High-molecular-weight kininogen is a critical component of host defense against Escherichia coli sepsis

Abstract

High-molecular-weight kininogen (HK) is known to bind lipopolysaccharides (LPS) with high affinity and serves as a crucial LPS carrier in circulation, supporting endotoxemia. However, its role in host defense against Gram-negative bacterial infection remains unclear. Here we demonstrate that HK directly binds to Escherichia coli (E. coli) via LPS and rapidly localizes to sites of infection. HK-deficient mice (Kng1 ^{– /–}) showed increased susceptibility to infection, with increased bacterial dissemination, lung injury, and proinflammatory cytokine production. In contrast, endogenous expression of human HK in Kng1 ^{– /–} mice restored survival, limited bacterial spread, and reduced tissue damage. Mechanistically, HK promoted neutrophil antimicrobial responses by enhancing reactive oxygen species production and microbicidal activity. Consistently, liver-specific HK deficiency recapitulated the impaired bacterial clearance and reduced survival upon E. coli challenge, highlighting the importance of plasma HK. Together, these findings identify HK as a new soluble pattern recognition molecule that senses E. coli invasion and initiates neutrophil-mediated antimicrobial responses, revealing a previously unrecognized protective function of the contact system in innate immunity.

Introduction

Sepsis, defined as a life-threatening syndrome caused by a dysregulated host response to infection, leads to a global health burden, with high mortality rates despite substantial advances in understanding its pathogenesis and management.^1,2 Bacterial infections are the predominant cause of sepsis, with Gram-negative bacteria accounting for approximately 60% of the infections.^3,4

A growing body of evidence suggests that activation of the contact system contributes to the pathophysiology of human sepsis.^5–9 The contact system consists of high molecular weight kininogen (HK) and three serine proteases-factor XII (FXII), factor XI (FXI), and plasma prekallikrein (PK). In circulation, HK forms complexes with either FXI or PK.^10,11 When blood “contacts” with negatively charged artificial or biological surfaces, contact factors bind to it and FXII becomes auto-activated. ¹² Activated FXII subsequently activates FXI and/or PK. Activation of FXI triggers the intrinsic coagulation pathway, whereas activation of PK (PKa) cleaves HK to release the potent inflammatory mediator bradykinin (BK), initiating the kallikrein–kinin system.^13,14 BK has been shown to recruit neutrophils,^15,16 and stimulate alveolar macrophages ¹⁷ and dendritic cells, ¹⁸ thereby potentiate host response against invading pathogens. Moreover, HK can bind to the surfaces of major bacterial pathogens, where it undergoes proteolytic cleavage to release antimicrobial peptides (AMPs).^19,20 These AMPs, derived from HK domains D3 and D5, exhibit broad-spectrum antibacterial activity. ²¹ Due to the release of AMPs and BK, which trigger inflammatory reactions, the contact system functions as an important branch of innate immune defense against microbial invasion. However, excessive BK-mediated vascular permeability may promote systemic dissemination of pathogens^22–24 and contribute to septic shock by inducing hypotension.^21,25 Thus, HK appears to play a dual role in host response to bacterial infection.

Lipopolysaccharides (LPS), the major structural element of Gram-negative bacteria outer membrane, triggers strong immune responses. ²⁶ Sensing of LPS by innate immune cells is vital for host defenses against Gram-negative infections. We have previously reported that HK binds LPS in a high affinity and serves as a crucial LPS carrier in circulation, thereby supporting endotoxemia. ²⁷ However, knowledge on how the HK influences the host response during the progressing Gram-negative bacterial infection is limited. A recent study suggested that HK plays a limited role in local host defense during pneumonia sepsis by a Gram-negative bacteria Klebsiella pneumoniae. ²⁸ Whether HK influences the host response during E. coli infection—the most frequently isolated Gram-negative bacteria in human sepsis, remains unknown. In the present study, we investigated the role of HK in the systemic host response to peritonitis-induced sepsis caused by E. coli.

