivo, we introduced sublethal concentrations of Klebsiella pneumoniae into the lungs of rats via intranasal inoculation and followed with multiple doses of predatory bacteria over 24 h. Predatory bacteria were able to reduce K. pneumoniae bacterial burden, on average, by more than 3.0 log10 in the lungs of most rats as measured by CFU plating. The work presented here provides further support for the idea of developing predatory bacteria as a novel biocontrol agent.

, in vitro (11)(12)(13)(14). Predatory bacteria are unable to invade and prey on mammalian cells (15), mitigating potentially harmful offtarget effects. Additionally, predatory bacteria are nonpathogenic in a variety of animal models, including mice, rabbits, guinea pigs, and chicks (14,(16)(17)(18)(19). Furthermore, development of genetically stable resistance to predation has yet to be confirmed (20). However, most studies examining the ability of predatory bacteria to control human pathogens have been performed in vitro; thus, the efficacy with which predatory bacteria can control a bacterial infection in a live host is still not known.
In this study, first, the safety of intranasal inoculation of predatory bacteria in Sprague-Dawley (SD) rats was assessed. Then, lungs of rats were exposed to Klebsiella pneumoniae and were treated with multiple doses of predatory bacteria to determine their ability to reduce bacterial burden within the lungs. To our knowledge, this is the first study that demonstrated the ability of predatory bacteria to attenuate the bacterial burden of a key human pathogen in an in vivo mammalian system. The work presented here further supports the potential development of predatory bacteria into a biocontrol agent.

RESULTS
Host morbidity. While the safety of administering predatory bacteria into the lungs of mice has already been demonstrated (19), we began by investigating if the predators are compatible with the animal model being used in this study, Sprague-Dawley (SD) rats. To examine the effect on host morbidity of introducing predatory bacteria into the respiratory tract of SD rats, we performed intranasal inoculations of 6.0 ϫ 10 8 PFU/rat of B. bacteriovorus 109J, 1.1 ϫ 10 9 PFU/rat of B. bacteriovorus HD100, or 5.0 ϫ 10 7 PFU/rat of M. aeruginosavorus ARL-13 into rats in three groups containing six rats each. Another group of six rats was inoculated as a control with the vehicle, phosphate-buffered saline (PBS).
The animals were housed and monitored for up to 10 days postinoculation for any signs of discomfort, illness, or infection. At 10 days, all 18 rats inoculated with PBS or either strain of B. bacteriovorus were visually healthy with no apparent signs of illness (Table 1). One rat inoculated with M. aeruginosavorus died immediately after inoculation; however, veterinary consultation proposed asphyxiation as the probable cause of death. Blood analysis of the specific rat confirmed no elevated concentrations of white and red blood cell counts, and red blood cells were found with normal morphology (data not shown), suggesting that death was not caused by an acute infection resulting from M. aeruginosavorus inoculation. The remaining five rats inoculated with M. aeruginosavorus were found to be healthy at 10 days postinoculation (Table 1).
Histological examination of lung tissue from the 10-day exposure experiment revealed no pathology due to inoculation with predatory bacteria in any treatment group (Fig. 1). All specimens were characterized by lungs showing partial collapse of alveolar sacs without signs of inflammation or other abnormalities. Together, the data suggest that when inhaled, predatory bacteria have no adverse effects on rat morbidity.
Inflammatory response and histopathology. To examine the effect of predatory bacteria on the host immune response within the lungs of SD rats, we introduced PBS, 3.8 ϫ 10 8 PFU/rat of B. bacteriovorus 109J, 2.5 ϫ 10 8 PFU/rat of B. bacteriovorus HD100, or 2.0 ϫ 10 8 PFU/rat of M. aeruginosavorus ARL-13 into four groups of 36 rats each. An additional 28 rats were inoculated with a sublethal dose of 2.8 ϫ 10 7 CFU/rat of K. pneumoniae, a known respiratory pathogen. Animals were again visually monitored for any signs of illness for up to 48 h. All 108 animals inoculated with any strain of predatory bacteria were found to be healthy for the duration of the experiment (Table 1). Twelve animals from each group (eight for the K. pneumoniae group) were sacrificed at 1, 24, and 48 h postinoculation, and the lung samples were harvested for histological examination and to assess inflammatory cytokine levels.
