Prekallikrein inhibits innate immune signaling in the lung and impairs host defense during pneumosepsis in mice

Abstract Prekallikrein (PKK, also known as Fletcher factor and encoded by the gene KLKB1 in humans) is a component of the contact system. Activation of the contact system has been implicated in lethality in fulminant sepsis models. Pneumonia is the most frequent cause of sepsis. We sought to determine the role of PKK in host defense during pneumosepsis. To this end, mice were infected with the common human pathogen Klebsiella pneumoniae via the airways, causing an initially localized infection of the lungs with subsequent bacterial dissemination and sepsis. Mice were treated with a selective PKK‐directed antisense oligonucleotide (ASO) or a scrambled control ASO for 3 weeks prior to infection. Host response readouts were determined at 12 or 36 h post‐infection, including genome‐wide messenger RNA profiling of lungs, or mice were followed for survival. PKK ASO treatment inhibited constitutive hepatic Klkb1 mRNA expression by >80% and almost completely abolished plasma PKK activity. Klkb1 mRNA could not be detected in lungs. Pneumonia was associated with a progressive decline in PKK expression in mice treated with control ASO. PKK ASO administration was associated with a delayed mortality, reduced bacterial burdens, and diminished distant organ injury. While PKK depletion did not influence lung pathology or neutrophil recruitment, it was associated with an upregulation of multiple innate immune signaling pathways in the lungs already prior to infection. Activation of the contact system could not be detected, either during infection in vivo or at the surface of Klebsiella in vitro. These data suggest that circulating PKK confines pro‐inflammatory signaling in the lung by a mechanism that does not involve contact system activation, which in the case of respiratory tract infection may impede early protective innate immunity. © 2019 Authors. Journal of Pathology published by John Wiley & Sons Ltd on behalf of Pathological Society of Great Britain and Ireland.


Introduction
Sepsis is a life-threatening syndrome featured by an unbalanced host response to infection [1,2]. In spite of substantial advancement in the knowledge of pathophysiological mechanisms involved, sepsis still causes a great health burden and mortality remains high [3,4]. Pneumonia is responsible for an inordinate disease burden worldwide [5] and represents the most common cause of sepsis [3,4]. Klebsiella pneumoniae is one of the most frequent Gram-negative causative agents of pneumonia and sepsis [6,7].

C Ding et al
The host response to bacterial infection involves activation of the contact system [8,9]. The contact system consists of three zymogen factors, Factor (F) XII, FXI, and prekallikrein (PKK), and one cofactor, high-molecular-weight kininogen (HK). PKK, once activated by FXIIa to kallikrein, exerts its serine protease activity by cleaving its substrate HK to produce the potent inflammatory mediator bradykinin, which is often referred to as the kallikrein-kinin system. Plasma kallikrein also facilitates the activation of FXII, which initiates the intrinsic pathway of coagulation [8,9]. The contact system may play a dual role in the host response to bacterial infection. Activation of the contact system can occur at the surface of bacterial pathogens and inhibition of the contact system after intraperitoneal infection with Streptococcus pyogenes in mice resulted in enhanced dissemination to the spleen [10,11]. In contrast, during acute overwhelming infections, contact system activation may contribute to mortality, at least in part by inducing septic shock [12,13]. Knowledge on how the contact system influences the host response during a gradually progressing bacterial infection, from a local infectious source to systemic dissemination, is limited.
We recently reported that FXII-deficient mice had lower bacterial burdens and an improved survival in a model that uses a low infectious dose of K. pneumoniae administered via the airways, initially resulting in a localized infection contained within the lungs and subsequently in sepsis with distant organ injury [14]. Remarkably, in the same model of pneumonia-derived sepsis, more downstream interventions in the contact system targeting kininogen or bradykinin did not modify the host response, suggesting that components of the contact system may influence innate immunity independent of their established role in the kallikrein-kinin system [15,16]. The present study aimed to determine the role of PKK in the host response to Gram-negative sepsis caused by pneumonia.