Materials and methods

Animals

Mice deficient in Kng1 (Kng1^−/−) and wild-type (Kng1^+/+) littermates were used in this study. ²⁷ H11-Alb-hKng1-polyA knock-in mice were made via CRISPR/Cas9 system. Cas9 mRNA, sgRNA and donor were co-injected into zygotes. sgRNA direct Cas9 endonuclease cleavage at H11 locus and create a double-strand break. Such breaks were repaired and resulted in Alb-hKng1-polyA inserted into H11 locus. The knock in pups were genotyped by PCR (H11-polyA-tF1: 5'-CCTGCTGTCCATTCCTTATTCCATA-3ʹ, H11-tR1: 5'- ATATCCCCTTGTTCCCTTTCTGC-3ʹ (WT: none, KI: 329 bp). H11-tF2: 5'-CCTCCAAGTCTTGACAGTAG-3ʹ, H11-R2: 5'-CTCTCTAGCGTGTCTATACAC-3ʹ. (WT: 418 bp, KI:5264 bp), followed by sequencing and southern blotting analysis. Then the positive knock-in mice (hKng1^ki/ki) were crossed with Kng1^−/− mice to generate HK humanized (hKng1^+/+) mice. Alb-cre/Kng1^fl/fl mice were generated by mating Alb-cre mice with Kng1^fl/fl mice. Age-matched (8 to 12-weeks) male littermates were used. Experiments with mice were performed in accordance with protocols approved by the Institutional Animal Care and Ethics Committee of Soochow University.

Animal infection model

Bacterial intraperitoneal model: E. coli bacteria (ATCC 25922) were cultured in Luria Bertani (LB) medium at 37 °C, harvested at mid-log phase (OD₆₀₀≈0.5), washed and diluted in PBS to a concentration of 1 × 10⁹ CFU/ml. 1 × 10⁸ CFU Escherichia coli (E. coli) suspension were intraperitoneally injected into mice. To assess survival, the mouse status was observed for indicated days. To determine bacterial load, animals were killed 16 h after initiation of infection. The organs were harvested and homogenized, peritoneal lavage fluids and a citrated blood sample were collected, and the 10-fold serial dilutions were plated on blood agar plates and incubated overnight at 37 °C. The plasma cytokines levels were determined using a commercially available cytometric beads array multiplex assay were performed as previously described. ²⁷

Cecal ligation and puncture (CLP) model ²⁹ : Mice were anesthetized with isoflurane. The cecum was exteriorized through a midline celiotomy, ligated with 4–0 silk suture 1 cm from the distal end (low-grade injury), then punctured through with a sterile 21-gauge needle. After a drop of cecal contents was extruded, the cecum was returned to the abdominal cavity, and the incision was closed with surgical staples. In sham control animals, the cecum was exposed but not ligated or punctured and then returned to the abdominal cavity. All mice received 1 ml warm 0.9% NaCl subcutaneously (SC) post-surgery. After surgery, mice were evaluated every 6 h, and received 1 ml warm 0.9% NaCl SC daily.

Histopathology

Lung was isolated from mice sacrificed 16 h post-E. coli infection or PBS control, fixed with 4% paraformaldehyde and then processed to paraffin, embedding. Sections were cut at 5 μm and stained with hematoxylin-eosin. Images were captured with a Leica DM2000 microscope (Leica Microsystems, Wetzlar, Germany).

HK immunoblot analysis

Mouse citrated plasma was separated on SDS-PAGE (10% separating gel) followed by immunoblotting with horseradish peroxidase-linked monoclonal anti-human HK domain 5 antibody (referred as 6C9G4). The antibody was raised against a peptide derived from the human HK domain 5 sequence “DHGHKHKHGHGHGKH” (the sequence is responsible for the binding of HK to LPS). ²⁷ Blots were imaged using chemiluminescence substrate NcmECL Ultra (NCM Biotech) and a chemiluminescent Imager 600 (Amersham Biosciences).

HK and E. coli binding assay

In order to detect whether HK is in cooperation with E. coli in vivo, purified HK protein (Enzyme research laboratories) was radiolabeled with ¹²⁵I-Na using Iodogen (Pierce) according to previously described, ³⁰ by which radiolabeled HK (¹²⁵ I-HK) retained its procoagulant activity and antigenic property. ¹²⁵1-bovine serum albumin (BSA) was prepared by a similar procedure using lodogen. The specific radioactivity of the protein was about 10 μCi/μg and > 97.4% of them were iodinated. Kng1^−/− were anesthetized and SC inoculated with 1 × 10⁸ CFU E. coli suspension or PBS into left and right ear respectively, following by ¹²⁵ I-HK (100 μCi/ ouse) or ¹²⁵ I-BSA(100 μCi/mouse) i.v injection. 2 h later, radioactivity was assessed by In-Vivo Imaging System FX Pro (Eastman Kodak Company).