All rats that were inoculated with predatory bacteria did not show any visual signs of illness or discomfort. Histological examination of lung tissue revealed no abnormal pathology due to B. bacteriovorus 109J or M. aeruginosavorus at 24 h postinoculation ( Fig. 1). At 48 h, examined lungs from rats inoculated with B. bacteriovorus 109J or M. aeruginosavorus showed only mild acute inflammation with stromal eosinophil infiltration and bronchioles containing a greater than normal amount of a proteinaceous substance consistent with mucus ( Fig. 1). Lungs inoculated with B. bacteriovorus HD100 exhibited reactive lymphoid hyperplasia and acute inflammation at 24 h postinoculation; however, no pathological abnormalities were seen at 48 h ( Fig. 1). Lungs inoculated with K. pneumoniae showed no abnormalities at    (Fig. 2B). No levels of other inflammatory cytokines were found to be increased more than 3.0-fold at any time point examined in any group treated with predatory bacteria (Fig. 2). In contrast, rats inoculated with K. pneumoniae still exhibited 14.4-, 11.0-, and 11.4-fold increases in the levels of IL-13, IL-1␤, and TNF-␣, respectively, at 24 h postinoculation ( Fig. 2B). At 48 h postinoculation, levels of IL-1␤ and TNF-␣ were still found to be substantially increased (3.6-and 6.3-fold, respectively) (Fig. 2C). Collectively, the data indicate that administering predatory bacteria at high concentrations into the lungs of rats does not result in adverse tissue pathology or provoke a sustained inflammatory response.
Predatory bacterial dissemination. To determine the predatory bacterial load within the lungs and other organs at 1, 24, and 48 h and 10 days post-intranasal inoculation, lung, liver, kidney, and spleen samples were also harvested during the previously described experiments and analyzed for the presence of predatory bacteria 16S rRNA using quantitative PCR (qPCR). Rats sacrificed at 1, 24, or 48 h were inoculated as described for the inflammatory response and histopathology experiment, while rats sacrificed at 10 days were treated as described above (see "Host morbidity").
Predatory bacteria 16S rRNA was detected in only a limited number of the kidneys, livers, and spleens of rats inoculated with either B. bacteriovorus HD100 or M. aeruginosavorus at 1 h postinoculation ( Fig. 3B to D). The kidneys, livers, and spleens of only six rats per treatment group at each time point were assessed. The trend exhibited a decrease in the concentration of predatory bacteria, as well as a decrease in the number of animals with predators that disseminated from the lungs to other organs examined. By 48 h postinoculation, B. bacteriovorus HD100 was detected in the kidneys of only 1/6 rats, in the livers of 0/6 rats, and the kidneys of only 2/6 rats (  continued to be detectable at similarly high levels at 48 h ( Fig. 3B to D). In conclusion, the data suggest that predatory bacterial dissemination to other organs after respiratory inoculation is limited, while predators are quickly and efficiently cleared from the host by 10 days postinoculation.
Pathogen inoculation and treatment. To determine whether predatory bacteria are able to reduce bacterial burden in the lungs, 36 rats were introduced with 3.3 ϫ 10 7 CFU/rat of K. pneumoniae via intranasal inoculation ("experimental group"). Thirty-six more rats were inoculated with PBS ("control group"). Twelve rats each from the control and experimental groups were dosed four times with PBS, 4.6 ϫ 10 8 PFU/rat of B. bacteriovorus 109J, or 6.6 ϫ 10 7 PFU/rat of M. aeruginosavorus ARL-13 over a 24-h period postinfection. Dosages of PBS or predatory bacteria were administered through intranasal inoculation at 30 min and 6, 12, and 18 h before animals were sacrificed and organs harvested at 24 h postinoculation.
As before, examination of lung tissue revealed no significant histological abnormalities due to predatory bacteria at 24 h postinoculation. Tissue from rats inoculated with K. pneumoniae and treated with either B. bacteriovorus 109J or M. aeruginosavorus was similar to the tissue from the PBS-inoculated control group and generally showed peribronchiolar aggregates of eosinophils and macrophages, though a few samples showed partial collapse of the alveolar sacs (Fig. 4). Rats inoculated with K. pneumoniae and not treated with predatory bacteria exhibited bronchiole-associated  lymphoid tissue with reactive lymphoid hyperplasia, as well as peribronchiolar infiltration of eosinophils (Fig. 4).