Animals
Male C57Bl/6J mice were purchased from Charles River Inc (Maastricht, The Netherlands) and used at 7-8 weeks of age. The Institutional Animal Care and Use Committee approved all experiments.

Oligonucleotides
Antisense oligonucleotides (ASOs) were synthesized using an Applied Biosystems 380B automated DNA synthesizer (Applied Biosystems, Waltham, MA, USA) and purified as described previously [17]. PKK ASO [18] and a non-specific, scrambled control (Ctrl) ASO were administered subcutaneously, twice weekly, at a dose of 40 mg/kg per week for 3 weeks prior to infection. The ASO dose and treatment regimen were established according to previous mouse studies [18,19].

Experimental design
Pneumonia was induced 4 days after the final treatment with PKK or Ctrl ASO by intranasal inoculation with K. pneumoniae [∼7000 colony-forming units (CFU) in 50 μl of isotonic saline] as described previously [21][22][23]. Mice were euthanized 12 or 36 h after induction of pneumonia (n = 8 per group at each time point) or observed for 10 days (n = 20 per group). During the latter experiment, clinical signs were scored as described previously [21]. At predefined endpoints, blood was collected in sodium citrate prefilled syringes (volume ratio 4:1); organs were harvested and homogenized; and bacteria quantified as described previously [21][22][23]. In some experiments, bronchoalveolar lavage fluid (BALF) was collected through flushing the lungs via the trachea three times with 300 μl of 0.2 mM EDTA/PBS.

Measurement of PKK (Klkb1) mRNA
RNA was isolated using an RNA isolation kit (Nucleospin RNA, Macherey-Nagel, Düren, Germany) and then stored at −80 ∘ C until used for reverse transcription and polymerase chain reaction (PCR). Total RNA was transcribed to cDNA using oligo dT primer and Moloney murine leukemia virus reverse transcriptase (Promega Benelux, Leiden, The Netherlands). Gene expression levels were quantified with SYBR Green reactions on a LightCycler system (LC480; Roche Applied Science, Mannheim, Germany). Gapdh was used as a reference. The primer sequences were as follows: Klkb1: AGTACCGGAAGAAGTGCCTG (forward) and GTGAAGAAAAGGCAGTTGGG (reverse); Gapdh: CTCATGACCACAGTCCATGC (forward) and CACATTGGGGGTAGGAACAC (reverse). Real-time PCR data were analyzed using LinRegPCR software (v.2014.4) (available at http://LinRegPCR.nl).

Histology
Four-micrometer tissue sections were stained with hematoxylin and eosin (H&E) and scored as previously described [24] with respect to the following parameters: bronchitis, edema, interstitial inflammation, intra-alveolar inflammation, pleuritis, endothelialitis, and percentage of the lung surface demonstrating confluent inflammatory infiltrate. Each parameter was graded 0-4, with 0 being 'absent' and 4 being 'severe'; the total pathology score was expressed as the sum of the score for all parameters. Granulocyte staining used FITC-labeled rat anti-mouse Ly-6 mAb (clone 1A8; BD Pharmingen, San Diego, CA, USA; catalog number 551460; 1:20 000). The amount of Ly-6G positivity was expressed as a percentage of the total surface area. Fibrinogen staining on lung sections was performed as described previously [25] using a biotin-labeled goat anti-mouse fibrinogen antibody (Accurate Chemical & Scientific, Westbury, NY, USA; catalog number YNGMFBGBio; 1:500). Slides were scanned using an Olympus dotSlide scanner (Olympus, Tokyo, Japan) to generate TIFF images of the full tissue section. Ly-6G and fibrinogen positivity was measured using ImageJ; the amount of positivity was expressed as a percentage of the total lung surface area [21,25].

