FGF8 Protects Against Polymicrobial Sepsis by Enhancing the Host's Anti-infective Immunity

Abstract

Background

Sepsis is characterized by a life-threatening syndrome caused by an unbalanced host response to infection. Fibroblast growth factor 8 (FGF8) has been newly identified to play important roles in inflammation and innate immunity, but its role in host response to sepsis is undefined.

Methods

A cecal ligation and puncture (CLP)-induced mouse sepsis model was established to evaluate the immunomodulatory function of FGF8 during sepsis. The underlying molecular mechanisms were elucidated by cell models using relevant molecular biology experiments. The clinical value of FGF8 in the adjuvant diagnosis of sepsis was evaluated using clinical samples.

Results

FGF8 protein concentrations were elevated in CLP-induced septic mice compared to controls. In vivo, FGF8 blockade using anti-FGF8 antibody significantly increased mortality and bacterial burden and was paralleled by significantly aggravated tissue injury after CLP. Therapeutic administration of recombinant FGF8 (rFGF8) improved the bacterial clearance and mortality of septic mice in a FGFR1-dependent manner. In vitro, FGF8 directly enhanced bacterial phagocytosis and killing of macrophages by enhancing the phosphorylation of the ERK1/2 signaling pathway, which could be abrogated with the ERK1/2 pathway inhibitor U0126. Clinically, serum FGF8 levels in both adult and pediatric patients with sepsis in an intensive care unit were significantly higher than those in healthy controls.

Conclusions

These results present a previously unrecognized role of FGF8 in improving survival of sepsis by enhancing host immune defense. Therefore, targeting FGF8 may provide new strategies for the diagnosis and immunotherapy of sepsis.

The underlying mechanism for high mortality associated with sepsis remains to be resolved. These results present a previously unrecognized role of FGF8 in improving survival of sepsis by enhancing host immune defense, providing novel strategies for sepsis diagnosis and immunotherapy.

Sepsis is a pathological condition caused by an unregulated response to pathogen invasion [1] and is the leading cause of death worldwide in cases of unsuccessful infection treatment [2]. Early in sepsis, the inability of innate immunity to eradicate the invading pathogen leads to immunosuppression and an altered inflammatory response that can result in increased risk of secondary infection and multiple organ failure and death [3]. Postmortem analyses have indicated that most patients with sepsis admitted to intensive care units (ICU) had unresolved septic foci suggesting a compromised host immune system is strongly associated with sepsis-related death [4]. The identification of host factors driving the unbalanced immune response in sepsis constitutes an important area of research that could offer effective treatment strategies.

Fibroblast growth factors (FGF) are circulating protein ligands that operate in a paracrine or endocrine manner to perform pleiotropic activities in development, tissue homeostasis, and metabolism [5, 6]. In particular, FGF8 plays key regulatory roles in cell growth [7] and survival [8–10], as well as immune regulation [11], via cognate receptor interactions with FGF receptors (FGFR1–4). FGF8 levels in healthy adults are maintained at a low level but are elevated during inflammation [12] and this FGF8 upregulation can suppress apoptosis in multiple tissues to thereby preserve normal organ function. This process could potentially serve a protective role in mitigating tissue damage in sepsis [13–15]. However, the mechanism of FGF8 signaling in bacterial clearance and immune regulation in sepsis remains unknown.

In this study, we examined the effects of FGF8 on systemic infection and immune regulation from the perspective of anti-infective immune defense, using a mouse model of polymicrobial sepsis induced by cecal ligation puncture (CLP) in vivo and a primary mouse macrophage model in vitro.

METHODS

Study Populations

Adult patients with sepsis and pediatric patients with sepsis were recruited from the First Affiliated Hospital of Chongqing Medical University (Chongqing, China) ICU and the Children's Hospital of Chongqing Medical University (Chongqing, China), respectively, and the corresponding control groups were obtained from healthy physical examination subjects. All patients met the clinical criteria of Sepsis-3 [1]. Clinical and demographic characteristics of all patients were recorded (Supplementary Tables 1 and 2). Adult patients were included if they had known or suspected infection plus an increase in the sequential (sepsis-related) organ failure assessment (SOFA) score ≥ 2 points for organ dysfunction. Similarly, pediatric patients with infection combined with Phoenix sepsis score (PSS) ≥ 2 were included in the study. Patients who were pregnant or breastfeeding, who had malignancy, organ transplantation, chronic viral infections (hepatitis, HIV), autoimmune diseases, or using immunosuppressive medication were excluded from the study. Healthy blood donors were age- and sex-matched to patients with sepsis. Plasma was isolated from healthy donors and patients with sepsis and stored at −80°C until analysis.