Lung samples were homogenized and plated on MacConkey agar. As expected, no colonies of K. pneumoniae from the lungs of rats in the PBS-inoculated control group were isolated on agar plates (Fig. 5). In the experimental group, a median of 1.6 ϫ 10 4 CFU/ml (mean ϭ 4.7 ϫ 10 5 CFU/ml) of K. pneumoniae was isolated from the lungs of 75.0% of rats initially inoculated with K. pneumoniae and treated with PBS (Fig. 5). In contrast, we recovered colonies of K. pneumoniae from the lungs of only 58.3% of rats inoculated with K. pneumoniae and treated with B. bacteriovorus 109J, with a median of 1.9 ϫ 10 2 CFU/ml (mean ϭ 1.8 ϫ 10 4 CFU/ml) (Fig. 5). K. pneumoniae was also recovered from the lungs of 66.6% of rats treated with M. aeruginosavorus, with a median of 1.2 ϫ 10 2 CFU/ml (mean ϭ 1.7 ϫ 10 5 CFU/ml). However, for both the B. bacteriovorus 109J and M. aeruginosavorus treatments, levels of K. pneumoniae CFU showed high variability. Strikingly, 83.3% of B. bacteriovorus 109J-treated rats and 66.6% of M. aeruginosavorus-treated rats exhibited greater than 3.0 log 10 reductions in copy numbers of K. pneumoniae recovered compared to the mean of the results from PBS-treated rats (Fig. 5). Furthermore, we recovered no K. pneumoniae from 41.6% of B. bacteriovorus 109J-treated rats and 33.3% of M. aeruginosavorus-treated rats. Taken together, the data indicate that B. bacteriovorus 109J and M. aeruginosavorus are able to efficiently reduce bacterial burden in the lungs of rats.

DISCUSSION
The prevalence of antibiotic-resistant infections has climbed to frightening levels over the last decade (2). Compounding this problem is the fact that since the development of linezolid in the 1970s, virtually no new class of antibiotics, particularly those that can target Gram-negative bacterial pathogens, has been discov-ered through traditional drug screening techniques. Recently, the first MDR infection caused by a member of the Enterobacteriaceae harboring the mcr-1 plasmid-borne colistin resistance gene was detected in the United States, potentially signaling the emergence of truly pan-drug-resistant bacteria (21). For these reasons, among others, it is crucial that we begin to develop new treatments to combat bacterial infection. Researchers have already begun investigating potential new antimicrobial strategies (22), such as the use of antimicrobial peptides (23,24), phage therapy (25,26), and gene-editing enzymes (27,28). Here, we report another promising novel approach: the use of predatory bacteria to control Gramnegative human pathogens.
In this study, we first determined the safety of intranasal inoculation of predatory bacteria in rats. High doses (Ͼ10 7 PFU/ml) of B. bacteriovorus 109J or HD100 or of M. aeruginosavorus ARL-13 were administered via intranasal inoculation into the lungs of SD rats. Two different B. bacteriovorus strains were examined to ensure that the results were not strain specific; unfortunately, we were unable to obtain additional M. aeruginosavorus strains. Of the total of 126 rats inoculated with predatory bacteria during the safety experiments, 125 rats were healthy, with no visual signs of adverse effects. Histological examination of tissue also revealed no adverse pathology associated with respiratory inoculation of predatory bacteria.
The lack of toxicity due to inoculation with predatory bacteria is in agreement with previous studies demonstrating the safety of introducing predatory bacteria in a variety of animal models (16)(17)(18)(19). In particular, we observed similar results in our previous study, in which C57BL/6 mice were intranasally inoculated with high concentrations of viable or heat-killed B. bacteriovorus or M. aeruginosavorus (19). In that study, no mouse exhibited either morbidity or any histological abnormalities at up to 50 days postinoculation, signaling that predatory bacteria are nontoxic when inhaled.
We next looked to determine the immune response to predatory bacteria within the lungs of rats. A slight increase in levels of inflammatory cytokines (IL-1␤, IL-6, IL-13, TNF-␣, KC/GRO) after inoculation with predatory bacteria was observed at the earlier time points of 1 and 24 h; however, inflammatory protein levels returned to baseline levels by 48 h postinoculation. The cytokine response in rats inoculated with the positive control, K. pneumoniae, was not as highly elevated as the response reported in other studies (29,30). An explanation for this is the fact we used a sublethal dose to ensure that the animals would not succumb to the infection before administration of predatory bacteria in the lung infection model. Furthermore, we demonstrated that predatory bacteria do not disseminate efficiently to other organs and are quickly cleared from the lungs of the host by 10 days postinoculation. As the aforementioned inflammatory cytokines are key players in the primary immune response, we suspect that the predatory bacteria are promptly and efficiently cleared through innate immunity mechanisms.