Western blotting
For PKK and HK immunoblotting analyses, citrated plasma was fractionated on 4-12.5% polyacrylamide gels under reducing conditions and separated proteins were transferred to polyvinylidene difluoride membrane (Roche Diagnostics, Almere, The Netherlands) followed by immunoblotting with goat anti-mouse PKK antibody (AF2498; R&D Systems, Minneapolis, MN, USA) or mouse monoclonal (13G11) anti-human PKK (GTX21006; GeneTex, Irvine, CA, USA). Rabbit polycolonal antibody against human full-length HK was used for HK immunoblotting (Abnova, Taipei City, Taiwan). Horseradish peroxidase-linked rabbit IgGs were used as secondary antibodies and blots were imaged using chemiluminescence and an ImageQuant LAS 4000 imager (GE Healthcare Life Sciences, Little Chalfont, UK).

Bacterial binding/activation of contact system
Binding and activation of contact system factors by bacteria were determined essentially as described previously [26]. In brief, bacteria in buffer C (containing 15 mM HEPES, 135 mM NaCl, 50 μM ZnCl 2 , pH 7.4) were incubated with citrated mouse or human plasma (1:1 v/v) on a roller for 1 h at room temperature. Bacteria (∼4 × 10 10 CFU) were then pelleted and washed twice with buffer C. Bound proteins on bacterial surfaces were eluted with 0.1 M glycine (pH 2.0) and analyzed by SDS polyacrylamide gel electrophoresis for detection of PKK and HK and derived activation products.

RNA microarrays and bioinformatics
RNA was isolated from lung and liver homogenate using a Nucleospin RNA isolation kit (Macherey-Nagel). Only RNA samples with an integrity number (RIN) greater than 6 (Agilent Bioanalyzer; Agilent Technologies, Amstelveen, The Netherlands) were used for microarray analyses. RNA was hybridized to the mouse Clariom S Assay HT chip (Thermo Fisher Scientific, Waltham, MA, USA). Pre-processing and quality control of the scans were performed by using the oligo method (version 1.44) [27] and probes were annotated using the platform design information for Affymetrix Clariom_S_Mouse_HT available through Bioconductor [28]. Array data were background-corrected using Robust Multi-array Average (RMA) and quantiles-normalized. Microarray quality control was performed by means of the array quality metrics method (version 3.36.0) [29]. The occurrence of non-experimental chip effects was evaluated by the surrogate variable analysis method (version 3.28.0) [30] and corrected using the combat method [31]. Probes were filtered by means of a 0.5 variance cut-off using the genefilter method (version 1.62.0) [32], resulting in 14 564 expressed transcripts. Comparison between groups was performed by moderated t-statistics implemented in the empirical Bayesian linear models method limma (version 3.36.2) [33]. Throughout, Benjamini-Hochberg multiple comparison adjusted probabilities (adjusted p < 0.05) were used to determine significance. All analyses were performed in the R statistical computing environment (version 3.5.0). Pathway analysis was performed using Ingenuity pathway analysis (Ingenuity Systems, Qiagen Bioinformatics, Redwood City, CA, USA; https://www .qiagenbioinformatics.com). Significance was evaluated using Fisher's exact test and Benjamini-Hochberg adjusted P values (adjusted p < 0.05). Normalized and non-normalized array data are accessible through the Gene Expression Omnibus with accession number GSE121970.

Statistical analysis
Data are expressed as described in the figure and table legends. Mann-Whitney U-tests were performed for comparisons between groups. Kruskal-Wallis one-way analysis of variance test followed by a Dunn's multiple comparison test was used for groups of three or more. Survival rates are depicted by Kaplan-Meier plots and were compared by log-rank test. Clinical observation scores were compared by calculating areas under the curves for each mouse followed by a t-test. A P value below 0.05 was considered statistically significant. All analyses were performed using GraphPad Prism 5 (GraphPad Inc, San Diego, CA, USA).