The experimental protocol was reviewed and approved by the Clinical Research Ethics Committee of The First Affiliated Hospital of Chongqing Medical University (No. K2023-384) and the Medical Research Ethics Committee of Children's Hospital of Chongqing Medical University (No. 2024-278). The study was conducted in accordance with the principles of the Helsinki Declaration. No informed consent was needed as the blood samples used in this study were all residual blood samples discarded after routine examination. Nevertheless, nonopposition to inclusion in the protocol was recorded from every patient or their guardian.

Animals and Polymicrobial Sepsis Model

C57BL/6 (wild type) mice (age, 6–8 weeks) were obtained from and raised at the Laboratory Animal Center of Chongqing Medical University. All animal studies were approved by the Ethics Committee of Chongqing Medical University, and all animal experiments were performed in accordance with the guidance of Chongqing Medical University Laboratory Animal Management and Use Committee (No. IACUC-CQMU-2024–0494), and also meet the requirements of “Chongqing Laboratory Animal Management Regulations” and other relevant laws and regulations.

CLP-induced polymicrobial sepsis model was established as described previously [16, 17]. In brief, C57BL/6 mice were anesthetized intraperitoneally (IP) with a mixture of xylazine (4.5 mg/kg) and ketamine (90 mg/kg). A 1-cm longitudinal incision was made at the median abdominal line, and then the cecum was exposed, ligatured at its external third, and punctured through with a 21-gauge needle (severe CLP, resulting in 0% to 20% survival) or with a 24-gauge needle (resulting in 50% to 60% survival). The cecum was then returned to the peritoneal cavity, and incisions were closed. Sham-operated (control) animals underwent identical laparotomy; the cecum was exposed, but not ligated or punctured, and was then replaced in the peritoneal cavity. Following surgery, mice were injected subcutaneously with 1 mL of Ringer's solution, including analgesia (buprenorphine, 0.05 mg/kg). Animals were placed in a temperature-controlled incubator and monitored every 6 hours for the first 48 hours, and then every 8 hours until the end of the experiments.

Treatment of Mice With rFGF8, Anti-FGF8, and Inhibitors

Recombinant murine FGF8 (rFGF8; 423-F8; R&D Systems) was administered IP to mice at 5 and 12.5 μg/kg at the indicated times following CLP and phosphate-buffered saline (PBS) was administered to controls in the same manner. FGF8 blockade was performed by IP administration of 25 μg/kg of mouse monoclonal anti-FGF8 antibody (MAB323, R&D Systems) immediately after CLP. Mouse immunoglobulin G 2b (IgG2b; 360107, R&D Systems) was used as isotype control. For in vivo blocking of FGFR1 and ERK1/2, PD173074 (1 mg/kg; HY-10321, MedChem Express) and U0126 (10 mg/kg; HY-12031A, MedChem Express) were dissolved in dimethyl sulfoxide (DMSO) solution (10% DMSO + 90% corn oil) and administrated immediately following CLP, respectively.

Statistical Analysis

Nonparametric Mann-Whitney U test was used for comparisons between 2 groups and the Kruskal-Wallis test followed by the Dunn test were used for multiple comparisons test to determine statistical significance. For survival studies, Kaplan-Meier analyses were performed followed by the log-rank test. All analyses were done using GraphPad Prism 8.0. P < .05 was considered statistically significant.

RESULTS

Kinetics of FGF8 Production in Polymicrobial Sepsis

In vivo data from the CLP-induced polymicrobial sepsis model indicated that the septic mice possessed significant elevations of FGF8 in the serum, peritoneal lavage fluid (PLF), lung, spleen, and kidneys compared to the controls (Figure 1A). The presence of FGF8 in septic mice was also directly demonstrated in kidney, spleen, and lung using immunofluorescence, and levels were increased over that of control tissues (Figure 1B). In addition, we found no significant increase in FGF8 in the liver even though this organ is the primary FGF8 source under normal conditions (Supplementary Figure 1A and 1B). Interestingly, defects in Toll-like receptor 2 (TLR2) and Toll-like receptor 4 (TLR4) did not affect sepsis-induced FGF8 production and indicated that the TLR2/TLR4 signaling pathway was not required for FGF8 induction during polymicrobial sepsis (Figure 1C). To identify the cells producing FGF8 during sepsis, we investigated the expression of FGF8 in various primary cells following bacterial stimulation. FGF8 was substantially elevated in macrophages (Figure 1D) but not in neutrophils, lymphocytes, or epithelial cells (Supplementary Figure 1C).

