Anti-HMGB1 Antibody Decreases Surgery-Induced Upregulation of Necroptosis-Associated Proteins in Aged Rats

the molecular mechanism on how systemic HMGB1 neutralization improves POCD is not fully elucidated. Necroptosis could cause sterile inflammation and negatively associates with the cognitive score in Alzheimer ’ s disease. Thus we detected the effects of anti-HMGB1 antibody on the necroptosis-associated protein expressions dur-ing the POCD of aged rats. under sevoflurane anesthesia and analgesia were performed. Immunoglobulin G (1 mg/kg) and HMGB1 neutralizing antibody (1 mg/kg) were injected via tail right before and 6 hours after surgery. The expression of necroptosis-associated proteins (HSP90, CDC37, and RIP3) in the prefrontal cortex of brain was detected by western blot and immunofluorescence. Oxidative stress was measured by dihydroethidium staining. Results: Systemic administration of anti-HMGB1 antibody decreased the levels of reactive oxygen species (ROS) and reduced the expression of HSP90, CDC37, and RIP3 in prefrontal cortex neurons of brains after surgery (P < 0.05, respectively). Moreover, acetylated HSP90 (ACHSP90) which is a negative regulator of necroptosis was significantly decreased by treatments of anti-HMGB1 neutralization antibody (P < 0.05). Conclusion: Systemic administration of anti-HMGB1 antibody may improve POCD through reducing reactive oxygen species and decreasing necroptosis in the prefrontal cortex of the aged brain. (Funded by the National Natural Science Foundation of China and Central South University Postdoctoral Foundation in Changsha, China.) ABSTRACT

P ostoperative cognitive dysfunction (POCD) usually happens in elderly patients following surgery, especially in patients under critical condition. Most of the elderly patients present impaired memory and damaged thinking skills (1). Up to 74% of elderly patients with critical illness experience POCD in early days following surgery and up to 34% of elderly patients still demonstrate symptoms of POCD one year after surgery (2). POCD is a disorder complicated with multifactors and its occurrence can be affected by multiple risk factors that may present before, during, and after surgery. Among these risk factors, the increased vulnerability of brain is an essential factor intends to induce POCD (3)(4)(5)(6)(7). Neuroinflammation is the major pathological mechanism (8,9). Possession of POCD largely extended the length of hospital stay and cost more expenses for patients. Even after discharged from hospital life independence is significantly decreased. Combined with fast-growing geriatric population, these consequences of POCD put a heavy burden on patients, their families, and the society (10). Therefore, exploring mechanisms and developing treatments for POCD is one of the priorities in geriatric research (11).
High mobility group box 1 protein (HMGB1) which is a crucial member of damage-associated molecular pattern molecules (DAMPs), is quickly released after tissue trauma and plays an essential role in immune cell recruitment and activation, cytokine release, and cell death (12). HMGB1 had been identified as an important mediator for sterile inflammation (13). In previous research, we find out HMGB1 significantly elevated in the peripheral blood of elderly patients after gastrointestinal surgery and the elevated levels of HMGB1 were positively associated with increased level of human POCD (14). Furthermore, systemic neutralization of HMGB1 in peripheral blood of aged rats after surgery remarkably decreased the level of postoperative neurocognitive dysfunction (15). Meanwhile, Maze and colleagues find out singledose injection of HMGB1 caused memory decline and blocking HMGB1 with monoclonal antibody before surgery reduced postoperative memory decline (16). These data revealed the pivotal role of systemic HMGB1 in mediating development and progression of POCD. Howev-er, the molecular mechanism on how systemic HMGB1 neutralization improves POCD is not fully elucidated.
Necroptosis is a common form of programmed cell death that often seen in ischemic injuries of the nervous system, heart, and kidney. The occurrence of necroptosis causes sterile inflammation (17,18). The HSP90 / CDC37 / RIP3 signaling pathway is the classic pathway inducing necroptosis (19)(20)(21)(22)(23)(24). A recent study has shown that necroptosis was activated in Alzheimer' s disease and necroptosis activation inversely correlated with cognitive scores and increased cell loss (25). The prefrontal cortex is closely involved in cognition and memory. Herein, we assessed the effects of systemic HMGB1 neutralization on necroptosis pathway proteins (HSP90, CDC37, and RIP3) in prefrontal cortex of the brain and its role in postoperative cognitive dysfunction in aged rats after surgery. These effects of systemic HMGB1 neutralization may provide some new evidence for pathophysiological mechanisms of POCD. gery + HMGB1 neutralizing antibody (S + anti-HMGB1) (1 mg / kg, i. v.). For each group, five aged rats were used. Treatments were implemented through tail vein injection right before surgical incision and 6 hours after surgery; dosage and timing were previously described by Okuma et al (26). Mouse IgG2b (Sigma, M1395-5MG) was used as an isotype control. Anti-HMGB1 monoclonal antibody (2G7, mouse IgG2b) was kindly provided by Dr. H. E. Harris' s laboratory, Stockholm, Sweden. The HMGB1 neutralizing efficacy and efficiency of Anti-HMGB1 monoclonal antibody have been extensively characterized in a series of in vitro and in vivo studies (26)(27)(28)(29)(30)(31)(32). The 2G7 anti-HMGB1 monoclonal antibody neutralizes both HMGB1stimulated cytokine/chemokine releases and chemotactic activities.