Inhibition of PKK gene expression and protein production by PKK ASO treatment
In order to suppress expression of Klkb1 (the gene encoding PKK), mice were treated twice weekly with a PKK ASO for 3 weeks according to a previously described dosing schedule [18,19]; control mice were treated with a scrambled Ctrl ASO. After 3 weeks of treatment (designated t = 0 in the infection experiments described below), Klkb1 mRNA expression in the liver was inhibited by over 80% in PKK ASO-treated mice compared with mice treated with Ctrl ASO ( Figure 1B). Consistently, plasma PKK protein and activity were almost completely abolished in PKK ASO-treated mice ( Figure 1A,C). In accordance, PKK ASO administration was associated with a strong prolongation of the aPTT ( Figure 1D), while the PT remained unaffected ( Figure 1E). Klkb1 mRNA was not detectable in lungs. Having established that PKK ASO treatment efficiently reduced PKK expression, we went on to infect PKK ASO-and Ctrl ASO-injected mice with Klebsiella via the airways. The effect of PKK ASO administration on liver Klkb1 mRNA, plasma PKK activity, and the aPTT was sustained for at least 36 h after infection ( Figure 1A-D). Notably, in Ctrl ASO-treated mice, liver Klkb1 mRNA and plasma PKK activity levels decreased during the infection ( Figure 1B,C), which was accompanied by a prolongation of the aPTT over time ( Figure 1D). The activities of other components of the contact system did not decrease (FXII) or even increased (HK) during infection (supplementary material, Figure S1A,B). PKK ASO treatment was associated with elevated plasma FXII activity at t = 0 and t = 12 h, and with reduced plasma HK activity at 36 h. PKK ASO administration diminished activation of coagulation, as indicated by an attenuated rise in plasma TATc levels (supplementary material, Figure S1C).

PKK depletion improves host defense during pneumosepsis
To obtain a first insight into the role of PKK in host defense during pneumosepsis, we infected mice pretreated with either PKK or Ctrl ASO with Klebsiella via the airways and observed them for 10 days. Ctrl ASO mice started to show symptoms 1 day after induction of pneumonia (Figure 2A), and most of them died between 40 and 72 h after infection, with a median survival time of 55 h ( Figure 2B). In contrast, mice pretreated with PKK ASO died mostly between 72 and 120 h after infection (median survival time 96 h; p < 0.0005 versus Ctrl ASO mice). In accordance with the delayed mortality in PKK ASO-treated mice, these animals displayed a postponed onset of symptoms (p < 0.05 versus Ctrl ASO mice). We assessed whether the protective effect of PKK deletion was associated with an altered pattern of bacterial growth and dissemination. To this end, we determined the bacterial loads at the primary site of infection (lungs) and distant organs (blood, spleen, and liver) at an early (12 h) and a late time point (36 h; i.e. before the first deaths occurred) after infection. Consistent with the symptom scores, at 12 h, bacterial loads in the lungs were similar in PKK-and Ctrl ASO-treated mice ( Figure 2C); at this early time point, the vast majority of mice did not show evidence of dissemination of the infection (Figure 2D-F). At 36 h after infection, however, lung bacterial loads were approximately 100-fold lower in PKK ASO-treated mice relative to Ctrl ASO-treated animals (p < 0.01); similarly, bacterial loads in blood, spleen, and liver were also 10-to 100-fold lower in PKK ASO-treated mice (p < 0.01 to p < 0.05). These results suggest that PKK depletion prolongs survival by reducing bacterial growth and dissemination during Gram-negative pneumonia-derived sepsis.

