High-mobility group box 1 protein initiates postoperative cognitive decline by engaging bone marrow-derived macrophages

Abstract

Background: Aseptic trauma engages the innate immune response to trigger a neuroinflammatory reaction that results in postoperative cognitive decline. The authors sought to determine whether high-mobility group box 1 protein (HMGB1), an ubiquitous nucleosomal protein, initiates this process through activation and trafficking of circulating bone marrow-derived macrophages to the brain.

Methods: The effects of HMGB1 on memory (using trace fear conditioning) were tested in adult C57BL/6J male mice; separate cohorts were tested after bone marrow-derived macrophages were depleted by clodrolip. The effect of anti-HMGB1 neutralizing antibody on the inflammatory and behavioral responses to tibial surgery were investigated.

Results: A single injection of HMGB1 caused memory decline, as evidenced by a decrease in freezing time (52 ± 11% vs. 39 ± 5%; n = 16-17); memory decline was prevented when bone marrow-derived macrophages were depleted (39 ± 5% vs. 50 ± 9%; n = 17). Disabling HMGB1 with a blocking monoclonal antibody, before surgery, reduced postoperative memory decline (52 ± 11% vs. 29 ± 5%; n = 15-16); also, hippocampal expression of monocyte chemotactic protein-1 was prevented by the neutralizing antibody (n = 6). Neither the systemic nor the hippocampal inflammatory responses to surgery occurred in mice pretreated with anti-HMGB1 neutralizing antibody (n = 6).

Conclusion: Postoperative neuroinflammation and cognitive decline can be prevented by abrogating the effects of HMGB1. Following the earlier characterization of the resolution of surgery-induced memory decline, the mechanisms of its initiation are now described. Together, these data may be used to preoperatively test the risk to surgical patients for the development of exaggerated and prolonged postoperative memory decline that is reflected in delirium and postoperative cognitive dysfunction, respectively.

Figures

Figure 1

Study design. (A) First experiment; mice were divided in 5 groups treated with IP injection of control liposome versus clodrolip 1h before HMGB1 Ag versus saline injection. Control animals received saline injections. The training session of the memory test was performed 30 min after the clodrolip/control liposome injection and 30 min before HMGB1 Ag/saline injection; and the context session was performed 72 h later. (B) Second experiment; mice were divided in 4 groups treated with anti-HMGB1 versus saline 1 h before tibia fracture. The training session of the memory test was performed 30 min after the IP injection and 30 min before tibia fracture. (C) Third experiment; Mice were divided in 4 groups treated with anti-HMGB1 versus saline 1h before tibia fracture and sacrificed 1 h and 24 h after the tibia fracture. Anti-HMGB1 = neutralizing HMGB1 monoclonal antibody; CT-lip = control liposome; HMGB1 = high-mobility group box 1 protein; HMGB1 Ag = HMGB1 antigen; IP = intraperitoneal.

Figure 2

Effects of HMGB1 on Cognitive Decline. Contextual fear response reveals hippocampal-dependent memory impairment at postoperative day 3 (Arm A). Quantification of the freezing time percentage according to the five groups (n = 15-17; * p = 0.012 control versus HMGB1 Ag, and # p = 0.039 HMGB1 Ag versus clodrolip+HMGB1 Ag, with one-way ANOVA and Bonferroni post hoc analysis). HMGB1 = high-mobility group box 1 protein; HMGB1 Ag = HMGB1 antigen.

Figure 3

Effects of preoperative administration of HMGB1 neutralizing monoclonal antibody on postsurgical memory impairment. Quantification of the freezing time percentage on contextual testing according to the four groups at postoperative day 3 (Arm B). (n = 15-16; **** p < 0.001 control versus surgery, and ** p = 0.001 surgery versus anti-HMGB1+surgery and with one-way ANOVA and Bonferroni post hoc analysis). Anti-HMGB1 = neutralizing HMGB1 monoclonal antibody; HMGB1 = high-mobility group box 1 protein.

Figure 4

Effects of preoperative administration of HMGB1 neutralizing monoclonal antibody on the HMGB1 and IL-6 serum concentration at 1 and 24 h after tibia fracture (Arm C). The dotted lines represent the average values for respectively for HMGB1 and IL-6 concentration of 6 control mice without any treatment or surgery (n = 6). (A) Levels of HMGB1 serum concentration 1 h and 24 h after surgery. After log transformation of the raw data, we observed with two way ANOVA significant time and treatment effects (p = 0.002 for the time effect; p = 0.005 for the treatment effect and p = 0.010 for interaction between time and treatment) and a significant difference between the 2 groups at 1 h (22.00 ± 16.54 vs. 5.29 ± 4.13 surgical control vs. anti-HMGB1+surgery p = 0.040 with two way ANOVA with Bonferroni's post-hoc analysis). (B) Levels of IL-6 serum concentration 1 h and 24 h after surgery. After log transformation of the raw data, we observed with two way ANOVA significant time and treatment effects (p = 0.015 for the time effect; p < 0.001 for the treatment effect and p = 0.015 for interaction between time and treatment) and a significant difference between the 2 groups at 1 h (24.24 ± 12.87 vs. 4.90± 4.47 surgical control vs. anti-HMGB1+surgery p = 0.006 with two way ANOVA with Bonferroni's post-hoc analysis). Anti-HMGB1 = neutralizing HMGB1 monoclonal antibody; HMGB1 = high-mobility group box 1 protein; IL-6 = interleukin 6.

Figure 5

Effects of HMGB1 neutralizing monoclonal antibody on hippocampal transcription of IL-6, TNF-α, IL-1β and MCP-1 24 h after tibia surgery (Arm C). (A) IL-6 mRNA (B) TNF-α mRNA (C) IL-1β and (D) MCP-1 (n = 6; * p = 0.023; ** p = 0.009; *** p < 0.001, # p = 0.003, ## p = 0.005 with one-way ANOVA and Bonferroni post hoc analysis). Anti-HMGB1 = neutralizing HMGB1 monoclonal antibody; HMGB1 = high-mobility group box 1 protein; IL-1=interleukin 1; IL-6 = interleukin 6; TNF= tumor necrosis factor; MCP-1 = monocyte chemotactic protein-1.