As detection the binding of HK to E. coli by flowcytometry, HK protein were labeled using FITC labeling kit (Invitrogen™) according to manufacturer's instruction. E. coli (1.5 × 10⁵ CFU per reaction) were incubated with FITC-HK in the presence or absence of LPS (Sigma-Aldrich) in PBS containing 50 μM Zn²⁺, 0.01% Tween-20, pH 7.4 (100 μL) for 30 min at 4 °C. After extensively washed with PBS, the degree of FITC positive was determined on a FACS Calibur (BD Biosciences).

Neutrophil isolation and bactericidal assay

Mice whole blood was subjected to red cell lysis by adding 3 ml of cold ammonium chloride lysis solution pH 7.2, the reaction was stopped immediately by the addition of serum-free media and the mixture was centrifuged at 400 g for 10 min at 4 °C. Cells were filtered through a sterile MACS 30-μm pre-separation filter to remove cell clumps, and neutrophils were separated by negative selection using the MACS magnetic bead separation system (Miltenyi Biotec) according to the manufacturer's instructions. The isolated neutrophil cells were identified by fluorescent staining with anti-Ly6G-PE and anti-CD11b-FITC antibodies (ebioscience), followed by analysis using flow cytometry. Neutrophil cells (5 × 10⁵ cells) were incubated with 5 × 10⁶ CFU E. coli for 30 min at 37 °C in RPMI 1640 medium supplemented with 10% of autologous serum, after spinning bacteria onto the cells at 300 g for 1 min at 4 °C to synchronize infection. After 30 min, cells were extensively washed and treated with gentamicin (500 μg/ml) for 30 min to kill extracellular bacteria, but leave intracellular bacteria viable. To measure bacterial killing, cells were then washed and lysed in 1% Triton-X 100, lysates were serially diluted and plated directly onto LB agar plates. Total viable bacteria CFU values were counted after overnight incubation at 37 °C.

Detection of neutrophil reactive oxygen species production

200 μl of whole blood (collected with heparin) form mice was equilibrated at 37 °C for 20 min with 10 μM DCFH-DA (Beyotime Biotechnology). 5 × 10⁶ CFU E. coli were then added and incubated for 15 min at 37 °C in a shaker. Samples were immediately placed on ice and subjected to red cell lysis. The fluorescence of DCF in Ly6G (ebioscience) positive neutrophil was determined on a FACS Calibur.

Statistical analyses

Data were expressed as mean ± SEM from at least three independent experiments. Statistical analysis was performed using GraphPad Prism (Version 8.2.1). The normality of data distribution was assessed using the Shapiro-Wilk test in GraphPad Prism. Data that fitted the normal distribution were analyzed using parametric statistical tests. Two-tailed Welch's t-test were performed for comparison of two groups. One-way analysis of variance (ANOVA) analysis followed by a Dunn's multiple comparison was used for groups of three or more. Survival rates were depicted by Kaplan–Meier curves and were compared by log-rank test. P < 0.05 were considered as statistically significant.

Results

HK identifies Gram-negative bacteria by recognizing LPS

It has been reported that HK binds to a variety of Gram-negative bacteria in vitro. To determine whether HK also plays this function in vivo, autoradiography imaging system was applied to trace colocalization of iodinated HK and E. coli at the sites of infection. When radiolabeled HK was injected into Kng1^−/− mice bearing E. coli infection, it was rapidly recruited into the site of infection (Figure 1A), indicating its in vivo association with E. coli. Given growing evidence that systemic contact activation is a hallmark feature in septic patients,^5,6,8,22,28 we next analyzed systemic HK level in a mouse model with E. coli peritonitis. Sixteen hours post-infection, a marked decrease of plasma HK level was observed, as shown by a decrease in full-length HK (110 kDa) and an increase in HK degradation product HK light-chain (56 kDa) (Figure 1B-C). Together, these findings indicate that HK interacts locally and systemically with E. coli during infection.

Figure 1.