Effect of PKK depletion on lung pathology and inflammation during pneumosepsis
The extent of lung inflammation increased in both groups after infection, and no difference was observed between PKK-and Ctrl ASO-treated mice (supplementary material, Figure S2). Likewise, PKK depletion did not impact fibrin deposition in lung tissue   (supplementary material, Figure S3). Considering that influx of neutrophils into the respiratory tract is a hallmark feature of bacterial pneumonia [34] and that contact system activation products can enhance neutrophil recruitment [9,35], we next quantified neutrophil numbers in lung tissue using immunohistopathology and digital imaging (supplementary material, Figure  S4A,B). Pneumonia was associated with a gradual increase in neutrophils in lung tissue, which was not different between PKK-and Ctrl ASO-treated mice.
Consistently, neither total cell counts (supplementary material, Figure S4C) nor neutrophil numbers (supplementary material, Figure S4D) in the bronchoalveolar space differed between treatment groups. To further study the effect of PKK depletion on lung inflammation after infection, we measured the levels of cytokines (TNFα, IL-1β, IL-6) and chemokines (CXCL1, CXCL2) in lung homogenates (Table 1). Directly prior to infection, PKK ASO-treated mice tended to have higher lung levels of these pro-inflammatory mediators when compared with Ctrl ASO-injected mice (although not significantly higher). After infection, pulmonary cytokine and chemokine levels were lower in PKK ASO-treated mice than in Ctrl ASO-treated mice, albeit with a large inter-individual variation; significant differences were found for IL-6 and CXCL1 at 12 h after infection, and for IL-6, IL-1β, and CXCL2 at 36 h after infection.

PKK depletion attenuates distant organ damage during pneumosepsis
To obtain further insight into the mechanism by which PKK depletion affected outcome, we measured plasma ALT and AST as markers of hepatocellular injury and plasma LDH to indicate cellular injury in general. Plasma ALT, AST, and LDH levels increased markedly during the late stage of sepsis in both treatment groups, but significantly less so in PKK ASO-treated mice ( Figure 3). Moreover, Ctrl ASO-but not PKK ASO-injected mice showed thrombus formation in liver tissue 36 h after infection (supplementary material, Figure S5).

Klebsiella does not activate the contact system
To study activation of the contact system and the role of PKK herein, we measured cleaved PK and HK by immunoblotting analysis of plasma from infected mice and human plasma exposed to K. pneumoniae in vitro. Kaolin (positive control) activated PKK to α-PKa, which is immediately captured by C1 inhibitor in plasma, as detected by the appearance of α-PKa/C1 inhibitor complexes (185 kDa product; non-reduced condition) [36] and α-PKa heavy chain generation (reduced condition; heavy chain is liberated of the serine protease domain containing light chain that is captured by C1 inhibitor) in mouse plasma ( Figure 4A). In contrast, neither α-PKa/C1 inhibitor nor α-PKa heavy chain was detected in the plasma of Ctrl ASO-or PKK ASO-treated mice infected with Klebsiella ( Figure 4A). We further tested whether the K. pneumoniae strain used in our in vivo studies can activate the contact system in vitro, using E. coli Ymel (able to bind both PKK and HK) and the E. coli mutant Ymel-1 (not able to bind PKK or HK) as comparators [20]. Neither full length PKK nor α-PKa heavy chain was detectable on K. pneumoniae incubated with plasma, indicating that PKK was not absorbed/activated on the cell surface of this bacterium ( Figure 4B). PKK and activation products could be detected at the surface of E. coli Ymel and, to a limited extent, of the E. coli mutant Ymel-1. Likewise, HK could be bound and cleaved at the surface of E. coli Ymel (cHK), while not so on the surface of Klebsiella or E. coli Ymel-1 ( Figure 4C).