Figure 1.

Kinetics of FGF8 production in polymicrobial sepsis. A, Male C57BL/6 mice (n = 5 per group) were subjected to sham or CLP using a 24-gauge needle. PLF, serum, lung, spleen and kidney were collected at the indicated times (6, 24, 48 hours) after CLP. FGF8 concentrations were measured by ELISA. B, Representative fluorescence images of FGF8 expression in kidney, spleen, and lung after CLP. Scale bar = 50 μm. Quantitative results are shown (n = 3 per group). C, FGF8 concentrations in serum and PLF of TLR2^−/−, TLR4^−/−, TLR2/4^−/−, and WT mice 24 hours (n = 3–4 per group) after CLP. D, Supernatant of heat-killed Pseudomonas aeruginosa (MOI = 1:100) challenged macrophages was collected at the indicated times (6, 12, 24 hours). FGF8 levels were measured by ELISA (n = 4 per group). A–D, Data are representative of 3 independent experiments; Kruskal-Wallis test followed by Dunn multiple comparisons posttest. *P < .05, **P < .01, ***P < .001, ****P < .0001 compared within 2 groups. Abbreviations: CLP, cecal ligation and puncture; Ctrl, control; ELISA, enzyme-linked immunosorbent assay; FGF, fibroblast growth factor; MOI, multiplicity of infection; ns, not significant; PLF, peritoneal lavage fluid; TLR, Toll-like receptor; WT, wild type.

Inhibition of Endogenous FGF8 Exacerbated Polymicrobial Sepsis

These initial results indicated that FGF8 production was increased in polymicrobial sepsis so we next examined whether systemic neutralization of FGF8 would impact the immunopathology of sepsis. The experimental mice were treated with the anti-FGF8 antibody immediately following CLP and this treatment increased mortality compared with mice treated with control IgG (Figure 2A). Consistent with these results, increased bacterial burdens were found in the liver, PLF, lung, spleen, and blood of the experimental mice compared to the controls (Figure 2B). Histological analyses also indicated that the anti-FGF8–treated mice possessed more severe pathological changes when compared with controls (Figure 2C). Serum levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST), lactate dehydrogenase (LDH), creatinine, and urea in anti-FGF8–treated mice were also significantly higher than in controls (Figure 2D). Furthermore, IL-6 levels in PLF and serum of anti-FGF8–treated mice were also significantly increased compared with controls. In contrast, we found no differences in CCL2, CXCL1, interleukin 1β (IL-1β), IL-17A, or tumor necrosis factor-α (TNF-α) in serum and PLF of septic mice treated with anti-FGF8 and the IgG controls (Supplementary Figure 2).

Figure 2.

Inhibition of endogenous FGF8 exacerbated polymicrobial sepsis. A, Survival of septic mice (n = 20 per group) with or without FGF8 neutralization after CLP using a 24-gauge needle. B, Dilutions of PLF, blood, liver, lung, and spleen tissues were obtained from septic mice (n = 10–11 per group) treated with or without monoclonal anti-FGF8 antibody at 24 hours after CLP; samples were cultured on blood agar plates and the numbers of bacterial colonies were then determined. C, Representative examples of hematoxylin and eosin-stained tissues as indicated from mice (n = 5 per group) treated or not treated with monoclonal anti-FGF8 antibody at 24 hours after CLP. The pathology scores are shown on the right side of histological images. D, Serological markers of organ injury including ALT, AST, LDH, creatinine, and urea in septic mice (n = 11–12 per group) treated with or without monoclonal anti-FGF8 antibody at 24 hours after CLP. Each dot represents an individual mouse. B–D, Nonparametric Mann-Whitney U test; (A) Kaplan-Meier analysis followed by log-rank test. *P < .05, **P < .01, ***P < .001 compared within 2 groups. Abbreviations: ALT, alanine transaminase; AST, aspartate transaminase; CLP, cecal ligation and puncture; FGF, fibroblast growth factor; IgG, immunoglobulin G; LDH, lactate dehydrogenase; ns, not significant; PLF, peritoneal lavage fluid.