HK directly binds to E. coli through LPS. (A) Representative autoradiography images of colocalization of ¹²⁵I-HK and E. coli in mKng1^−/− mice.¹²⁵I-BSA were used as negative control. (B) Plasma HK levels before and after intraperitoneal injection with E. coli by Western blot. (C) Relative quantification of HK (110 kDa), or degradation products (56 kDa) were determined by densitometry of the bands from Figure 1B. *P ＜ 0.05 and ***P ＜ 0.001. (D) FITC-HK in different concentration were incubated with washed E. coli (1.5 × 10⁵ CFU) at 4°C for 30 min. The binding of HK to E. coli was determined by flow cytometry. *P ＜ 0.05 and ***P ＜ 0.001. (E) For LPS competition test, E. coli (1.5 × 10⁵ CFU) were incubated with FITC-HK (50 nM) in the presence or absence of LPS at the indicated concentration. The degree of FITC positive HK binding were analyzed by flow cytometry. ^***P ＜ 0.001. E. coli: Escherichia coli; LPS: lipopolysaccharides; BSA: bovine serum albumin; HK: high-molecular-weight kininogen.

LPS is the most abundant component of the outer membranes of Gram-negative bacteria and previously we have verified that HK binds to LPS with high affinity, ²⁷ so we determined whether HK binds E. coli through LPS. As detected by flow cytometry, HK binding to E. coli is dose-dependent and reached saturation at approximately 150 nM, about one-quarter of its plasma concentration ³¹ (Figure 1D). Excess LPS competitively inhibited this binding in a dose-dependent manner (Figure 1E). These results indicated that HK recognizes not only purified LPS but also intact E. coli cells, underscoring its potential role in the host response to Gram-negative bacterial infection.

HK deficiency impairs survival in CLP and E. coli bacteria induced septic mice

To explore the role of HK in host response to Gram-negative bacterial infection, Kng1^−/− mice were subjected to E. coli peritoneal infection or CLP. In both peritonitis sepsis models, HK-deficient mice showed a significant decrease in survival rate than WT mice (Figure 2A). In the low-grade CLP model, which typically produces mild disease, 90% of WT mice survived, whereas survival dropped to 45.5% in Kng1^−/− mice. In E. coli induced septic peritonitis, WT mice showed a survival rate of 44% with a median survival time of 3.9 days. Strikingly, all Kng1^−/− mice died within 32 h with a strongly reduced median survival time of only 19 h. These results indicated that HK is required for host defense against Gram-negative bacterial sepsis.

Figure 2.

HK deficiency increases susceptibility to CLP and E. coli induced peritonitis. (A) Survival of mice were followed after challenged with low-grade CLP (n = 11) and 1 × 10⁸ CFU E. coli (n = 15). (B) Representative images (100× magnification) of paraffin-embedded sections of lung stained with H&E before, and 16 h after intraperitoneal infected with 1 × 10⁸ CFU E. coli in Kng1^+/+and Kng1^−/− mice. (C) Plasma samples from mice per group ((n = 5) were collected before, and 6 h after E. coli infection. The plasma concentration of proinflammatory cytokines TNF-α, IL-6, IL-1β, MCP-1 and HMGB1 were measured using Multi-Plex immunoassay. *P ＜ 0.05; **P ＜ 0.01; ***P ＜ 0.001. E. coli: Escherichia coli; CLP: cecal ligation and puncture; H&E: hematoxylin and eosin.

HK-deficient mice have more severe lung injury and systemic inflammation

To examine the relationship between HK deficiency and organ injury, lung histology was examined in Kng1^+/+ and Kng1^−/− mice after E. coli infection. H&E staining of lung tissue at 16 h revealed more severe pathology in Kng1^−/− mice, including increased thickness of the interstitial space of alveoli, infiltration of inflammatory cells, and congestion/hemorrhage (Figure 2B). Consistence with these histological findings, Kng1^−/− mice displayed markedly elevated plasma levels of proinflammatory cytokines TNF-α, IL-6, MCP-1, and HMGB1compared with WT mice post E. coli infection (Figure 2C). Taken together, these results suggested that HK may be an important plasma factor conferring a survival advantage during Gram-negative bacterial infection.

HK deficiency increases bacterial dissemination and outgrowth

To elucidate the mechanism underlying HK-mediated protection, bacterial burdens were determined in peritoneal fluid, blood, liver and spleen 16 h after E. coli infection. CFU analysis revealed significantly higher bacterial loads in their peritoneal fluid, blood, liver and spleen of Kng1^−/− mice compared to wild-type mice, suggesting an ability for HK in recognition and clearance of E. coli (Figure 3).

