Frontline Science: HMGB1 induces neutrophil dysfunction in experimental sepsis and in patients who survive septic shock

Abstract

Sepsis is accompanied by the initial activation of proinflammatory pathways and long-lasting immunosuppression that appears to contribute to late-occurring mortality. Although high-mobility group box 1 (HMGB1) is involved in many aspects of inflammation, its role in sepsis-induced immune suppression remains unclear. In this study, we examined HMGB1's contribution to neutrophil NADPH oxidase activity dysfunction and associated neutrophil-dependent bacterial clearance in mice subjected to sepsis and in patients who survive septic shock. Using a murine model of polymicrobial septic peritonitis, we demonstrated that treatment with anti-HMGB1 Ab significantly diminished sepsis-induced dysfunction of neutrophil NADPH oxidase activity. In a subsequent set of experiments, we found that blocking HMGB1 preserved the ability of neutrophils from patients recovering from septic shock to activate NADPH oxidase. Taken together, our data suggest that HMGB1 accumulation in the late phase of sepsis plays a specific role in the development of postsepsis immunosuppression and specifically affects neutrophil-dependent antibacterial defense mechanisms. Thus, blocking HMGB1 may be a promising therapeutic intervention to diminish the adverse effects of sepsis-induced immunosuppression.

Keywords: HMGB1; NADPH oxidase; immunosuppression; neutrophils; sepsis.

Figure 1.. Mice subjected to intra-abdominal polymicrobial sepsis developed neutrophil dysfunction.

(A) Mice subjected to CLP or sham surgery were treated with antibiotics for 5 d, followed by measurement of neutrophil function at d 7. (B) Extent of NADPH oxidase activation in PMA-stimulated bone marrow neutrophils obtained from control (sham) or CLP mice. (C) Bacterial clearance was determined after culture of neutrophils (CLP or control) with E. coli for 16 h (n = 6). Means ± sd. *P < 0.05; **P < 0.01.

Figure 2.. HMGB1 diminished neutrophil oxidative burst.

(A) Western blot and quantitative analysis showed the amount of HMGB1 in mouse sera obtained before or 24 h or 7 d after CLP or sham (n = 5). Means ± sd; **P < 0.01. (B) Representative rates of cytochrome c reduction showed NADPH oxidase inactivation in neutrophils treated with HMGB1 (0 or 300 ng/ml) for 15 min. The oxidative burst was determined after inclusion of E. coli (10⁶/ml) for 15 min. (C) Sham mice and mice with CLP-induced sepsis were treated with HMGB1-neutralizing Ab (125 µg) or control isotype-specific IgG for (5 d). NADPH oxidase subunit p-phox40 (left) or oxidative burst (right) was than determined in bone marrow neutrophils stimulated with PMA (0 or 10 nM) for 15 min (n = 3). Means ± sd; **P < 0.01.

Figure 3.. Human PMNs have diminished NADPH oxidase activation and ability to kill bacteria in patients who survived septic shock.

(A) PMNs purified from blood of normal patients or of patients in ICU without sepsis or 7 d after septic shock. (B) The amount of HMGB1 was determined in plasma of healthy donors, patients without sepsis, or patients after septic shock (n = 6–8). Means ± sd; *P < 0.05. (C) The extent of RAGE and TLR4 mRNA expression was determined in PMNs of control and postsepsis patients. RQ-PCR was performed on neutrophils at admission (T0) and when vasopressor therapy was stopped (T1) to analyze RAGE and TLR4 expression. Each sample was normalized to 18S expression and compared to expression levels on T0 neutrophils. (D) PMNs purified from blood of normal patients or patients in ICU without sepsis or 7 d after septic shock were loaded with H₂DCF-DA and treated with PMA (0 or 10 μM) for 30 min. Measurement of ROS production occurred before and after stimulation with PMA (10 µM) for 15 min. Representative FACS analyses (left) showed the extent of ROS production in unaltered (white) or stimulated (gray) PMNs. (E) Measurement of ROS production after stimulation with PMA (10 nM) for 15 min. PMA-stimulated NADPH oxidase activation was determined in PMNs isolated from patients with septic shock, patients without sepsis, and healthy donors (right) (n = 6–7). Means ± sd; *P < 0.05. (F) PMNs (10⁶ cells) obtained from healthy or postsepsis patients were incubated with E. coli or MSSA for 16 h and the amount of viable bacteria determined with a CFU assay (n = 6–7). Means ± sd; *P < 0.05).

Figure 4.. HMGB1 in postsepsis plasma promoted dysfunction of PMN oxidative burst.

(A) Healthy donor PMNs were incubated for 30 min with plasma of healthy patients or plasma of those with or without sepsis, and then oxidative burst was determined after exposure to PMA (n = 6–8) Means ± sd; *P < 0.05. (B) Plasma obtained from postsepsis patients was treated with HMGB1-neutralizing Ab (50 µg) or IgG (isotype control) for 2 h, and the treated plasma was used in culture with healthy donor PMNs for an additional 30 min. The oxidative burst was measured after inclusion of PMA (0 or 10 µM) for 15 min (n = 6). Means ± sd; *P < 0.05.