Chronic unpredictable stress exacerbates surgery-induced sickness behavior and neuroinflammatory responses via glucocorticoids secretion in adult rats

Accumulated evidence indicates that stress sensitizes neuroinflammatory responses to a subsequent peripheral immune challenge. The present study investigated whether chronic unpredictable stress (CUS) aggravated surgery-induced sickness behavior and neuroinflammatory processes via glucocorticoids secretion in the adult brain. Methods Sprague-Dawley adult male rats (12–14 weeks old) were exposed to 14-day CUS and then subjected to partial hepatectomy 24 h after the last stress session. The rats were pretreated with an antagonist of the glucocorticoids (GCs) receptor RU486 (30 mg/kg, i.p.) 1 h prior to stress exposure. The behavioral changes were evaluated with open field test and elevated plus-maze test. The hippocampal cytokines interleukin (IL)-1β and IL-6 were measured on postoperative days 1, 3 and 7. Ionized calcium binding adaptor protein (Iba)-1, microglial M2 phenotype marker Arg1, brain derived neurotrophic factor (BDNF) and CD200 were also examined at each time point. Results CUS exacerbated surgery-induced sickness behavior. Exposure to CUS alone failed to alter the levels of pro-inflammatory cytokines in the brain. However, CUS exaggerated surgery-induced pro-inflammatory cytokines expression (e.g. IL-1β and IL-6) and upregulated the levels of Iba-1 on postoperative days 1 and 3. An additional significant decreased BDNF, CD200 and a lower level of Arg1 were also observed in the stressed rats following surgical procedure. Pretreatment with RU486 blunted the potentiating effects of CUS on surgery-induced sickness behavior and neuroinflammatory responses. Conclusion Chronic unpredictable stress enhanced surgery-induced sickness behavior and neuroinflammatory responses. Stress-induced GCs played a pivotal role in enhancing surgery-induced neuroinflammatory processes by modulation of microglia functions.


Introduction
Postoperative cognitive dysfunction (POCD), characterized by the progressive deterioration of intellectual/cognitive function, is a major complication following various surgical procedures [1]. Mounting evidence indicates that POCD is associated with surgery-induced neuroinflammatory processes, which may influence neuronal functioning either directly or through modulation of intraneuronal pathways, such as brain derived neurotrophic factor (BDNF) mediated pathway [2,3]. High levels of neuroinflammatory cytokines, such as interleukin (IL)-1β, IL-6, and tumor necrosis factor (TNF)-α play a pivotal role in surgery-induced cognitive deficits [1,4].
Surgical trauma caused sickness behavior and triggered neuroinflammatory responses in the brain of rats [1,4,10]. Psychological stress is common prior to the major surgery. It was reported to affect 60-80% of surgical patients [11]. The main objective of this study was to investigate whether CUS aggravated surgery-induced sickness behavior and neuroinflammatory responses in the adult rats. We also explored whether stress and the consequent increase of circulating GCs modulated the immunophenotype of microglia, thereby sensitizing neuroinflammatory responses to the subsequent surgical challenge.

Methods Animals
Sprague-Dawley adult male rats (12-14 weeks old) were randomly divided into a total of six groups: control group (n = 30), CUS group (n = 36), RU486 group (n = 30), surgery group (n = 30), CUS+surgery group (n = 30), and RU486+CUS+surgery group (n = 30). All animals were housed in groups of four per cage except the day of surgery and had free access to food and water. Colony conditions were maintained at 25˚C on a 12-h light/dark cycle (lights on at 07:00 A.M.). All rats were adapted to their environment for a minimum of 7 days before the experiments. The control rats stayed in their home cage. Partial hepatectomy was performed under general anesthesia (3% isoflurane in O 2 at 0.6 L/min) in the surgery group. Briefly, the liver was exposed through a 1-2 cm midline abdominal incision. The left lateral lobes of the liver (approximately corresponding to 30% of the organ) were excised. The wound was then infiltrated with 0.25% bupivacaine, and closed by sterile suture. To limit variability, all surgeries were performed by the same person. Animals in RU486 and RU486+CUS+surgery groups were intraperitoneally injected with a daily dose of RU486 (30 mg/kg, dissolved in DMSO) 1 h prior to stress exposure. All procedures were conducted in accordance with the Declaration of the National Institutes of Health Guide for Care and Use of Laboratory Animals and approved by China Medical University Animal Care and Use Committee (No: IACUC-2017001).
body weight was measured every two days during the 14-day CUS. The behavioral changes were evaluated with relatively low-stress methods-open field test and elevated plus-maze test on postoperative days 1, 3 and 7. The rats were euthanized and the hippocampi were harvested for biochemical analyses after the behavioral test. The remaining rats were transcardially perfused with ice-cold saline (0.9%) for 3 min, followed by 4% paraformaldehyde for immunohistochemistry at each time point. The cardiac blood was collected on postoperative days 1, 3 and 7 in the morning (09.00 h).