Figure 3.

HK deficiency enhances bacterial dissemination and growth in E. coli peritonitis mice. Kng1^+/+ mice and Kng1^−/− mice (n = 16) received an i.p. injection of 1 × 10⁸ CFU E. coli. Bacterial counts were determined by CFU in peritoneum, blood, spleen and liver homogenates 16 h after infection with E. coli. **P ＜ 0.01; ***P ＜ 0.001. E. coli: Escherichia coli; HK: high-molecular-weight kininogen

Endogenous expression of human HK in Kng1^–/– restore host defense against E. coli infection

To evaluate the relevance of HK's protective effects in humans, we generate human Kng1 knock-in mice using the CRISPR/Cas9 system by inserting Alb-hKNG1-polyA into H11 locus. Positive knock-in mice (hKng1^ki/ki) were crossed with Kng1^−/− mice to obtain humanized HK mice (hKng1^+/+). As detected by WB, humanized HK mice endogenously expressed human HK protein in the absence of murine HK protein (Figure 4A), and the expression levels of murine PK and FXII were comparable to those in wild-type mice, indicating that the humanization of HK did not alter the baseline expression of other contact system components (Supplemental Figure). Following E. coli infection, hKng1^+/+ mice exhibited significantly improved survival rate compared with Kng1^−/− mice (Figure 4B). Bacterial CFU revealed persistent viable bacteria in plasma, peritoneal fluid, liver and spleen of Kng1^−/− mice, whereas WT and humanized HK mice had almost completely cleared bacteria in these tissues (Figure 4C). In addition, compared with WT mice, Kng1^−/− mice exhibited extensive leukocytes infiltration and widespread destruction in the lung. However, the lung pathology was significantly ameliorated in humanize HK mice (Figure 4D). These findings demonstrate that plasma HK provides a clear survival advantage and that human HK can effectively restore host defense in HK-deficient mice.

Figure 4.

Endogenous expression of human HK inhibits bacterial outgrowth and showed a prolonged survival. (A) Western Blot analysis of mouse plasma with a rabbit polyclonal antibody raised against a peptide derived from the human HK “CWKSVSEINPTTQMK” (anti-human HK), which exclusively target human HK protein. Immunoblotting with an anti-BK antibody served as the loading control. (B) Survival was measured in Kng1^+/+, Kng1^−/− and hKng1^+/+ mice (n = 10 per group) after intraperitoneal infection with 1 × 10⁸ CFU E. coli. *P < 0.05. (C) Bacterial burden 16 h after infection (n = 6): CFUs were measured in the peritoneum, blood, spleen and liver. *P ＜ 0.05; **P ＜ 0.01; ***P ＜ 0.001. (D) Representative histological images of H&E-stained lung sections from mice 16 h after infection. E. coli: Escherichia coli; HK: high-molecular-weight kininogen.

Binding of plasma HK to E. coli accelerated neutrophil reactive oxygen species production and microbicidal activity

To further investigate how HK supports host defense, we analyzed neutrophil function, as the early mortality observed in Kng1^−/− mice suggested defects in innate immunity. Neutrophils kill pathogens primarily through reactive oxygen species (ROS) generation; therefore, we compared ROS production across genotypes. Upon E. coli challenge, neutrophils from Kng1^+/+ mice produced robust ROS, whereas Kng1^−/− neutrophils showed markedly reduced ROS generation. In contrast, neutrophils from humanized HK mice exhibited ROS production comparable to Kng1^+/+ mice (Figure 5A). In line with ROS level, Kng1^−/− mice displayed a reduced capacity of neutrophils to kill invading bacteria (Figure 5B). To confirm the plasma HK ability on neutrophil ROS production, neutrophils from Kng1^+/+ and Kng1^−/− mice were incubated with plasma from different genotypes. Neutrophils from Kng1^−/− mice failed to respond to E. coli in the absence of plasma HK, but responded normally when supplemented with plasma from WT or humanized HK mice (Figure 5C). Despite having the inherent ability to respond to bacteria, Kng1 ^−/− neutrophils fail to do so due to a lack of soluble HK in their plasma. Thus, plasma HK increases neutrophils ROS production and bactericidal activity, thereby promoting efficient infection control.