rFGF8 Treatment Improves Outcomes of Mice With Sepsis

The adverse effects of anti-FGF8 treatment on sepsis in animal models led us to consider using recombinant FGF8 protein (rFGF8) for reverse validation. As expected, in a severe sepsis model, mice treated with rFGF8 (5 and 12.5 μg/kg) had significantly improved survival compared to PBS-treated controls (Figure 3A). Moreover, FGF8 protected against CLP-induced mortality in a sex-independent manner (Figure 3A). Furthermore, effective and rapid bacterial clearance is a fundamental determinant of outcomes in sepsis [18–20]. In our experiments, we observed significant reductions in bacterial loads in the blood, lungs, PLF, and spleens of septic mice treated with rFGF8 (Figure 3B). Besides, we evaluated the role of FGF8 in single-bacterial sepsis induced by intraperitoneal injection of Staphylococcus aureus (a gram-positive bacterium) and Pseudomonas aeruginosa (a gram-negative bacterium). Data showed that rFGF8 treatment reduced bacterial load in the blood and PLF of mice with sepsis (Supplementary Figure 3A and 3B) and protected the mice from sepsis-mediated death (Supplementary Figure 3C and 3D). Moreover, we also evaluated the effect of FGF8 on severe bacterial pneumonia induced by S. aureus and P. aeruginosa. As shown in Supplementary Figure 3E–H, rFGF8 treatment significantly reduced the bacterial load in blood and lung tissue of mice (Supplementary Figure 3E and 3F), thereby improving bacterial pneumonia-related death (Supplementary Figure 3G and 3H).

Figure 3.

FGF8 treatment improves outcomes in a murine sepsis model. A, Different levels of rFGF8 (0, 5, 12.5 μg/kg) were injected intraperitoneally into mice (n = 20 per group) immediately after CLP and survival was monitored for 14 days. B, Dilutions of blood, lungs, PLF, and spleens obtained from septic mice (n = 9–10 per group) treated with or without rFGF8 (12.5 μg/kg) 48 hours after CLP. C, Representative examples of hematoxylin and eosin-stained lung, liver, spleen, and kidney tissues from CLP mice treated with or without rFGF8 (12.5 μg/kg) after CLP. Scale bars = 400 μm. Histological scores (n = 5 per group) are shown. D, Serological markers of organ injury in septic mice (n = 13 per group) treated with or without rFGF8 (12.5 μg/kg) at 48 hours after CLP. B–D, Data are representatives of 3 independent experiments; nonparametric Mann-Whitney U test; (A) Kaplan-Meier analysis followed by log-rank test. *P < .05, **P < .01 compared within 2 groups. Abbreviations: ALT, alanine transaminase; AST, aspartate transaminase; CLP, cecal ligation and puncture; FGF, fibroblast growth factor; LDH, lactate dehydrogenase; ns, not significant; PBS, phosphate-buffered saline; PLF, peritoneal lavage fluid.

Tissue inflammation and damage as assessed by hematoxylin and eosin staining was as expected and rFGF8 treatment alleviated tissue damage in the lungs, livers, kidneys, and spleens of septic mice and was accompanied by reduced histopathological scores (Figure 3C). Serum levels of ADH, ALT, AST, and creatinine were also reduced in mice treated with rFGF8 (Figure 3D). In addition, rFGF8 treatment resulted in decreased IL-6 levels in PLF and serum 6 hours following CLP, although no significant differences were found at 24 hours. Similarly, no significant differences were observed in the levels of TNF-α, IL-1β, IL-10, IL-17A, HMGB1, CCL2, or CXCL1 between rFGF8- and PBS-treated mice at 6 and 24 hours after CLP (Supplementary Table 3). To further clarify the role of FGF8 in alleviating tissue damage and hyperinflammation, an LPS-induced endotoxemia model was established. Compared with the control group, rFGF8 treatment significantly improved the survival rate of LPS-challenged mice (Supplementary Figure 4A), and correspondingly decreased their serum concentrations of IL-6 (Supplementary Figure 4B) and injury-related serological markers (Supplementary Figure 4C).

FGF8 Enhances Bacterial Phagocytosis and Killing by Macrophages

Macrophage chemotaxis to infection sites [17, 21] and phagocytosis and pathogen killing are key to the elimination of bacteria by macrophages during sepsis [22, 23]. We therefore examined whether FGF8 altered these antibacterial functions of macrophages in vivo and in vitro. We found no significant differences in the numbers of F4/80⁺ macrophages in the peritoneum of rFGF8-treated mice versus controls (Supplementary Figure 5). However, rFGF8-treated macrophages displayed significantly enhanced phagocytosis against bacteria compared to controls (Figure 4A). Furthermore, the bactericidal activity of rFGF8-treated macrophages was also significantly increased compared to the controls (Figure 4B).

Figure 4.

