Pseudomonas expression of an oxygen sensing prolyl hydroxylase homologue regulates neutrophil host responses in vitro and in vivo

Background: Pseudomonas species are adapted to evade innate immune responses and can persist at sites of relative tissue hypoxia, including the mucus-plugged airways of patients with cystic fibrosis and bronchiectasis. The ability of these bacteria to directly sense and respond to changes in local oxygen availability is in part consequent upon expression of the 2-oxoglutarate oxygenase, Pseudomonas prolyl hydroxylase (PPHD), which acts on elongation factor Tu (EF-Tu), and is homologous with the human hypoxia inducible factor (HIF) prolyl hydroxylases. We report that PPHD expression regulates the neutrophil response to acute pseudomonal infection. Methods: In vitro co-culture experiments were performed with human neutrophils and PPHD-deficient and wild-type bacteria and supernatants, with viable neutrophil counts determined by flow cytometry. In vivo consequences of infection with PPHD deficient P. aeruginosa were determined in an acute pneumonia mouse model following intra-tracheal challenge. Results: Supernatants of PPHD-deficient bacterial cultures contained higher concentrations of the phenazine exotoxin pyocyanin and induced greater acceleration of neutrophil apoptosis than wild-type PAO1 supernatants in vitro. In vivo infection with PPHD mutants compared to wild-type PAO1 controls resulted in increased levels of neutrophil apoptosis and impaired control of infection, with higher numbers of P. aeruginosa recovered from the lungs of mice infected with the PPHD-deficient strain. This resulted in an overall increase in mortality in mice infected with the PPHD-deficient strain. Conclusions: Our data show that Pseudomonas expression of its prolyl hydroxylase influences the outcome of host-pathogen interactions in vitro and in vivo, demonstrating the importance of considering how both host and pathogen adaptations to hypoxia together define outcomes of infection. Given that inhibitors for the HIF prolyl hydroxylases are in late stage trials for the treatment of anaemia and that the active sites of PPHD and human HIF prolyl hydroxylases are closely related, the results are of current clinical interest.


Introduction
Pseudomonas aeruginosa is an opportunistic pathogen, which colonizes the airways of patients with chronic inflammatory lung diseases including cystic fibrosis (CF) and bronchiectasis 1,2 , and is an important pathogen in the setting of acute ventilation -associated pneumonia 3 . In the 2004 US CF patient registry, 57% of patients were found to be colonized with P. aeruginosa 4 , whilst children with CF who have sputum positive for P. aeruginosa experience more frequent hospitalisation and higher mortality 5 . These patients have chronic sputum production, with areas of mucus 'plugging', resulting in local hypoxia, a condition in which bacteria thrive 6 . Despite high levels of neutrophilic inflammation, P. aeruginosa continues to survive in these patients, evidence that the bacteria employ a number of effective immune-evasion strategies.
A key mechanism by which P. aeruginosa impairs host neutrophil function is by generation of phenazine metabolites, particularly pyocyanin, which contributes to the characteristic blue-green colour of infected sputum 7 . Pyocyanin has previously been shown to accelerate neutrophil apoptosis through activation of the lysosomal death pathway, a process dependent upon the generation of reactive oxygen intermediaries within the neutrophil 8-10 and thus on the availability of molecular oxygen 11 . More recently, the possibility Pseudomonas itself can directly sense and respond to changes in local oxygen availability was raised by the observation that Pseudomonas spp contain a 2-oxoglutarate (2OG)-dependent Pseudomonas prolyl hydroxylase (PPHD), which acts on the abundant translation elongation factor Tu (EF-Tu) and is homologous to the oxygen sensing hypoxia inducible transcription factor (HIF) prolyl hydroxylase (PHD) enzymes described in eukaryotes 12 .
Importantly, an insertional mutant strain of P. aeruginosa lacking PPHD manifests increased production of pyocyanin under normoxic (room oxygen) standardized broth culture conditions. Moreover, growth of P. aeruginosa under conditions of hypoxia has been observed to reduce the pathogenicity of P. aeruginosa through repression of production of the siderophores pyoverdine and pyochelin and the secreted virulence factor Exotoxin A 13,14 . Thus, the outcomes of host-pathogen interactions may be in part defined by adaptation of both the host and the pathogen to local oxygen availability. In this context, we hypothesised that (1) a PPHD-deficient P. aeruginosa sp might demonstrate a survival advantage in vivo as a consequence of increased pyocyanin production, leading to accelerated neutrophil apoptosis and impaired neutrophil mediated bacterial killing, and (2) that these effects would be influenced by oxygen availability.