Our results seen with the rat model align with our previous study, where intranasal or intravenous inoculation of high doses of either B. bacteriovorus or M. aeruginosavorus in mice caused no tissue pathology and induced only a modest inflammatory response that returned to baseline levels within 24 h postinoculation (19). Furthermore, predatory bacteria were completely cleared from the animals by an innate immune response (possibly via neutrophils) within 48 h postinoculation (19). The consistency in our results provides further evidence that B. bacteriovorus and M. aeruginosavorus are inherently nonpathogenic in mammalian models. The lack of a productive and sustained immune response to predatory bacteria may be partially a result of the presence of an altered lipopolysaccharide (LPS) (31). The negatively charged phosphate residues of the LPS on the surface of Gram-negative bacteria are primarily responsible for antagonizing an immune response by strongly binding to Toll-like receptors on immune cells. B. bacteriovorus HD100 is known to express a neutrally charged LPS which has been shown to provoke only a weak inflammatory response in vitro (31), which may explain the lack of immunogenicity in our in vivo models. It is still unknown if M. aeruginosavorus expresses an altered LPS, as well.
The main objective of our study was to determine whether predatory bacteria are able to reduce the bacterial burden in the lungs in an in vivo mammalian system. We intranasally inoculated rats with K. pneumoniae and treated them with PBS, B. bacteriovorus 109J, or M. aeruginosavorus at 30 min and 6, 12, and 18 h postinoculation. In order to limit the number of animals being sacrificed, only B. bacteriovorus 109J was used in efficacy experiments. It was predicted that B. bacteriovorus 109J would provide a greater chance of reducing the bacterial burden than strain HD100 due to its weak innate immune response profile and its ability to remain longer in the lungs of the rats, as measured in dissemination experiments. Histological examination of lung tissue exhibited no significant pathological abnormalities due to treatment with predatory bacteria at 24 h postinoculation. Lung samples were homogenized and plated on MacConkey agar, a selective medium used to isolate Gram-negative and enteric bacteria, in order to determine K. pneumoniae concentrations (32). While the average reduction of K. pneumoniae bacterial burden due to treatment with B. bacteriovorus 109J or M. aeruginosavorus was approximately Յ1 log 10 , 83.3% of B. bacteriovorus 109J-treated rats and 66.6% of M. aeruginosavorus-treated rats exhibited greater than 3.0 log 10 reductions in the levels of K. pneumoniae recovered compared to the mean of the results seen with PBS-treated rats. Furthermore, 5/12 B. bacteriovorus 109J-treated rats and 4/12 of M. aeruginosavorus-treated rats had no detectable K. pneumoniae, suggesting clearance of the pathogen from the host. The observed amount of K. pneumoniae reduced by predatory bacteria in vivo is similar to the reported change in vitro. One such study demonstrated the ability of predatory bacteria to reduce levels of five different strains of K. pneumoniae (10). B. bacteriovorus 109J and M. aeruginosavorus were able to reduce K. pneumoniae levels in vitro by averages of 3.4 and 3.0 log 10 CFU/ml, respectively (10). Although active predation could be the sole contributor to the reduction of the K. pneumoniae load in the animals treated with predatory bacteria, one might suggest that additional immune response elements, elicited by the presence of the predatory bacteria, also played a role in reducing the microbial burden. Thus, the potential synergistic effects of predatory bacteria and the immune system should be a basis of future studies.
We are aware of only one other study that has tested the in vivo efficacy of treatments using predatory bacteria. That previous study determined whether oral administration of B. bacteriovorus could reduce the level of colonizing Salmonella in the guts of young chicks (18). The authors observed an average of 0.64 to 1.09 log 10 reduction of levels of Salmonella (18). While the mean reductions in the chick study were similar to the mean values that we obtained, a major difference in our study was that the majority of rats treated with predatory bacteria showed a greater than 3.0 log 10 reduction in K. pneumoniae levels. Note that, in the previous study, chicks were dosed once with predatory bacteria, whereas we dosed the rats four times over 24 h, modeling the dosing regimen in a typical antibiotic dosing schedule. Another major difference between the studies was in the physiological conditions of the animals, as birds have a physiological body temperature of 42°C. Predatory bacteria may prey upon the host bacteria and grow more efficiently in a mammal whose physiological temperature is 37°C, which is closer to the optimum predation conditions for predatory bacteria at 30°C. Furthermore, it is difficult for predatory bacteria to survive passage through the acidic environment of the stomach, and while an antacid was administered concurrently with predatory bacteria to increase the pH in the chick study, it is possible that the overall conditions of the chick gut are not optimal for predation.
