Background
Sepsis is defined as life-threatening acute organ dysfunction secondary to infection in the body and bloodstream that most commonly originates in the lung, urinary tract, and abdomen [
1,
2]. Sepsis progresses when the initial host response fails to control the infection, resulting in widespread inflammation and multiple organ injury [
3]. Sepsis-induced mortality is closely associated with the failure to eradicate invading pathogens [
4]. In a post-mortem study of 235 patients in surgical intensive care who were admitted with sepsis, about 80% of patients had unresolved septic foci at death [
5]. Therefore, new strategies to treat sepsis should boost host immunity, thereby leading to a more rapid resolution of the infection and prevention of death.
Toll-like receptors (TLRs) are central for host defense against pathogenic microorganisms by recognizing conserved motifs in pathogens termed pathogen-associated molecular patterns [
6]. TLR5 is the receptor for extracellular flagellin, which is a component of motile bacteria [
7]. Clinical data have shown that the expression of TLR5 on monocytes predicted systemic inflammatory response syndrome (SIRS) [
8]. Animal data have demonstrated that TLR5 agonist flagellin restored antibiotic-impaired innate immune defenses and restricted colonization with vancomycin-resistant
Enterococcus (VRE) [
9]. However, whether TLR5 plays a role in controlling infection during sepsis has yet to be addressed. In this study, we investigated the role of flagellin-induced TLR5 activation in controlling the infection during sepsis using the cecal ligation and puncture (CLP) model of abdominal sepsis in mice. We also measured TLR5 expression levels in septic patients and analyzed their relationship with clinical phenotypes.
Methods
Patient and healthy control demographics
Patients who met the clinical criteria for Sepsis-3 were screened for eligibility within the first 24 h after they were admitted to the Department of Infectious Diseases of The First Affiliated Hospital of Chongqing Medical University or the Intensive Care Unit of The Second Affiliated Hospital of Chongqing Medical University between January 2017 and February 2018 [
1,
10]. A total of 53 septic adult patients (Additional file
1: Table S1) were enrolled. Patients were included if they had known or suspected infection plus an increase in the Sequential [Sepsis-related] Organ Failure Assessment (SOFA) score of 2 or more points for organ dysfunction. Patients who are pregnant or breast-feeding; patients with malignancy, organ transplantation, chronic viral infections (hepatitis, HIV), liver cirrhosis, chronic kidney insufficiency, and autoimmune diseases; and patients using immunosuppressive medication were excluded from the study. Twenty-three non-septic patients but in critical conditions of trauma injury (poly-trauma or cerebral trauma) were recruited as controls. The clinical data, such as Acute Physiology and Chronic Health Evaluation II (APACHE II) score, SOFA score, causes of sepsis, microbial culture result, length of intensive care unit stay, and mortality, were recorded. There was no difference in end-stage renal failure between septic patients and non-septic patients. Healthy control samples were obtained from 37 healthy donors with no medical problems in the medical examination center of The First Affiliated Hospital of Chongqing Medical University. Nine milliliters of venous peripheral ethylenediaminetetraacetic acid (EDTA) blood was collected at the time of enrollment, and blood samples were also obtained from 9 patients with sepsis within 1 h of death. Aliquots of whole blood were processed immediately for peripheral blood mononuclear cell (PBMC) isolation. PBMCs were prepared by centrifugation of blood using a density gradient (Ficoll-Paque Plus; GE Healthcare Life Sciences). This protocol was approved by the Clinical Research Ethics Committee of Chongqing Medical University, and informed consent was obtained from all participants according to the Declaration of Helsinki.
Sepsis model
Pathogen-free 6–8-week-old female C57BL/6 mice were obtained from and raised at Chongqing Medical University. Polymicrobial sepsis was provoked by CLP as described in our previous studies [
11,
12]. C57BL/6 mice were anesthetized with a mixture of xylazine (4.5 mg/kg) and ketamine (90 mg/kg) intraperitoneally (i.p.). CLP was performed by making a midline incision ~ 2.5 cm in length to expose the cecum. A 3-0 silk ligature was placed at the base of the cecum without causing bowel obstruction. The cecum was then punctured twice with a 21-gauge needle (lethal CLP) or a 26-gauge needle (sublethal CLP). The cecum was then placed back in the peritoneal cavity, and the incision was closed with surgical staples. Sham-operated (control) animals underwent identical laparotomy, the cecum was exposed but not ligated or punctured and was then replaced in the peritoneal cavity. Mice received saline (5 ml per 100 g body weight) subcutaneously for resuscitation, and no antibiotic treatment was used. Survival was monitored twice daily for 14 days. For the
Escherichia coli model, 5 × 10
8 E.
coli was injected intraperitoneally. The animal experiments were approved by the local Animal Care and Use Committee.