Ethical approval
All participants gave written informed consent in accordance with the Declaration of Helsinki principles, with AMREC approval for the study of healthy human volunteers through the MRC/University of Edinburgh Centre for Inflammation Research blood resource (15-HV-013). Human peripheral blood neutrophils were isolated from whole blood using dextran sedimentation and discontinuous Percoll gradients 15 .

Bacterial growth curves
A Columbia blood agar culture plate (VWR International, UK) was inoculated with a single bead from a thawed master stock vial of either wild-type (PA01) or PA0310 insertional knockout mutant strain (PPHD knockout) pseudomonas and then incubated overnight at 37°C. The following day, ten colonies were taken from the plate using a sterile inoculating loop and used to inoculate 15 ml of sterile Luria-Bertani (LB) broth (Sigma, UK) in a 50ml Falcon tube. The tube was then incubated at 37°C on a shaking platform with the lid loosened. Optical density at 595 nm was measured regularly until plateau.

Intratracheal pneumonia model
All animal experiments were conducted under an Home Office approved project license in accordance with the Home Office Animals (Scientific Procedures) Act 1986 and University of Edinburgh guidelines in line with the NC3Rs. Six to eight week male C57Bl6J mice were group-housed under standard 12hr light/dark cycles with access to food and water ad librium. All efforts were made to ameliorate any suffering of the animals. Mice were closely monitored over the course of the experiments and humanely culled once threshold of severity was reached.
Mice were anaesthetised with ketamine (76mg/kg, Willows Francis Veterinary, UK) and medetomidine (1mg/kg, Orion Pharma, UK) intraperitoneally. Once adequately anaesthetised, the animals were suspended from a frame by the upper incisors and a blunt needle was passed into the trachea via the orotracheal route. Each mouse then had 1×10 7 cfu of either PA01 (wild-type) or PPHD knockout out (mutant) pseudomonas instilled in 50μl PBS via the endotracheal. Twenty minutes after anaesthesia, the mice were given atipamezole (2mg/kg, Orion Pharma, UK), an anaesthetic reversal agent, and recovered for six hours. At indicated time points (6, 12, 24, 36 and 48h after instillation) mice were assessed and tissues harvested. For the Kaplan-Meier plots, mice were culled once the threshold of sickness was reached.

Assessment of lung injury
Bronchoalveolar lavage (BAL) was obtained by cannulation of the trachea. Total cell counts were calculated using haemocytometer counts and differential cell counts assessed on cytocentrifugation slides. IgM levels were quantified using commercially available kits (Mouse IgM ELISA quantitation set, Bethyl Laboratories Inc, Montgomery, USA; EnzChek Elastase Assay Kit, Molecular Probes Europe BV, Leiden, The Netherlands).
For histological analysis, lungs were fixed with 10% buffered formalin and embedded into paraffin blocks. Tissues slices were fixed and stained with haematoxylin and eosin.
Flow cytometry for BAL neutrophil apoptosis BAL cells were counted and 1×10 6 cells were centrifuged at 300g for 5 minutes at 4°C. Cell pellets were resuspended in 50uL of FC block (1:100 anti-CD16/32 Ab, RRID:AB_312801; Biolegend) and 1:10 mouse serum in FACS buffer (PBS with 0.5% BSA and 0.02mM EDTA) and incubated on ice for 15 minutes. Subsequently, cells were stained with 50ul anti-Ly6G Ab (RRID:AB_ 1236494; BioLegend) at 1:200 final concentration and incubated on ice for 30 minutes in the dark. Following a wash with FACS buffer and centrifugation at 300g for 5 minutes at 4°C, cells pellets were resuspended in Annexin-binding buffer and Annexin-V PE stain (Becton Dickinson) for 15 minutes at room temperature in the dark. Prior to flow cytometry acquisition, cells were stained with Topro3 APC (Molecular Probes). Neutrophils were gated based on Ly6G expression and Annexin-V and Topro3 expression was quantified.
Cells were acquired using a BD Calibur machine and analysed using FlowJo version 10 software (Tree Star).
Quantification of viable bacterial counts 10-fold serial dilutions were performed on whole blood aliquots and lungs homogenized in sterile tubes following collection of BAL fluid. Three 10μl drops from each of 6 dilutions were then plated onto blood agar plates and cultured overnight in 37°C to calculate viable bacterial counts, which were normalized to count per ml of blood or per pair of lungs.