It is important to emphasize the limitations of our model for extrapolation to the treatment of humans. We acknowledge that treatment with predatory bacteria at 30 min post-K. pneumoniae inoculation does not represent abrogation of an established infection. Rather, this study was an experimental exercise designed to determine whether predatory bacteria have the ability to reduce bacterial burden in an in vivo mammalian system. Our treatment scheme is similar to that of a recent study examining the efficacy of structurally nanoengineered antimicrobial peptide polymers (SNAPPs) in reducing MDR infections in mice (33). Mice were "infected" with a pathogen (Acinetobacter baumannii) and, similarly to our study method, were treated with SNAPPs at 30 min and 4 and 8 h postinoculation before being euthanized at 24 h.
In conclusion, our results indicate that predatory bacteria are safe to administer intranasally to a mammalian host, are able to attenuate pathogen burden in the lungs of rats, and may provide a novel way to combat infection caused by Gram-negative pathogens. Future studies will focus on using established in vivo models of infection to further determine whether the use of predatory bacteria is a viable treatment for Gram-negative infections, including MDR infections.

MATERIALS AND METHODS
Bacterial strains and growth conditions. Predatory bacteria examined in this study were Bdellovibrio bacteriovorus strain 109J (ATCC 43826), Bdellovibrio bacteriovorus strain HD100 (ATCC 15356) (34), and Micavibrio aeruginosavorus strain ARL-13 (9). As the pathogen, Klebsiella pneumoniae ATCC 43816 was used and grown in Luria-Bertani (LB) medium (35). Predatory bacteria were cultured and processed as previously described (11,15). Escherichia coli WM3064, a diaminipimelic acid (DAP) auxotroph, was used as prey and grown overnight in LB medium supplemented with 0.3 mM DAP. Predator stock lysates were made by coculturing the predators with prey cells in HEPES buffer (25 mM) supplemented with 3 mM MgCl 2 and 2 mM CaCl 2 . Cocultures were incubated at 30°C until the culture cleared (stock lysates). In order to obtain high concentrations of B. bacteriovorus for inoculation experiments, 10 ml of a washed overnight culture of E. coli WM3064 cells (~1 ϫ 10 9 CFU/ml) was resuspended in 80 ml of HEPES medium containing 10 ml of predatory bacteria from the stock lysates and incubated for 24 h on a rotary shaker. To obtain higher M. aeruginosavorus concentrations, M. aeruginosavorus cocultures were prepared in 200 ml of HEPES medium containing 25 ml of prey and 25 ml of M. aeruginosavorus stock lysates and incubated on a rotary shaker for 72 h. Cocultures were passed two times through a 0.45m-pore-size Millex filter (Millipore) to remove residual prey and cell debris (filtered lysate). Filtered lysates were pelleted three times by centrifugation at 29,000 ϫ g for 45 min using a Sorvall LYNX 4000 centrifuge (Thermo Fisher Scientific Inc.) to further purify and concentrate predator samples. Each time, the pellet was washed and resuspended in 50 ml of phosphate-buffered saline (PBS). For the final wash, the predator pellet was resuspended in 1 to 2 ml of PBS solution to reach final optical densities at 600 nm (OD 600 ) of 0.2 Ϯ 0.02 for B. bacteriovorus and 0.1 Ϯ 0.02 for M. aeruginosavorus, which corresponded to PFU values of between~5 ϫ 10 9 and 5 ϫ 10 10 PFU/ml and between~5 ϫ 10 8 and 5 ϫ 10 9 PFU/ml, respectively. The standard double-layered agar method was used to determine predator cell concentrations (36). Fifty microliters of the predator samples was plated on DAP-supplemented LB agar and tryptic soy broth (TSB)-blood plates to confirm that the samples were free of prey cells and contaminants. Since the predatory bacteria were used directly after isolation, the actual viable predator dose was known only a few days after each experiment, as the PFU appeared. Therefore, in some experiments, mainly involving M. aeruginosavorus, the inoculation sizes differed somewhat. The actual predator inoculation doses are specified for each experiment.
Rats. Wild-type male Sprague-Dawley (SD) rats (4 to 6 weeks old) were purchased from Charles River Laboratories (Wilmington, MA). All rats were housed under pathogen-free conditions at the Rutgers New Jersey Medical School animal facility. Guidelines from the Rutgers New Jersey Medical School Institutional Animal Care and Use Committee (protocol 15012) and the Animal Care and Use Review Office of the U.S. Army Medical Research and Material Command were followed in handling the animals.