PKK depletion is associated with an upregulation of multiple innate immune pathways in the lungs
To gain insight into the mechanism by which PKK modulates the host response during pneumonia-derived sepsis, we chose an unbiased approach and investigated the effect of PKK depletion on lung and liver transcriptomes. When comparing lung transcriptomes of uninfected PKK ASO-treated mice with those of Ctrl ASO-administered mice, 1237 significantly altered genes were revealed ( Figure 5A). Pathway analysis revealed that overexpressed genes in PKK ASO-treated mice were significantly associated with pathogen recognition and several innate/adaptive immunity pathways including TREM-1 signaling, interferon signaling, Th1 pathway, and crosstalk between dendritic cells and natural killer cells ( Figure 5B,C). Infection with Klebsiella caused a profound change in the lung transcriptome relative to uninfected mice characterized by enhanced expression of several innate immune pathways (supplementary material, Figure S6A,B). The overexpression of several pathways in naïve mice treated with PKK ASO was sustained at 12 h after infection, most notably so for pathogen recognition, crosstalk between dendritic cells and natural killer cells, and natural killer signaling, while at this time point also complement signaling was enhanced in PKK ASO-injected mice ( Figure 6A-C). Pathways that were underexpressed in PKK ASO-treated mice, amongst others, related to fibrosis and agranulocyte adhesion and diapedesis.
The array analyses at t = 0 and t = 12 h confirmed that Klkb1 mRNA was not expressed in the lungs (data not shown). In the liver, transcriptome analysis also pointed to a marked activation of several innate and adaptive immune pathways in PKK-depleted mice, both directly prior to and 12 h after infection (supplementary material, Figures S7A,B and S6C,D, respectively). By comparison, infection per se induced less pronounced upregulation of mRNAs in liver 12 h after infection (supplementary material, Figure S7E,F).