Production of bacterial supernatants
Ten colonies were taken from the blood agar culture plate and used to inoculate plates containing 20ml of pseudomonas isolation agar (Difco). These plates were incubated overnight at 37°C and then placed in direct sunlight for 48h to allow pigment to develop. Each plate was then flooded with 6ml RPMI media (Sigma, UK) and left at room temperature for 2 hours. The RPMI was removed, spun at 4000g for 15 min, twice, and filter sterilised through a 0.22μm filter to remove any bacteria. To ensure sterility, 100μl of each supernatant was used to inoculate a blood agar plate and cultured for 48h at 37°C. Supernatants were stored at -80°C.
Quantification of pyocyanin concentration PPHD mutant and wildtype colonies were inoculated into 10 ml of LB broth and incubated overnight at 37°C in a shaking incubator. 1ml of the overnight cultures were then inoculated into 9ml of LB broth and incubated for 2 hours at 37°C in a shaking incubator. 100 μl of each strain was then pipetted onto Pseudomonas Isolation agar (Difco) plates and incubated overnight under conditions of normoxia (21% O 2 ) and hypoxia (3% O 2 ) and supernatants produced as detailed above. 4.5ml of chloroform was added to 7.5ml of sterile bacterial supernatant and vortexed. Samples were centrifuged at 2000g for 10 minutes. 3ml of the chloroform layer was transferred to a clean tube and 1.5ml 0.2M hydrochloric acid was added. Tubes were vortexed and spun at 2000g for 2 minutes. 1ml of the top layer was removed, absorbance at 520nm measured and pyocyanin concentrations determined 16 .
Isolation and culture of human neutrophils Human peripheral blood neutrophils were isolated from whole blood using dextran sedimentation and discontinuous Percoll gradients. Neutrophils were resuspended in RPMI with 20% fetal calf serum (Lifetech, Paisley, UK) at 10x10 6 /ml. 75μl of this suspension was cultured with 75μl of either wild-type (PA01) or mutant (PPHD knockout) pseudomonas supernatant for five hours in either normoxia (room air) or hypoxia (1% oxygen, in vivo 400 hypoxia workstation, Ruskinn). After 5 hours, cells were removed from the culture plate and pelleted at 400g for 5 minutes. The pellets were resuspended in 95μl annexin binding buffer and 5μl annexin V/PE (Becton-Dickinson) and incubated on ice for 20 minutes. 100μl of Topro3/APC (Molecular Probes) and 5×10 4 Countbright™ absolute counting beads (ThermoFisher, UK) were added to each sample, and samples run using a BD FACSCalibur (BD Biosciences, UK).

Statistical analysis
Data were analysed using Prism 7.0 software (GraphPad Software Inc., San Diego, CA). Unpaired t-tests were used for comparisons between wild-type and knockout sample means. Two-way ANOVA with Bonferroni's post-test comparisons was performed if multiple time points were used. For comparison of viable bacterial counts, Mann-Whitney test was performed. Survival was analysed using log-rank test. Statistical significance was accepted when p<0.05.