Chronic unpredictable stress
Chronic variate stress was adapted from other models of variate stress with modifications [12]. Individual stressors and length of time applied each day are listed in Table 1. In all stress experiments, the stressed rats were exposed to the same order of stressful stimuli. Stress was applied at a different time of each day to minimize predictability. Restraint was carried out by placing the animal in a 21×6 cm plastic tube and adjusting it with a plaster tape on the outside, so that the animal could not move. Forced swimming was carried out by placing the animal in a glass tank with 25 cm of water depth at 23 ± 2˚C. The control rats were manipulated every day for 10 min in the home cage to control for nonspecific handing effects.

Open-field test
The open-field observation cage consisted of a square wooden arena (100 × 100 × 50 cm) with its inside walls covered in black. It was divided into 25 equal squares, including peripheral area (16 around squares) and central area (9 middle squares). The rats were placed individually into the center of the area and allowed to explore freely. The test was conducted in a quiet room in the morning (8:00-12:00 A.M.). The animals were monitored for 5 minutes with a video tracking software (Smart, San Diego Instruments, San Diego, CA, USA). Spontaneous locomotor activity was assessed by the total amount of distance traveled in the chamber. The time in the central area was taken as measures of anxiety and exploratory behavior.

Elevated plus-maze test
The elevated plus-maze is an apparatus widely used to evaluate the level of anxiety in rodents. The level of anxiety is measured by variables such as the percentage of time spent and the

Plasma GCs concentration
Concentrations of GCs in plasma were quantified by using an enzyme-linked immunoassay (ELISA). In brief, cardiac blood was collected by cardiac puncture into EDTA coated syringes and centrifuged (3000 rpm for 10 min at 4˚C). Plasma was collected and stored frozen (-80˚C) until assaying. GCs titers were assessed by using a competitive enzyme immunoassay kit, following the manufacturer's instructions (R&D Systems, Minneapolis, MN).

Western blot analysis
Hippocampus tissues were collected and homogenized in ice-cold lysis buffer for 30 minutes. The homogenate was centrifuged (12000 g for 10 min 4˚C) and the quantity of protein in the supernatants was determined using a BCA protein assay kit (Fdbio science, China). Equal amounts of protein were loaded per well on a 12% SDS-polyacrylamide gel electrophoresis and transferred to PVDF membranes, which were then blocked with 5% skim milk at room temperature for 2 h. Membranes were then incubated using the following antibodies: rabbit monoclonal anti-IL-1β (1:2000; Abcam, Cambridge, UK), mouse polyclonal anti-IL-6 (1:2500; Abcam, Cambridge, UK), rabbit polyclonal anti-BDNF (1:2000; Abcam, Cambridge, UK) overnight at 4˚C. The membranes were then incubated in appropriate secondary antibodies diluted in TBST for 2 h at room temperature. Chemiluminescence detection was performed using an ECL Western Blotting kit (cat. # 170-5060; Bio-Rad Laboratones. Inc.). Relative expression levels of protein were normalized by the ratio of target protein (IL-1β, IL-6 and BDNF) to β-actin.