Figure 5.

Plasma HK accelerated neutrophil ROS production in response to E. coli. (A) DCFH-DA assay was performed to measure ROS production by neutrophils in whole blood from Kng1^+/+, Kng1^−/− and hKng1^+/+ mice after 15 min incubation with E. coli. *P ＜ 0.05; **P ＜ 0.01. (B) Blood neutrophils from Kng1^+/+, Kng1^−/− and hKng1^+/+ mice were incubated for 30 min with E. coli in the presence of 10% homologous plasma. Cells were washed and incubated with gentamicin (500 μg/mL) for 90 min to kill extracellular bacteria. Lysates were then plated to count viable bacteria. *P ＜ 0.05; **P ＜ 0.01. (C) ROS production of neutrophils form Kng1^+/+ and Kng1^−/− mice incubated with E. coli in the presence of serum from Kng1^+/+, Kng1^−/− and hKng1^+/+ mice. *P ＜ 0.05; **P ＜ 0.01; ***P ＜ 0.001. E. coli: Escherichia coli; ROS: reactive oxygen species.

Plasma HK deficiency affects neutrophils responses to E. coli infection

To directly confirm the protection role of plasma HK, we generated liver-specific Kng1 deficiency mice (Kng1^Alb−/−) and determined their susceptibility to E. coli infection. As expected, absence of expression of the HK antigen were observed in plasma from Kng1^Alb−/− mice (Figure 6A). Kng1^Alb−/− mice exhibited worsened survival rate compared to that of Kng1^f/f control mice, which was comparable to that of Kng1^−/− mice (Figure 6B). Neutrophils from Kng1^Alb−/− mice exhibited attenuated ROS production in response to E. coli (Figure 6C), confirming impaired microbicidal activity. These results indicating that the absence of plasma HK compromises the ability of neutrophils to sense and respond to bacteria during the early phase of infection, leading to uncontrolled bacterial proliferation and fatal sepsis.

Figure 6.

Effects of plasma HK in host responses to E. coli infection. (A) Western Blot analysis of plasma from Kng1^f/f and Kng1^Alb−/− mice with an anti-HK antibody. (B) Survival of Kng1^f/f and Kng1^Alb−/− mice (n = 7 pre group) challenged with 1 × 10⁸ CFU E. coli. ^**P ＜ 0.01. (C) ROS production by Kng1^f/f and Kng1^Alb−/− mice neutrophils after incubation with E. coli. ^**P ＜ 0.01. E. coli: Escherichia coli; ROS: reactive oxygen species; HK: high-molecular-weight kininogen.

Discussion

The current study demonstrates the importance of HK in protecting innate immune response against a Gram-negative infection. Our data show that HK limits systemic bacterial dissemination and inflammation by facilitating neutrophil-mediated microbicidal activity. Mechanistically, plasma HK facilitates the effective generation of ROS in neutrophils during E. coli infection, thereby initiating antimicrobial defense mechanisms.

Our results differ from previous studies in which depletion of HK using antisense oligonucleotides (ASOs) reduced inflammation and bacterial spreading during Streptococccus pyogenes sepsis ²⁸ or where HK deficiency did not affect disease outcome in K. pneumoniae induced pneumonia sepsis. ³² These discrepancies can, at least partially, be explained by differences in bacterial species and the vigor of the induced innate immune response. The E. coli strain used in our study (ATCC 25922) does not produce verotoxin, ³³ and its LD50 is approximately 1 0⁸ CFU. When administered at a non-lethal dose, E. coli is likely to be cleared by complement mediated bacteriolysis or neutrophils killing, processes that require moderate rather than excessive inflammation. In contrast, S. pyogenes and K. pneumoniae are highly virulent and induce a massive proinflammatory cytokines release, which likely overwhelms the protective effects of HK.

In fact, upon exposure to E. coli, bacteria multiplied dramatically in HK-deficient mice, and mice died of overwhelming infection. Thus, HK appears to play a protective role against less virulent E. coli infection, but had no protective effect against virulent K. pneumoniae and, in some cases, exacerbated the outcome when injected with S. pyogenes due to release of BK from HK. The contribution of HK, therefore, seems to depend on the mode and intensity of activation by different infectious agents, analogous to the distinct roles reported for other coagulation factors, such as the tissue factor (TF) and factor XI (FXI) pathways, which exert strain-specific effects in infection models.^34–38 Collectively, these results suggest that HK plays a role in the elimination of avirulent or low numbers of less virulent bacteria through specific mechanisms.