Results
Neutrophil co-culture with PPHD mutant bacterial supernatants induced cell loss, which was reversed with hypoxic culture To directly address whether expression of the hypoxia sensing prolyl hydroxylase PPHD by P. aeruginosa would affect rates of neutrophil apoptosis, freshly isolated human peripheral blood neutrophils were cultured for 5h with sterile supernatants harvested from wild type PA01 and PPHD-deficient bacterial cultures in vitro, a time-point at which pyocyanin markedly accelerates neutrophil apoptosis 8 . Total neutrophil numbers and neutrophil viability were assessed by flow cytometry ( Figure 1A). In normoxia, both PA01 and PPHD-deficient supernatants caused significant loss of neutrophil numbers ( Figure 1B) and a reduction in Annexin V-/Topro 3-(viable) neutrophils recovered. Greater reductions in viable cell numbers (Annexin V-/Topro 3-) were observed when neutrophils were cultured in the presence of PPHD-deficient supernatants ( Figure 1C). Hypoxic cell culture reversed the increases in both cell loss and apoptosis observed with PAO1 and PPHD-deficient supernatants ( Figure 1A-C), in keeping with the dependence of pyocyanin-induced apoptosis on the availability of oxygen 11 .
To address whether the observed differences in neutrophil loss reflected altered pyocyanin production, we measured pyocyanin production over a 48 hour inoculation of blood agar following recovery into RPMI media. Under conditions of normoxia (21% O 2 ), the PPHD-deficient strain produced significantly higher levels of pyocyanin than the wild-type strain ( Figure 1D), in keeping with the greater loss of viable neutrophil numbers ( Figure 1C). Elevated pyocyanin production by the PPHD-deficient strain was abrogated entirely by use of a hypoxic (1% O 2 ) cell culture ( Figure 1D). This was not a consequence of differential bacterial growth rates, with equivalent 595 nm absorbance and bacterial counts being observed for PA01 and PPHD mutants at each oxygen tension studied ( Figure 1E, F). Hypoxia (1% O 2 ), whilst impairing bacterial growth, had no differential growth effects on PAO1 compared with the PPHD-deficient strain.
PPHD-deficient P. aeruginosa infection results in increased mortality and lung injury during acute pneumonia and greater impairment of neutrophil-mediated host defense compared with wild-type PAO1 infection To define whether PPHD deficiency results in an altered course of acute P. aeruginosa infection in vivo, mice were challenged via the trachea with 1×10 7 cfu PAO1 (wildtype) or PPHDdeficient bacteria. 50% of animals receiving PPHD-deficient P. aeruginosa reached sickness thresholds requiring the animals to be culled by day 5 (Figure 2A). In contrast, all PAO1 infected mice were viable up to 5 days following infection challenge (Figure 2A, *p<0.05). Importantly, this increase in mortality was associated with a 2.5-fold greater bacterial burden in the lungs of PPHD mutant infected (11.1×10 4 CFU/lung± 5.99×10 4 ) compared to wildtype infected mice (4.48×10 4 CFU/lung ± 3.48×10 4 , Figure 2B, p<0.05).
In light of the equivalent growth of wild-type and PPHD mutant strains we had observed in vitro, we questioned whether the increase in bacterial numbers in PPHD mutants was a consequence of an impaired host response. Whilst the initial recruitment of inflammatory cells to the lungs (6h and 12h) was similar between wildtype and PPHD-deficient P. aeruginosa infected mice ( Figure 2C-F), significantly fewer cells were recovered from the airways of PPHD-deficient infected mice by 24h after infection ( Figure 2C, **p<0.01), as a consequence of reductions in both the percentage ( Figure 2D, *p<0.05) and total number of airway neutrophils ( Figure 2E and F, **** p<0.0001). In keeping with a more severe infection in PPHD-deficient infected mice, higher levels of IgM, an indirect marker of vascular leak and lung injury, were detected in mice infected with mutant PPHD ( Figure 2G, *p<0.05). Histological analysis of lungs of mice infected with P. aeruginosa supported the observed differences in BAL, with fewer neutrophils in the lungs of mice infected with mutant PPHD ( Figure 2H).
In light of the increased production of pyocyanin by PPHD mutant P. aeruginosa and the observed increase in neutrophil loss with PPHD supernatants, we hypothesised the reduction in neutrophil numbers observed at 24 hours to be a consequence of increased levels of neutrophil apoptosis. Ly6G+ airway recovered neutrophils were therefore dual stained with Annexin V/Topro3 to directly quantify the number of apoptotic cells following Pseudomonas infection. Infection with mutant PPHD P. aeruginosa resulted in higher detectable levels of apoptosis than infection with the PAO1 wild type strain ( Figure 2I, p<0.01).