Flagellin
For animal experiments, TLR5 ligand flagellin (InVivoGen), derived from Salmonella typhimurium, was used. Flagellin did not contain detectable lipopolysaccharide (LPS), as determined by the Limulus amoebocyte lyase assay (sensitivity limit 12 pg/ml; BioWhittaker Inc., Walkersville, MD). Flagellin was injected i.p. at 2–10 μg/injection at 2–8 h after CLP, and PBS was delivered in a similar fashion as control solutions.
Treatment with anti-TLR5 antibody and recombinant mouse TLR5 Fc Chimera
TLR5 neutralization was performed by i.p. administration of 50 μg of rat anti-mouse TLR5 monoclonal antibody (InVivoGen, clone: Q23D11) at 2 h before CLP. The neutralization activity has been determined for its ability to inhibit flagellin-induced NF-kB activation in HEK-293 cells transfected to express mouse TLR5. Rat IgG2a was used as isotypical IgG antibody. In some experiments, TLR5 inhibition was also performed by i.p. administration of 50 μg recombinant mouse TLR5 Fc Chimera (R&D systems) at 2 h before CLP. The inhibition effect has been confirmed for its ability to inhibit flagellin-induced IL-8 secretion in HT29 human colon adenocarcinoma cells.
Statistics
Data were expressed as scatter dot plots with medians unless otherwise specified. Comparisons between groups were tested using Student’s t test, Mann–Whitney U test, or Kruskal–Wallis test followed by Dunn’s multiple comparisons post-test as appropriate. Correlations were tested by Spearman’s rank correlation test. For survival studies, Kaplan–Meier analyses followed by log-rank tests were performed. All analyses were done using GraphPad Prism version 5.01 (GraphPad Software, San Diego, CA). p values less than 0.05 were considered statistically significant.
Additional details on the methods are available in the online supplement.
Discussion
In the present study, we demonstrated that the administration of flagellin effectively prevented the progression of sepsis in the CLP polymicrobial sepsis model in a macrophage-dependent manner. Flagellin activation of macrophages promoted the protective immunity against bacterial infection and improved survival in polymicrobial sepsis. We also clearly demonstrated that the therapeutic effects of flagellin required TLR5, and TLR5 deletion could abolish the beneficial effects of flagellin on sepsis. Therefore, flagellin acted through TLR5 to elicit antiseptic activity.
A number of studies have demonstrated the effectiveness of flagellin as an adjuvant for controlling infectious diseases [
15,
16]. A recent study has demonstrated that pretreatment of the mice with flagellin 4 h before CLP challenge significantly decreased the sepsis-induced lethality [
16]. To assess the therapeutic benefits of flagellin administration at a time when patients are more likely to be treated, flagellin was administered 2–8 h after CLP in this study according to our previous work and others [
11,
17,
18]. Importantly, the ability of flagellin administration after the onset of sepsis to significantly improve survival in mice suggests that flagellin is an effective rescue therapy. These results are consistent with a previous study which has demonstrated that administering flagellin to antibiotic-treated mice before VRE infection can reduce VRE colonization [
9].