FGF8 directly enhances bacterial phagocytosis and killing by macrophages. A, Peritoneal macrophages (1 × 10⁶ cells) were stimulated with or without rFGF8 (200 ng/mL) for 6 hours and then challenged with FITC-labeled Pseudomonas aeruginosa (MOI = 1:100) for 30 minutes at 37°C. Representative images from 3 independent experiments are shown. Dot plot depicts macrophage phagocytosis levels (n = 5 per group). B, Bacterial killing of P. aeruginosa in peritoneal macrophages (5 × 10⁵ cells) treated with PBS or the indicated doses of rFGF8. Dot plot depicts macrophage killing (n = 4 per group). C, Mortality of rFGF8-treated (12.5 μg/kg) septic mice in the presence or absence of macrophage depletion after CLP (n = 10 per group). D, Bacterial loads in PLF and blood of rFGF8-treated (12.5 μg/kg) septic mice in the presence or absence of macrophage depletion after CLP (n = 5 per group). E, Survival after transfer of rFGF8- or PBS-treated peritoneal macrophages in mice (n = 12 per group) after CLP. A, B, and D, Data are representatives of 3 independent experiments; (A) nonparametric Mann-Whitney U test; (B and D) Kruskal-Wallis test followed by Dunn multiple comparisons posttest; (C and E) Kaplan-Meier analysis followed by log-rank test. *P < .05, **P < .01, ***P < .001, ****P < .0001 compared within 2 groups. Abbreviations: CLP, cecal ligation and puncture; CFU, colony-forming unit; DAPI, 4′,6-diamidino-2-phenylindole; FITC, fluorescein isothiocyanate; MOI, multiplicity of infection; ns, not significant; PBS, phosphate-buffered saline; rFGF, recombinant fibroblast growth factor; TRITC, tetraethyl rhodamine isothiocyanate.

It is particularly worth mentioning that no direct antibacterial activity of FGF8 itself was observed by dynamically monitoring the effect of rFGF8 on bacterial growth (Supplementary Figure 6A and 6B). Thus, we hypothesized that macrophages might mediate the protective effects of FGF8 against sepsis. We therefore depleting macrophages using clodronate liposomes (Supplementary Figure 7) and this procedure abrogated the beneficial effects of rFGF8 on survival (Figure 4C) and bacterial clearance (Figure 4D) in the septic mice. Furthermore, survival of septic mice could be significantly enhanced by adoptive transfer of rFGF8-treated macrophages (Figure 4E). These effects were limited to the macrophage response because FGF8 did not alter bacterial phagocytosis and killing by neutrophils (Supplementary Figure 8A–C) nor did it alter neutrophil recruitment to the infection site (Supplementary Figure 8D).

FGFR1 Plays a Critical Role in FGF8-Induced Protection Against Experimental Sepsis

There are 4 described FGF8 receptors (FGFR1, 2, 3, 4) [7, 24–26] and a STRING database (https://string-db.org/) analysis indicated that FGF8 interacts with FGFR1–4 in both humans and mice (Supplementary Figure 9A and 9B). However, FGF8 immunofluorescence of tissues indicated that FGF8 was primarily colocalized with FGFR1 in bacteria-infected macrophages (Figure 5A). Further, the Human Expression Data in Immunological Genome Project (ImmGen, https://www.immgen.org) indicated that FGFR1 and FGFR2, but not FGFR3 and FGFR4, were expressed in CD14⁺ monocytes (Supplementary Figure 10). However, FGF8 has a stronger affinity with FGFR1 than FGFR2 [7]. We found that rFGF8 induced the phosphorylation of FGFR1 in a dose-dependent manner (Figure 5B and Supplementary Figure 11A).

Figure 5.

FGFR1 plays a critical role in FGF8-induced protection against experimental sepsis. A, Representative confocal images of the colocalization of FGF8 (Cy3) and FGFR (FITC) in peritoneal macrophages treated with rFGF8 (200 ng/mL). Scale bar = 20 μm. B, Peritoneal macrophages were pretreated with or without rFGF8 (200 ng/mL) followed by incubation with or without heat-inactivated Pseudomonas aeruginosa. Representative fluorescence images of phospho-FGFR1 expression are shown. Scale bar = 25 μm. C, Peritoneal macrophages (n = 4 per group) were pretreated with or without the FGFR1 inhibitor PD173074 (100 nM) for 1 hour followed by incubation with or without rFGF8 (200 ng/mL). In vitro bacterial phagocytosis and killing of P. aeruginosa. D, Mortality after blocking FGFR1 with PD173074 (1 mg/kg) and subsequent treatment with rFGF8 (12.5 μg/kg) or PBS control after CLP (n = 15 per group). Except for survival (D), data are presented as means and are representative of 3 independent experiments; (C) Kruskal-Wallis test followed by Dunn multiple comparisons posttest; (D) Kaplan-Meier analysis followed by log-rank test. *P < .05, ***P < .001, ****P < .0001 compared within 2 groups. Abbreviations: CLP, cecal ligation and puncture; FGFR, FGF receptor; ns, not significant; rFGF, recombinant fibroblast growth factor; FITC, Fluorescein Isothiocyanate; Cy3, Cyanine 3.