Discussion
We have used an established model of lower respiratory tract infection caused by the common human pathogen K. pneumoniae to investigate the role of PKK in host defense. This model is characterized by infection with a low bacterial dose via the airways initially resulting in a localized infection contained within the lungs and subsequently in sepsis with distant organ injury and death, thereby allowing for studies on both the early protective immune response and the late detrimental consequences of ongoing immune activation [21][22][23]. At the initiation of our investigations on the role of the kallikrein-kinin system in pneumonia-derived sepsis, we anticipated that an impairment of the function of this pluripotent pro-inflammatory system, such as by depletion of PKK, would impair the early innate immune response and result in enhanced bacterial growth. During later phases of the infection, PKK depletion in theory could have protected against organ injury, consistent with earlier studies that documented the harmful effects of activation products of the kallikrein-kinin system in models of acute fulminant infection or endotoxemia [12,13]. In contrast to these expectations, PKK depletion was associated with enhanced expression of multiple innate immune pathways in lungs and liver, together with an improved antibacterial defense. While hepatic PKK production and plasma PKK activity decreased during the progression of pneumonia-derived sepsis in control animals, mice in which PKK was depleted prior to infection showed delayed mortality together with reduced distant organ injury prior to the occurrence of the first deaths, most likely due to limitation of bacterial spread. Considering that we could not detect Klkb1 mRNA in lungs, these data suggest that circulating PKK, originating from constitutive PKK production in the liver, restrains pro-inflammatory signaling in the lung, which in the case of respiratory tract infection may impede early protective innate immunity.
We recently reported on the role of other components of the contact system in the host response during pneumonia-derived sepsis using the same model. Neither kininogen depletion (by ASO treatment) nor kininogen deficiency influenced bacterial growth or inflammatory responses after infection with K. pneumoniae via the airways [15]. Likewise, neither bradykinin receptor deficiency nor inhibition modified the host response in this model [16]. These earlier data and the fact that we were unable to detect activation of the contact system suggest that the mechanism by which PKK impacted host defense is unrelated to its role in the kallikrein-kinin system. While to the best of our knowledge zymogen PKK has not been linked with regulation of innate immunity, zymogen FXII has recently been shown to enhance several innate immune functions of neutrophils [37]. Other findings have also indicated that zymogen FXII may influence cell functions independent of its enzymatic action [38]. Notably, we previously documented that kininogen-depleted or -deficient mice have much lower plasma PKK levels (∼40% of the PKK levels in the respective control mice) [15]. We consider it likely that the markedly different phenotypes of kininogen-depleted/deficient mice versus PKK-depleted mice during Klebsiella pneumosepsis relate to the different extents of plasma PKK reduction, which was virtually complete in PKK ASO-treated mice.
PKK ASO treatment resulted in a sustained inhibition of Klkb1 mRNA expression in the liver and a profound reduction in plasma PKK activity. This effect was associated with induction of multiple innate immune pathways already prior to infection, especially in the lung. These immune-enhancing effects were not associated with changes in lung pathology or recruitment of white blood cells or neutrophils to the bronchoalveolar space (supplementary material, Figure  S4), or of neutrophils (supplementary material, Figure  S4), lymphocytes or macrophages (data not shown) to lung tissue, suggesting that these remained restricted to an extent not causing autoinflammation. Nonetheless, these pro-inflammatory modifications of innate immune signaling pathways are a likely explanation for the observed protective effect of PKK ASO treatment. Indeed, in general, early stimulation of innate immunity in the lung is associated with an improved defense against respiratory pathogens in pneumonia models [34], and specific pathways induced by PKK ASO administration such as pattern recognition receptors [39,40], TREM-1 signaling [41] and interferon signaling [42,43] have been implicated in protective immunity during Klebsiella pneumonia (supplementary material, Figure S8). A non-specific pro-inflammatory effect of PKK ASOs is unlikely considering that scrambled Ctrl ASOs were used in all experiments. In addition, unlike the strong improvement of host defense provided by PKK ASO treatment, ASOs targeting kininogen [15] or FXII [44] did not influence bacterial loads or inflammatory responses in the same model of Klebsiella-induced pneumosepsis.
Progressive infection was accompanied by a selective decline in plasma PKK activity; plasma FXII and plasma HK activity remained unaltered or increased, respectively. It is tempting to speculate that this decline in systemic PKK levels serves to facilitate the innate immune response in the lungs during pneumonia. PKK depletion was associated with elevated plasma FXII activity, which is in accordance with a previous study [18] and consistent with FXII being a substrate for activated PKK.
PKK depletion was associated with reduced distant organ injury. Although earlier investigations documented an injurious role for contact system activation in fulminant models of endotoxemia and sepsis [12,13] in the present investigation activation of the contact system could not be detected and the protective effect of PKK ASO administration likely was due to the lower bacterial burdens. Likewise, while previous studies have documented a role for PKK in thrombosis [18,45], diminished thrombus formation in livers of PKK ASO-treated mice could also be explained by the lower bacterial burdens in these animals, providing a reduced pro-coagulant stimulus. This would also explain the lower plasma TATc levels in PKK ASO-administered animals during late-stage infection. Alternatively, and not mutually exclusive, severe infection such as induced here is likely to result in significant hypoxia, which could explain differences in hepatocellular injury between groups, considering the less severe disease course in PKK-depleted mice.
Our study is limited by the fact that we do not provide a mechanism by which PKK influences innate immune pathways. In addition, gene expression profiles were studied in tissues and the effect of PKK depletion on specific cell types remains to be established. Moreover, we did not study the effect of PKK or PKK-depleted serum on the killing capacity of phagocytes; for the results presented here, this would have been difficult considering that the highly virulent Klebsiella strain used cannot be killed by phagocytes in vitro [46]. Additional experiments using purified PKK and purified cells are warranted to obtain more insight into the role of PKK in host defense and innate immune responses.
In conclusion, we have shown that constitutive production of PKK in the liver is diminished during pneumonia-derived sepsis caused by the common human pathogen K. pneumoniae, resulting in declining plasma PKK concentrations in the absence of detectable contact system activation. PKK depletion prior to infection resulted in activation of multiple innate immune pathways in the lungs, which upon inoculation of Klebsiella via the airways was associated with reduced bacterial growth and dissemination, attenuated distant organ injury, and an improved survival. These results point to a novel inflammation-confining role of circulating PKK, which during a gradually evolving infection can impede protective immunity.