Partial Hepatolobectomy
Rats were first rapidly anesthetized with 4.5% sevoflurane (Maruishi Pharmaceutical Co., Ltd., Japan) mixed with high flow of pure oxygen (6 L / min). When each individual rat lost sensation of right reflex, a 14G catheter was inserted into its trachea. 2%-2.5% sevoflurane were supplied with oxygen (80-85%) to maintain anesthesia and analgesia. The constant gas supplied was under surveillance by a multi-functional monitor (Datex-Ohmeda, Helsinki, Finland). Rat' s vital sign parameters such as respiratory rate (R), Pet-CO2, FiO2, and FiSev were continuously recorded. The depth of anesthesia was modulated according to the R (30-50 cycles / min) and the body movement of the rats. The partial hepatolobectomy was carried out with strict aseptic procedures and performed as previously described with some modifications (33,34). In brief, an incision about 2 cm long was made below the xyphoid; the left lobe of the liver was carefully isolated, ligated, and then removed. Finally, muscles and skin were closed with sterile sutures and 0.2 ml of 0.25% bupivacaine was subcutaneously administered to provide local postoperative analgesia. Animals were then allowed to recover for further investigation.

Immunofluorescence
For the immunofluorescence experiments, the localization and expression level of necroptosis associated proteins (HSP90, CDC37, and RIP3) were assessed on day 3 after surgery. Rats were euthanized with chloral hydrate (10%) and perfused transcardially with ice cold 0.01 M phosphate-buffered saline (PBS). The brain was dissected immediately following termination. One hemisphere was used for immunofluorescence and the other for western blot. The hemisphere used for immunofluorescence was immersed in 4% paraformaldehyde for fixation at 4 ºC for 24 hours. The brains were dehydrated with 15% sucrose for one day and 30% sucrose for two days respectively and later embedded in OCT (Sakura, Tissue-Tek, USA). Cross-sections were consecutively cut at 20 μm of thickness in Leica cryostat (CM1860, Leica Biosystems Inc., IL, USA). Totally thirty sections were prepared. Three sections of prefrontal cortex were randomly picked from 3 sets of serial sections from each rat at +1.70 mm to +1.10 mm anteroposterior to the bregma for immunostaining. Sections were washed three times with 0.01 M PBS (10 min/each time). After three washes, sections were blocked with 5% bovine serum albumin (BSA) in 0.01 M PBS plus 0.3% Triton X-100 for 50 min at room temperature and then incubated with primary antibodies HSP90 (1: 100 Table 1. Following three washes in 0.01 M PBS, the sections were incubated with secondary antibodies (Biotinylated Goat Anti-Rabbit IgG, 1:200, Vector, USA) for 2 hours at room temperature. Sections were finally washed three times with 0.01 M PBS and then covered with permount containing DAPI (Vector, USA, H-1200). For each stained section, pictures containing prefrontal cortex area of the brain were taken under the same magnification (40 X objective lens) by a fluorescent microscope (DS-Ril, Nikon, Japan). The mean fluorescence density of target proteins (AC-HSP90, CDC37, and RIP3) was measured and analyzed by using Image Pro-Plus Anti-HMGB1 Antibody Decreases Necroptosis-Associated Proteins Hongkang Zhou et al. 6.0 software (Media Cybernetics Inc., Rockville, MD, USA).