To verify whether FGF8 regulates the phagocytotic and killing function of macrophages through FGFR1, we transfected macrophages with 4 FGFR1 small interfering RNAs (siRNAs) or a negative control (Supplementary Figure 11B and Supplementary Table 4) and found that FGF8-mediated increases in bacterial phagocytosis and macrophage killing were suppressed by FGFR1-siRNA-3 (Supplementary Figure 11C). Therefore, we used the TransIT-QR prenatal Delivery Solution of FGFR1-siRNA-3 for the knockdown of FGFR1 in vivo. As expected, there was no difference in mortality rate (Supplementary Figure 11D) or bacterial load (Supplementary Figure 11E) in rFGF8-treated mice posttransfection with siFGFR1 compared to control mice. FGFR1 knockdown abolished the protective effect of FGF8. We therefore utilized a highly selective FGFR1 inhibitor, PD173074, and selected its optimal concentration to determine whether blocking FGFR1 phosphorylation could produce a similar effect (Supplementary Figure 12A and 12B) [27, 28]. PD173074 treatment significantly inhibited FGF8-mediated bacterial phagocytosis and killing (Figure 5C). Furthermore, PD173074 treatment abolished FGF8-induced protection against death in mice as demonstrated by increased mortality (Figure 5D), bacterial burden (Supplementary Figure 11F), serum levels of ALT, AST, LDH, and creatinine (Supplementary Figure 11G), as well as tissue pathology scores (Supplementary Figure 11H).

FGF8 Enhanced Antibacterial Functions of Macrophages Through ERK1/2 Signaling Pathways

These results indicated that FGF8 regulates macrophage functions during sepsis via FGFR1 binding so we further examined downstream signaling pathways triggered by FGF8/FGFR1. The addition of rFGF8 induced the phosphorylation of extracellular regulated protein kinases 1/2 (ERK1/2) in a time-dependent manner but not of P38 mitogen-activated protein kinase (p38 MAPK), signal transducer and activator of transcription 1 (STAT1), or protein kinase B (PKB, Akt) (Figure 6A). Bacterial stimulation combined with rFGF8 treatment increased the phosphorylation of ERK1/2 but not the phosphorylation of STAT, PI3 K, or p38 MAPK (Figure 6B). These results were further dissected with the use of the selective, noncompetitive MAPK kinase inhibitor U0126 that can effectively inhibit Erk1/2 phosphorylation [29–31]. Based on previous studies and toxicity threshold values from the 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT) assays (Supplementary Figure 12C), we used an optimal concentration of U0126 that could suppress ERK1/2 activation by FGF8 (Supplementary Figure 12D). FGF8-mediated increases of bacterial phagocytosis and killing (Figure 6C) were suppressed in macrophages by U0126. Furthermore, U0126 treatment impaired FGF8-induced protection against death of septic mice (Figure 6D).

Figure 6.

FGF8-enhanced antibacterial functions of macrophages are regulated by ERK1/2 signaling pathways. A, Peritoneal murine macrophages were treated with or without rFGF8 (200 ng/mL) for the indicated times and examined for the presence of phosphorylated ERK1/2, P38, STAT1, and Akt. Three independent experiments were performed with comparable results and representative blots are shown. B, Effects of rFGF8 on the activation of MAPK, STAT, PI3 K signaling pathways in murine macrophages. Peritoneal macrophages were treated with or without rFGF8 (200 ng/mL) and then incubated with heat-inactivated Pseudomonas aeruginosa. Total cellular proteins were extracted from murine macrophages for the detection of phosphorylated ERK1/2, P38, STAT1, and Akt with the indicated antibodies by western blot analysis. Experiments were performed in 3 independent experiments with consistent results and representative blots are shown. Peritoneal macrophages (n = 4 per group) were pretreated with the ERK1/2 inhibitor U0126 (20 μM) for 1 hour followed by incubation with or without rFGF8 (200 ng/mL). C, Analysis of in vitro bacterial phagocytosis and killing of P. aeruginosa. D, Mortality after blocking ERK1/2 signaling pathway with U0126 (10 mg/kg) and subsequent treatment with rFGF8 (12.5 μg/kg) or PBS control after CLP (n = 20 per group). Except for survival rate (D), data are presented as means and are representative of 3 independent experiments; (C) Kruskal-Wallis test followed by Dunn multiple comparisons posttest; (D) Kaplan-Meier analysis followed by log-rank test. *P < .05, **P < .01, ****P < .0001 compared within 2 groups. Abbreviations: CFU, colony-forming unit; CLP, cecal ligation and puncture; DMSO, dimethyl sulfoxide; ns, not significant; PBS, phosphate-buffered saline; rFGF, recombinant fibroblast growth factor.