Intranasal inoculation of bacteria. Predatory bacteria were introduced into the lungs of SD rats by intranasal inoculation to model a respiratory infection. After animals were anesthetized with 4% isoflurane-oxygen for 5 min using an isoflurane vaporizer, 50 l of purified bacterial suspension was applied at both nostrils using a pipette. Rats were inoculated with PBS, B. bacteriovorus strain 109J, B. bacteriovorus strain HD100, M. aeruginosavorus strain ARL-13, or K. pneumoniae. To avoid cross contamination, animals were caged according to treatment group and time point to be sacrificed. After every set of inoculations was performed in each experiment, animals were visually assessed for signs of infection, illness, and discomfort. At 1, 24, and 48 h and 10 days postinoculation, lung, liver, spleen, and kidney samples were harvested for histological examination, inflammatory protein analysis, and bacterial dissemination experiments.
Histological examination. Lung samples designated for histology were stored in formalin at 4°C before examination. All histopathological examinations were performed by a pathologist blind to each specimen's treatment group. Formalin-fixed organ segments from infected mice were paraffin embedded and stained with hematoxylin and eosin (H&E) for analysis of cellular composition as previously described. Stained sections were analyzed and photographed using an EVOS FL cell imaging system (Life Technologies, Carlsbad, CA).
Inflammatory protein analysis (ELISA). Lung samples were harvested in Lysing Matrix D tubes (MP Biomedicals) containing 1.0 ml of PBS with protease inhibitor. Samples were homogenized at 5.0 m/s for 30 s on a FastPrep-24 instrument (MP Biomedicals) before being stored at Ϫ80°C. At the time of analysis, homogenized tissues were thawed and centrifuged at Ͼ13,000 ϫ g relative centrifugal force (RCF) for 10 min at 4°C. The resulting supernatant was filtered through a 0.22 m-pore-size filter at 12 ϫ g RCF for 4 min. Inflammatory proteins were measured using a V-Plex proinflammatory Panel 2 (rat) kit (K15059D-1; Meso Scale Discovery) according to the manufacturer's instructions and read on a Sector Imager 2400 instrument (Meso Scale Discovery).
Nucleic acid extraction. Samples were prepared as previously described (37). Lung, liver, spleen, and kidney samples designated for nucleic acid extraction were harvested in Lysing Matrix D tubes containing 1.0 ml of TRIzol (Invitrogen). Samples were homogenized at 5.0 m/s for 30 s on a FastPrep-24 instrument before being stored at Ϫ80°C. Total RNA was extracted as previously described. Briefly, liquefied samples were spun down at Ͼ13,000 ϫ g RCF for 20 min at 4°C to remove tissue debris.
Two hundred microliters of chloroform was added to the supernatants, and the reaction mixture was centrifuged again at Ͼ13,000 ϫ g RCF for 15 min at 4°C. An equal volume of isopropanol was added to the aqueous phase, and then the reaction mixture was centrifuged at Ͼ13,000 ϫ g RCF for 15 min to pellet the precipitated RNA. After removal of the remaining isopropanol, pellets were washed twice with 500 ml of ice-cold 70% ethanol and then resuspended in 30 l of nuclease-free water. Total RNA was then purified using the "RNA Cleanup" protocol described in the instructions provided for the RNeasy minikit (Qiagen) and stored at Ϫ80°C.
Pathogen inoculation and treatment. Animals were anesthetized and inoculated intranasally with 50 l of K. pneumoniae as previously described. A 50-l volume of predatory bacteria was administered to rats at 30 min and 6, 12, and 18 h postinfection for a total of four doses over a 24-h period. At 24 h postinoculation, animals were euthanized and lung samples harvested for histological examination and CFU plating. Lung samples designated for CFU were stored in Lysing Matrix D tubes containing 1.0 ml of PBS and placed on ice. Samples were immediately homogenized at 6.0 m/s for 1 min on a FastPrep-24 instrument. Liquefied samples were then serially diluted and plated on MacConkey agar in order to determine K. pneumoniae concentrations. Histological examination was performed as described above.
Statistical analysis. ELISA data are presented as means Ϯ standard errors of the means; significant differences between the data from the treated samples and the data from the respective controls were determined using analysis of variance (ANOVA). K. pneumoniae reduction data are presented as medians; significant differences between treatment groups and respective controls were analyzed using the Mann-Whitney test. A P value of Ͻ0.05 was considered significant. Graphs were prepared and statistical analyses were performed using GraphPad Prism 6.05.