Western Blot
Western blot analysis was employed to further detect the expression of HSP90, CDC37, and RIP3 (necroptosis-associated proteins) in the prefrontal cortex of brain. In brief, frozen prefrontal cortex was homogenized in lysis buffer with protease inhibitors (cat. no. P8340; Roche, Germany) and nylmethanesulfonylfluoride (cat. no. p7626; PMSF, Sigma, USA). Protein samples were centrifuged at 14,000 g for 10 min at 4 ºC and supernatant was collected. The protein concentration was quantified by using a BCA

Statistical Analysis
The data were expressed as mean ± S. E. M.
(standard error of the mean). One-way ANOVA was used to analyze the data from immunofluorescence, western blot, and ROS measurements. Bonferroni multiple comparison test was performed to compare selected groups when ANO-VA showed significance. Statistical analysis was conducted by using Prism 5 (Graph Pad Software Inc., La Jolla, CA, USA). A P < 0.05 was considered as significant.

Results
Expressions of necroptosis-associated proteins are decreased by anti-HMGB1 treatment HSP90, CDC37, and RIP3 are classic necroptosis associated proteins and play critical roles in necroptosis. Prefrontal cortex is closely involved in cognition and memory. Thus we tried to detect the expressions of HSP90, AC-HSP90, CDC37, and RIP3 in prefrontal cortex by immunofluorescence labeling and western blot analysis. The immunofluorescent density of AC-HSP90, CDC37, and RIP3 were measured on stained sections of prefrontal cortex, which presenting in situ expression level of these proteins.  Results are presented as mean ± S.E.M. (n=5). * to normal control group (Con), *P < 0.05; # to S + IgG group, #P < 0.05; by one-way ANOVA. Scale bar = 50 μm. Compared to the control groups, the level of AC-HSP90 in S + IgG was distinctly decreased on day 3 after surgery (P < 0.05) ( Figure 1A, B). Rats treated with anti-HMGB1 antibodies (S + anti-HMGB1 group) had significantly higher AC-HSP90 expression than the S + IgG group (P < 0.05) ( Figure 1A, B). In addition, compared to corresponding control groups, the expression level of CDC37 (Figure 2A, B), and RIP3 ( Figure 3A, B) were all significantly higher in S + IgG groups at day 3 after surgery (P < 0.05, respectively). Treatment with anti-HMGB1 antibodies all apparently decreased CDC37 (Figure 2A, B), and RIP3 ( Figure 3A, B) expression at day 3 after surgery (P < 0.05, respectively). We also had assessed HSP90, AC-HSP90, CDC37, and RIP3 expression in lysate of pre-frontal cortex of aged rat brains after surgery. Compared to their corresponding controls, Expression of HSP90 ( Figure 4A), CDC37 ( Figure  4B), and RIP3 ( Figure 4C) were all increased in S + IgG groups at day 3 after surgery (P < 0.05, respectively). Furthermore, Systemic administration of anti-HMGB1 neutralizing antibodies significantly reduced the upregulation of HSP90 ( Figure 4A), CDC37 ( Figure 4B), and RIP3 (Figure 4C) in S + anti-HMGB1 groups on day 3 after surgery as compared to the corresponding S + IgG groups (P < 0.05, respectively). Acetylation of HSP90 has implicated an inhibitory role in necroptosis (19). Compared to control, the acetylated HSP90 (AC-HSP90) was obviously less in S + IgG group on day 3 after surgery (P < 0.05) ( Figure 4A), indicating an increased level of necroptosis. However, the anti-HMGB1

Figure 2. Anti-HMGB1 Treatment Decreases Expression of CDC37 in Aged Rats After Surgery.
(A) The intensity of CDC37 immunofluorescence (red) was used to show expression level of CDC37 in the prefrontal cortex of brain on postoperative day 3. Neurons (Neun) were stained with green fluorescence. Merged images of CDC37 and Neun double labeling present blue fluorescence. (B) The quantified mean density of CDC37 fluorescence in the prefrontal cortex (FC) shows increased expression of CDC37 in surgery + immunoglobulin G as control antibody group (S + IgG) and significantly decreased CDC37 expression level in anti-HMGB1 antibodies treated group (S + anti-HMGB1). Results are presented as mean ± S.E.M. (n=5). * to normal control group (Con), *P < 0.05; # to S + IgG group, #P < 0.05; by one-way ANOVA. Scale bar = 50 μm.