FGF8 Is a Candidate Biomarker for Sepsis

To demonstrate the clinical relevance of these basic findings above, 73 adult patients and 96 pediatric patients with sepsis and corresponding healthy control subjects were enrolled in our study. Both in the adult and pediatric cohort, patients with sepsis had significantly higher serum levels of FGF8 on the day of ICU admission versus healthy volunteers (Figure 7A). In receiver operator characteristic (ROC) curve analysis for diagnoses of sepsis in adults, the area under curve (AUC) value for FGF8 was 0.89 (95% confidence interval [CI], .83–.96) (Figure 7B). A cutoff value of 34.20 ng/mL provided high sensitivity (80.82%) while still maintaining reasonable specificity (93.75%) (Supplementary Table 5). Consistent with the results in the adult cohort, the AUC of FGF8 for the diagnosis of sepsis in children was 0.81 (95% CI, .74–.89) (Figure 7B). The cutoff value was 21.05 ng/mL with a sensitivity of 93.98% and a specificity of 60.71% (Supplementary Table 5). FGF8 abundance was independent of sex in patients with sepsis as well as in the healthy controls (Figure 7C). In addition, FGF8 levels were significantly lower in adults than in children (Figure 7D). However, there was no difference in serum FGF8 concentrations between survivors and nonsurvivors (both in pediatric and adult patients with sepsis) (Supplementary Figure 13A and 13B). ROC curve and Cox survival regression analysis showed that SOFA/pSOFA score had the best diagnostic efficacy in predicting death in patients with sepsis, while FGF8 showed no substantial advantage in predicting prognosis in patients with sepsis (Supplementary Figure 13C and 13D and Supplementary Tables 6–9).

Figure 7.

FGF8 is a candidate biomarker for sepsis. A, Levels of FGF8 at admission measured by ELISA in serum samples collected from 73 adult patients with sepsis, 96 child patients with sepsis, and corresponding healthy controls. B, Receiver operating characteristic analysis of FGF8 for diagnosis of sepsis (AUC = 0.89 for adult and AUC = 0.81 for children). C, Comparison of serum FGF8 levels between male and female patients with sepsis and healthy controls. D, Comparison of serum FGF8 levels between adults and children in healthy controls and patients with sepsis. A, C, and D, Nonparametric Mann-Whitney U test. ****P < .0001 compared within 2 groups. Abbreviations: AUC, area under the curve; CI, confidence interval; ELISA, enzyme-linked immunosorbent assay; FGF, fibroblast growth factor; ns, not significant.

DISCUSSION

Pathogen clearance and inflammatory response control are 2 key factors in determining the prognosis of sepsis. The present study reveals a critical role for FGF8 in the innate immune response to bacterial sepsis and pathogen clearance. In particular, we report the following 4 important findings: (1) The expression of FGF8 is increased in patients with sepsis and has good diagnostic efficacy; (2) FGF8 restoration therapy protected mice with polymicrobial infection from sepsis in a macrophage-dependent manner; (3) Inhibition of FGFR1 abrogated the protective effects of FGF8 against experimental sepsis; and (4) activation of the ERK1/2 signaling pathway was crucial for FGF8-mediated augmentation of antibacterial functions in macrophages and improvement of survival in experimental sepsis. Hence, FGF8 not only could serve as a biomarker for early diagnosis of sepsis, but also represents a potential host-directed therapy to restore immune homeostasis in combating sepsis.

In a hope for more successful patient outcomes, it is critical to promptly and accurately diagnosis sepsis and effectively determine the immune status of patients for targeted therapy [32, 33]. Our study indicated that FGF8 production was significantly elevated in patients with sepsis and the ROC curve demonstrated its diagnostic efficacy. The strength of this study is that the findings were replicated in cohorts of adults and children with sepsis. This phenomenon has also been observed for other cytokines such as elevation of circulating IL-36 levels associated with decreased risk of death following sepsis [34]. Another is Metrnl (IL-41) and patients with sepsis with low serum Metrnl levels had a 2.3-fold increased risk of death compared to patients with high serum Metrnl levels [21]. However, the clinical value of these 2 cytokines in sepsis has not been demonstrated in pediatric patients.