Discussion
A significant focus of research from our group and others has centered round defining the mechanisms by which hypoxia directly regulates immune cell function [17][18][19][20][21][22] . Innate responses to bacterial challenges are critically regulated by oxygen availability, with neutrophils in particular being adapted to survive in hypoxic tissues where they phagocytose and kill bacteria 11,21,23 . Until recently, the possibility that oxygen may also regulate the behavior of bacterial pathogens has not been considered. This is of particular relevance to Pseudomonas spp that persist in chronically inflamed tissues characterized by limited oxygen availability 24 and induce oxidant-dependent cell death via the production of the phenazine, pyocyanin 10,25 . The results described here reveal the importance of the Pseudomonas prolyl hydroxylase, PPHD, in regulating the effectiveness of neutrophil mediated host defenses in vivo, likely, at least in part, as mediated by variations in the levels of the toxic metabolite pyocyanin.
Suppression of PHD activity is described in eukaryotic systems in the context of hypoxia -indeed is central to regulation of the hypoxic response 26-30 . Diminished Pseudomonas aeruginosa pathogenicity in hypoxia has recently been described as a consequence of reduced expression of the virulence factors pyoverdine and exotoxin A 14 , with our work extending this to include production of pyocyanin. This is of interest, given that neutrophil respiratory burst activity, a key anti-microbial defence, is associated with promotion of a hypoxic niche 22 . It is also of relevance to the oxygen requiring process by which the pyocyanin induces ROI-mediated lysosomal dysfunction and neutrophil apoptosis 10 , as evidenced by the reduction in cell loss we observed when neutrophils were challenged with P. aeruginosa conditioned media in the context of hypoxia. Thus in clinical scenarios in which tissue oxygen availability is severely limited, oxygen dependent regulation of the balance between innate immune responses and bacterial virulence and replicative capacity may be critical in defining the outcomes of infection and host morbidity and mortality. This is, however, further complicated by the observations that differential expression of oxygen sensing prolyl hydroxylase enzymes either by immune cells 20,31,32 , or bacterial pathogens 12 can also directly regulate cellular function and bacterial virulence when oxygen is not a limiting factor. For example, neutrophil loss of PHD2 under physiological normoxia promotes a phenotype of excessive neutrophilic inflammation 33 , whilst deletion of PPHD is associated with increased production of pyocyanin 12 . In vivo therefore, the dominant phenotype is likely to be in part determined by the physiological Κ m , in which both PPHDs and PHDs function in both innate immune and bacterial cells. Of interest, kinetic analysis of the isolated PPHD enzyme has identified a lower apparent Κ m for O 2 than PHD2, but a higher Κ m for Fe(II), suggesting that iron regulation may also be of critical importance in defining the activity of PPHD enzyme activity in a physiological setting 12 . This is of particularly relevance to Pseudomonas spp where enhanced iron redox states enable competitive outgrowth from other bacterial species 34 .
In this work, we provide in vivo data, describing the clinical outcomes when mice are challenged with acute P. aeruginosa infection in the context of normal lung architecture and therefore relatively preserved local tissue oxygenation. In this setting we observe increased mortality with PPHD-deficient strains as a consequence of insufficient neutrophil host defense and failure to control bacterial replication. Thus, we can now extend the concept that immune cell loss of PHD2 in the context of preserved tissue oxygenation promotes a detrimental immune response to also include detrimental consequences of prokaryotic loss of PPHD expression. This has potentially important ramifications in light of the current development of relatively non-selective PHD inhibitors (which may well inhibit PPHD), as well as the use of iron chelators in the clinical arena, and how they may impact more widely on the host pathogen response with consequence both for the host and the pathogen.