To clarify HK involvement during Gram-negative infection, we first assessed its local and systemic level in E. coli-infected mice. Early and marked recruitment of HK to the infection sites (within 3 h) indicated that HK rapidly responds to invading bacteria. Conversely, plasma HK levels declined significantly by 24 h post-infection, accompanied by cleavage into degradation fragments detected by anti-human HK domain 5 antibody. Previous clinical studies have shown that systemic contact activation and HK consumption occur frequently in patients with severe sepsis,^6,8,39 and low levels of HK during sepsis associate with poor outcomes.^5,28 These observations, together with our current findings, support the concept that HK behaves as an acute phase protein and could be part of the host defenses against infection.

To effectively respond a wide spectrum of microbes, the innate immune system employs an array of proteins to identify distinctive surface features of microbial pathogens. Earlier reports have reported that more than 20 bacterial species bind and active contact factors on their surfaces. ²⁰ Previously, we demonstrated that HK associates with different LPS species with high affinity through LPS carbohydrate region. In the present study, we further show that HK binds directly and efficiently to E. coli via LPS. In addition to LPS, other bacterial surface structures-such as curli fibers from E. coli or Salmonella enterica and fimbriae from Porphyromonas gingivalis have also been reported to support HK binding.^40–42 Notably, both curli and fimbriae are outer membrane protein and belong to a category of structured and stable proteins known as bacterial amyloids.^43,44 The role of bacterial surface molecules in interaction of HK with other Gram-negative bacteria remains to be ascertained. Interestingly, the major plasma LPS-binding protein does not bind most live bacteria, ⁴⁵ likely because it LPS via lipid A, which is embedded within the bacterial lipid bilayer, unlike the carbohydrate region exposed on the bacterial surface. ⁴⁶ Our data therefore identify HK as a pattern recognition receptor capable of recognizing exogenous LPS structures and specific surface proteins on Gram-negative bacteria. This recognition may opsonize whole bacteria to be readily eliminated, thereby acting as a direct innate immune mechanism.

Opsonization of pathogens by soluble PRPs is known to modulate phagocytes and antimicrobial functions.^47–51 Here, we show that HK is essential for antimicrobial activity of neutrophils by sustaining adequate ROS production. HK deficiency was associated with increased susceptibility to E. coli infection, as evidenced by infection severity and pathological inflammation. Indeed, higher levels of organ bacterial burden, systemic pro-inflammation cytokines levels, and tissue damage were observed in Kng1^−/− mice compared to wild-type mice. And these phenotypes were largely reversed by the endogenous expression of human HK in plasma, although it did not appear to completely restore to wild-type mice. This partial rescue may indicate that human HK will not completely restore all murine kininogen functions. A recent paper by Mohammed et al. ⁵² showed than Human HK binds less efficiently to murine FXI than to human FXI, resulting in modestly reduced FXI activity in Kng1^–/– mouse plasma supplemented with human HK. Further investigation is warranted to fully elucidate these interspecies functional differences.

HK deficiency is a rare clinical condition, and its pathological implications in humans remain largely unknown. ³¹ large-scale epidemiological studies are needed to further determine whether HK deficiency is associated with increased susceptibility to specific pathogens. Notably, human neutrophils have been shown to contain and bind HK protein, ⁵³ to further determine the protective effect against bacterial infection is directly associated with plasma HK, the liver-specific Kng1 knock-out mice with plasma HK deficiency were employed. Results revealed that plasma HK is indispensable for neutrophils ROS production and bacterial clearance, likely plays a role in the ability to contain the infection. Thus, plasma-derived HK is required for survival benefit during E. coli infection.

In conclusion, our findings identify HK as a protective factor against E. coli-mediated peritonitis and highlight its role as a physiologic component of innate immune defense. During infections with avirulent Gram-negative bacteria, where a strong inflammatory response may not be required, HK enables neutrophil activation and microbicidal ROS production essential for pathogen containment. Despite limitations of any murine sepsis model, our study provides the first direct evidence that HK contributes to host protection during E. coli-mediated peritonitis, suggesting that HK may exert beneficial effects in human Gram-negative peritonitis.