We and others have shown that monocytes/macrophages can play a pivotal role in the resolution of sepsis [
11‐
13,
19,
20]. We have demonstrated that progranulin could protect sepsis by promoting macrophage recruitment [
11], while others have shown that mast cells aggravated sepsis by inhibiting macrophage phagocytosis [
20]. Furthermore, DJ-1, a well-established ROS scavenger, could impair ROS production for bacterial killing by macrophages, and DJ-1-deficient mice had improved bacterial clearance, reduced organ injury, and increased survival in CLP-induced polymicrobial sepsis compared with wide-type mice [
21]. In contrast, sphingosine-1-phosphate receptor 3 (S1PR3) signaling was essential for ROS production and phagolysosomal maturation, which mediated bacterial killing in macrophages. Enhancing endogenous S1PR3 activity using a peptide agonist potentiated ROS production and bactericidal function in macrophages, resulting in decreased bacterial burden, less tissue injury, and improved survival rates in multiple models of sepsis [
13]. In this study, we found that depleting macrophages could dramatically impair the survival and bacterial clearance of septic mice treated with flagellin, while adoptive transfer of flagellin-activated macrophages could protect mice against lethal sepsis. Furthermore, flagellin could directly enhance phagocytic function by promoting phagosome formation and bacterial killing by increasing ROS production in macrophages. These data indicate that flagellin administration is a viable therapeutic modality in sepsis by upregulating antimicrobial activity in macrophages. However, the improved outcomes after flagellin treatment of sepsis were not mediated by neutrophils. This result is consistent with a report by Lu et al., which showed macrophages, but not neutrophils, mediated the beneficial effect of leukocyte cell-derived chemotaxin 2 (LECT2) on bacterial sepsis [
19], but is at odds with a report by Muñoz et al., which showed that neutrophils were required for flagellin-elicited protection against
Streptococcus pneumoniae lung infection [
22]. One potential explanation for this apparent discrepancy may be the different animal models. We used a CLP-induced sepsis model, whereas Muñoz et al. used a
Streptococcus pneumoniae pneumonia model. Another explanation for this discrepancy is likely due to the use of different routes of flagellin administration. We used a systemic administration route (intraperitoneal injection of flagellin), whereas Muñoz et al. used a locally mucosal administration route (intranasal injection of flagellin). In addition, a recent study has also shown that endotoxin preconditioning could confer renal epithelial protection in various models of sepsis, which was mediated by macrophages [
23].
A recent study has shown that flagellin could suppress experimental asthma by generating regulatory T cells in mice [
24]. In this lethal CLP model, there was no significant difference in the regulatory T cell population between mice treated with flagellin and PBS control (data not shown). We acknowledge that the contribution of flagellin-mediated effects on regulatory T cells to long-term immunosuppression in sepsis should be investigated using another sublethal model of CLP followed by a secondary infection with external pathogens [
25,
26]. For example, IL-33 has been shown to attenuate sepsis by enhancing neutrophil influx to the site of infection in the lethal CLP model [
4]. In the sublethal CLP model followed by
Legionella pneumophila infection, IL-33 also contributed to sepsis-induced long-term immunosuppression by expanding the regulatory T cell population [
26]. Furthermore, TLR5 is also expressed on other innate cells and organ epithelium and endothelium cells [
6,
7]; the regulatory effects of flagellin on these cells in sepsis should be studied in the future work, which is beyond our present study.
In the development of adjuvant therapies for patients with sepsis, it is imperative that these findings from animal studies should be translated into humans. A recent study has demonstrated that the expression of TLR5 on monocytes on day 1 or 2 could predict SIRS after major abdominal surgery [
8]. Here, we further found that TLR5 expression on peripheral monocytes was significantly upregulated in septic patients when compared with healthy individuals. Importantly, we documented for the first time that those septic patients who did not survive had significantly higher expression of TLR5 on circulating monocytes than did the survivors. Our observation that higher expression levels of TLR5 on monocytes were associated with poorer outcomes and an increase in the incidence of bacteremia of septic patients suggests that monocyte TLR5 expression might be a potential indicator of immune dysfunction and mortality. Furthermore, monocytes/macrophages from patients who died of sepsis demonstrated reduced phagocytic activity and bacterial killing ability, while treatment of septic monocytes/macrophages with flagellin restored their phagocytosis and bacterial killing ability as observed in healthy individuals. These data suggest that measuring human monocyte TLR5 expression may guide the use of flagellin to contain the infection and improve survival in septic patients. However, we acknowledge that the clinical value of monocyte TLR5 expression in septic patients should be further validated in a larger size clinical trial.
Acknowledgements
The authors thank Dr. Fang Xu from Department of Emergency and Critical Care Medicine, The First Affiliated Hospital of Chongqing Medical University, Chongqing, China, for expert technical assistance and statistical analysis support.