Data availability
Raw data counts available In this manuscript, Dickinson and colleagues reveal the impact of P. aeruginosa PPHD on the bacteria virulence using in vitro and in vivo models. In particular, the PPHD mutation-dependent increased concentration of pyocyanin in bacterial supernatants was characterized and associated with increased neutrophil apoptosis. The manuscript is clear, well written and the presented data support the conclusions.
In order to better appreciate the contribution of pyocyanin on P. aeruginosa mediated neutrophil apoptosis, the following points should be addressed: The authors studied the impact of pyocyanin on neutrophil survival using bacterial supernatants form WT and PPHD-deficient mutant. In order to evaluate the contribution of pyocyanin on P. aeruginosa-dependent neutrophil apoptosis induction, the authors should additionally infect neutrophils with WT and PPHD-deficient mutant strains and assess neutrophil viability (w/wo oxygen). The comparison with neutrophil apoptosis levels induced with bacterial supernatant should be then discussed.
It would be informative to assess the ability of neutrophils to kill WT and MT strains (w/wo oxygen) and the results should be discussed to better interpret in vivo results.
In the discussion, the authors say that "the possibility that oxygen may also regulate the behavior of bacterial pathogens has not been considered". This statement is not true, additional references should be included regarding the O2-modulation of bacteria virulence, adhesion, secretion, etc.

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? Yes
No competing interests were disclosed. Competing Interests: 1.

I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.
Author Response 22 Nov 2017 , University of Edinburgh, UK

Sarah Walmsley
We thank Dr Marteyn for the review of our manuscript.
In vitro co-culture of neutrophils directly with pseudomonas species rapidly results in neutrophil loss. This is why we firstly explored the consequence of neutrophil culture with bacterial supernatants before undertaking in vivo experiments in which we were able to directly address the consequence of infection with the different strains of P. aeruginosa on neutrophil survival in a biological setting.
Due to the ability of pseudomonas to induce neutrophil apoptosis, this is a difficult question to directly address. We would however argue that it is the ability of Pseudomonas to evade the host response that is critical in defining the outcome of the infection challenge We are sorry for any confusion caused, in the discussion we do actually state that "until recently, the possibility that oxygen may also regulate the bacterial pathogens has not been considered" and provide a number of references specific to Pseudomonas throughout the text that reference that capacity of hypoxia to alter pathogenicity ( This manuscript further highlights the importance of understanding how the new generation of small molecular inhibitors that interact with the oxygen sensing pathway need to be comprehensively investigated not only at the cellular/tissues level but also at the point of host pathogen interaction.

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? Yes
No competing interests were disclosed.

Competing Interests:
I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.

Susanne Schlisio
Department of Cell and Molecular Biology, Ludwig Institute for Cancer Research, Karolinska Institute, Stockholm, Sweden The authors provide convincing evidence that prolyl hydroxylase-decient infection in vivo P. aeruginosa results in increased levels of neutrophil apoptosis, impaired control of infection, and consequently increased mortality in mice. The authors conclude that the expression of the oxygen sensing prolyl hydroxylase homologue (PPHD) in Pseudomonas regulates neutrophil host responses . The in vivo importance of this finding is in light of current development of relatively non-selective PHD inhibitors that most likely inhibit both, the host and Pseudomonas prolyl hydroxylase and thus impact both, host and pathogen.
The experiments and analyses presented in this work appear robust and of high quality. in vivo I only have a minor comment regarding studies in Fig 1B: in vitro It seems that under normoxia, there was no significant change in neutrophil numbers cultured with WT or PPHD mutant supernatant (Fig 1B normoxia -black bar-WT versus MT). Since pyocyanin was significantly increased in the MT under normoxia (Fig 1D), why was that not reflected in significant decrease of neutrophil counts under normoxic conditions in MT versus WT (FIG 1B)? Please provide comments. What is the p value: WT versus MT under normoxia in Fig 1B? Is the mild reduction of neutrophil counts significant under normoxia (WT vs MT)?

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? Yes
No competing interests were disclosed.

Competing Interests:
I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.
Author Response 16 Nov 2017