Volume 5
Number 3 May, 2018 antibodies significantly increased the AC-HSP90 level in S + anti-HMGB1 group (P < 0.05) (Figure 4A). Combined with immunofluorescence data of necroptosis-associated proteins, these results demonstrated a protective mechanism of anti-HMGB1 neutralizing antibodies in necroptosis.

Anti-HMGB1 treatment decreases oxidative stress in the prefrontal cortex of brain after surgery
In order to further investigate the mechanisms of systemic HMGB1 neutralization improving the postoperative neurocognitive decline, the reactive oxygen species (ROS) levels of prefrontal the cortex of aged rat brains after surgery were measured. Compared to the control group, the ROS level of S + IgG group was significantly elevated on day 3 after surgery, showing an higherthan-normal oxidative stress after surgery (P < 0.05) ( Figure 5A, B). Notably, Compared to the S + IgG group, the systemic HMGB1 neutralization significantly decreased the ROS level in S + anti-HMGB1 group (P < 0.05) ( Figure 5A, B). Moreover, the neurons of prefrontal cortex were labeled by immunofluorescence (anti-Neuron primary antibodies). The merged images of DHE and neurons double immunofluorescence labeling indicated that the changes of ROS levels mainly happened in the neurons of prefrontal cortex ( Figure 5A).

Discussion
In this study, we investigated the impact of anti-HMGB1 neutralizing antibodies on necroptosis associated proteins of brain prefrontal cortex in aged rats after liver surgery. Our data demonstrate that systemic administration of anti- HMGB1 antibodies significantly limited surgeryinduced up-regulation of HSP90, CDC37, and RIP3 expressions in prefrontal cortex of brain. Meanwhile, oxidative stress in prefrontal cortex of brain also distinctively reduced by anti-HMGB1 treatment in aged rats after surgery.
These results indicate that systemic HMGB1 neutralization may provide a protective mechanism of postoperative cognitive dysfunction through reducing ROS level and necroptosis. The high mobility group box (HMGB) protein family is the most represented protein family among the high mobility group proteins (36). There are four members in the HMGB protein family (HMGB1-4). HMGB1 is the most important member of HMGB protein family. As a discovered nuclear protein, HMGB1 has played a critical role in transcription, replication, DNA repair, recombination, nucleosome assembly, and genomic stability (37,38). Besides its nuclear function, HMGB1 involves in an essential extracellular activity; that is, HMGB1 belongs to the damage-associated molecular pattern molecules (DAMPs), which are released from damaged cells, causing immunoresponses locally and systemically (39). Surgical trauma-induced tissue damage leads to a rapid release of HMGB1 and increased systemic levels of HMGB1 have been detected in elderly patients after gastrointestinal surgery. Interestingly, these levels correlated with the development of POCD (14,16), suggesting that the release of HMGB1 may be critical in initiating inflammatory cascade leading to brain dysfunction. Furthermore, HMGB1 release after surgery has been characterized in some preclinical models of cognitive decline (27,40,41).
Increased levels of HMGB1 after surgery have been evidenced to compromising the bloodbrain barrier (BBB) and subsequently lead to neuroinflammation (42,43). The mechanism whereby peripheral HMGB1-induced inflammation contributes to neuroinflammation and cognitive dysfunction had been partially revealed in our previous study. Increased systemic HMGB1 proteins were able to access the brain and causing microglial activation. In the meantime, phosphorylated RREB which is critical in memory and synaptic plasticity of hippocampus was significantly decreased 3 days after surgery. Treat- Results are presented as mean ± S.E.M. (n=5). * to normal control group (Con), *P < 0.05; # to S + IgG group, # P < 0.05; by one-way ANOVA. Scale bar = 50 μm.
The important discovery of present study is that the acetylated HSP90 (AC-HSP90) was reduced by anti-HMGB1 antibody treatment (Figure 1 and 4A).
Postoperative cognitive dysfunction (POCD) is a multifactorial disorder prominently found in elderly patients after surgery. Previous research showed the ccurrences of POCD were related to elevated levels of oxidative stress and neuroinflammation, phosphorylation of Tau proteins, and loss of NMDA receptors (47). In previous study, our data suggest systemic HMGB1 neutralization improves surgery-induced neurocognitive dysfunction (15) and this process may function through regulating the expression of necrop-tosis-associated proteins and reducing levels of oxidative stress. These findings provide new evidence for neuroprotective mechanisms of anti-HMGB1 antibody treatment which may be beneficial for discovering defined mechanisms and treatments of POCD.