Our study highlights an immunomodulatory role of FGF8 in sepsis and provides new insights into the effects of the FGF family in sepsis by investigating the role of FGF8 in combatting bacterial infections. Previous studies have indicated that FGF2 [35], FGF5 [36], FGF9 [37], FGF19 [38], and FGF21 [39] can reduce inflammatory responses and ameliorate tissue damage in experimental sepsis, thereby improving mortality. Our study also found that rFGF8 treatment reduced multiple organ damage in septic mice and was involved in the regulation of apoptosis and autophagy in lung and intestinal tissue (data not shown). Autophagy and apoptosis play important roles in the pathogenesis and development of sepsis by regulating the pathogen-inflammation-apoptosis balance. This may be another important way that FGF8 plays a protective role in sepsis.

Early and effective infection control is a critical element of treatment for sepsis [40–42]. In this study, we found that FGF8 protein expression in macrophages increased upon the introduction of in vitro infectious stimuli, which suggests that FGF8 may have functions in the anti-infectious immune response. FGF8 could directly enhance the phagocytosis and killing effect of macrophages on bacteria and our macrophage depletion and adoptive transfer experiments demonstrated a contribution of FGF8-macrophage signaling as a host defense against infection. Conversely, inhibition of FGFR1 expression and its downstream ERK1/2 signaling eliminated the effect of FGF8 on macrophage bacterial phagocytosis and killing. Together, our results provide new insights into the function of FGF8 in targeting macrophages and combating bacterial infection. These findings offer a new therapeutic tool for treating sepsis and other bacterial infectious diseases.

Our study has several limitations. First, these findings will be enriched by comparisons with populations in Western countries and in general clinical practice. Second, as a cell growth factor, FGF8 plays an important regulatory role in injury repair, tissue maintenance, growth and development [43, 44]. It is speculated that mechanisms other than bacterial clearance may also participate in FGF8-mediated protection. Third, the activation of FGFR1 and ERK1/2 by FGF8 is responsible for the enhanced bacterial phagocytosis and killing function of macrophages. The regulation of FGFR1 and ERK1/2 by FGF8 has been widely reported [26, 45, 46] but how FGFR1-ERK1/2 signaling affects macrophage function is unclear. Lastly, we have established that macrophages release large amounts of FGF8 during infection, but the FGF8 source in sepsis remains unclear. Macrophages can express many types of pattern recognition receptor, including TLR, RLG-I-like receptors, NOD-like receptors, and C-type lectin receptors [47]. These pattern recognition receptors are stimulated to induce transcriptional expression of various cytokines in the cell. Interestingly, we found that the expression of FGF8 in sepsis was not mediated by either TLR2 or TLR4. The molecular mechanism of elevated circulating FGF8 in sepsis needs to be further explored.

CONCLUSIONS

In summary, our study revealed that FGF8 measurement could represent an attractive biomarker for identification of sepsis. In complementary experimental studies, recombinant FGF8 replacement therapy enhanced bacterial clearance and improved survival in sepsis by activating macrophages in a FGFR1 and ERK1/2-dependent manner. Thus, these data lay the foundation for future studies guiding the development of personalized FGF8-based immune therapies for sepsis (Supplementary Figure 14).

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online (http://jid.oxfordjournals.org/). Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.

Notes

Author contributions. Z. S. and K. C. contributed to the conception and design. K. C., Y. R., W. M., Y. H., and Y. L. performed the experiments. K. C., Y. R., and Z. S. analyzed data. K. C. and Z. S. wrote the manuscript. X. Y. and H. T. collected the blood sample from patients with sepsis. D. C., Y. Y., and X. Z. supervised the work and helped to revise the manuscript. All authors read and approved the final manuscript.

Data availability. The datasets generated and analyzed during the current study are available from the corresponding author on reasonable request.

Financial support. This work was supported by Chongqing Natural Science Foundation Joint Fund for Innovation and Development Municipal Education Commission/Key Project (grant number CSTB2022NSCQ-LZX0045 to Z. S.); the National Natural Science Foundation of China (grant number 82272399 to D. C.); Chongqing Young and Middle-Aged Medical Talents Project - Chongqing Municipal Health Commission (to Z. S.); Chongqing Medical University Program for Youth Innovation in Future Medicine (to D. C.); Chongqing Medical University, Children's Hospital YINGCAI Program (to Z. S.); and Chongqing Medical University, Children's Hospital Distinguished Young Scholars (to Z. S.).