Real-time PCR
Total RNA from the different treatment conditions was extracted from the whole hippocampus with TRIzol reagent (Takara, Otsu, Japan) according to the manufacturer's instructions. cDNA was synthesized from total RNA (1.0 μg) using the PrimeScript™ RT reagent Kit with gDNA Eraser (Takara, Otsu, Japan). Two microliters of cDNA were used to perform quantitative real-time PCR. The following primers were used to amplify the mRNA: CD200: 5 0 -CTGC ACACAACTGCATCCTT-3 0 (forward) and 5 0 -GGGCTTTGCTGTAAGTGACC-3 0 (reverse); Arg1: 5'-GCAGAGACCCAGAAGAATGG-3' (forward) and 5'-CACGATG TCCTTGGCAGATA-3' (reverse); Glyceraldehyde 3-phosphate dehydrogenase (GAPDH): 5'-GGGGCTCTCTGCTCCTCCCTG-3' (forward) and 5'-AGGCG TCCGATACGGCCAAA-3' (reverse). GAPDH was adopted as an internal control, which were obtained from Sangon Biotech. The reverse transcription reaction was carried out under the following conditions: 95˚C for 10 min (initial denaturation), followed by 40 cycles of 95˚C for 20 sec, 62˚C for 30 sec and 72˚C for 30 sec (amplification). The relative gene expression was determined by calculating the expression ratio of the gene of interest to GAPDH. The relative expression of mRNA was quantified using the 2 -ΔΔCt method.

Immunohistochemistry
The sections (5 μm) were pretreated with 3% H 2 O 2 for 20 min and then incubated with the specific primary antibody overnight at 4˚C as follows: goat anti-ionized calcium-binding adaptor protein-1 (Iba1) (1:300; Abcam, Cambridge, UK). An appropriate antibody was applied for 60 min at room temperature. After thoroughly washed, the reaction products were visualized using the DAB method. The sections were counterstained with hematoxylin, dehydrated, and mounted. Control samples were run in parallel omitting the primary antibody. The integrated optical density (IOD) of positively stained area was analyzed at 200× magnificationin CA1 region with image analysis software (Image-Pro Plus 6.0).

Statistical analysis
All data are presented as the mean ± SEM. Statistical comparisons were subjected to a multivariate analysis of variance (ANOVA) in which stress, surgery and intervention were dependent variables. Bonferroni's test was employed when ANOVA showed significance. A p-value < 0.05 was considered to be statistically significant.

Chronic unpredictable stress decreased the bodyweight of stressed rats
CUS exerted a negative effect on the bodyweight gain (Fig 1). While the controls maintained their weight through the protocol, the stressed animals lost 8.15% of weight (p < 0.001) during the 14-day CUS. Chronic unpredictable stress exacerbated surgery-induced spontaneous locomotor activity impairment and increased the levels of anxiety Two-way ANOVA of the total distance and the time in the central area in the open field test revealed significant effects of CUS (p = 0.010 and p = 0.003, respectively), surgery (p<0.001 and p<0.001, respectively) and CUS×surgery interaction (p = 0.022 and p = 0.017, respectively). There was no significant effect of the CUS on the total distance in the stressed animals compared with the controls 48 h post-stress (p = 0.794). The total distance was shorter in the surgery group compared to that in the control group on postoperative day 1 (p = 0.028). CUS produced an additive effect on the total distance in the surgical animals on postoperative days 1 and 3 (p < 0.001 and p = 0.019, respectively) (Fig 2A). Similarly, the time in the central area in the surgical rats was significantly shorter compared with the controls on postoperative day 1 (p < 0.001). Significant difference of the time in the central area was also observed between the surgery group and the CUS+surgery group on the postoperative days 1 and 3 (p = 0.021 and p = 0.003, respectively) ( Fig 2B). When pretreated with RU486, the distance moved was similar between the surgery group and the RU486+CUS+surgery group.

Chronic unpredictable stress exaggerated surgery-induced neuroinflammatory responses in the hippocampus
Surgical trauma significantly increased the levels of IL-1β and IL-6 on postoperative days 1 (p < 0.001 and p = 0.033, respectively) and 3 (p = 0.03 and p = 0.011, respectively). In the absence of surgical challenge, CUS failed to alter the levels of pro-inflammatory cytokines. However, CUS amplified surgery-induced pro-inflammatory cytokine IL-1β expression on postoperative day 3 (p = 0.002), and IL-1β expression returned to baseline on day 7 (p = 0.248) compared with the day-matched surgery group (Fig 4). RU486 alone failed to significantly alter hippocampal pro-inflammatory cytokine levels when compared to the naive controls at any time point. However, administered with RU486 (30 mg/kg) prior to CUS and surgery, the levels of IL-1β were similar in the RU486+CUS+surgery group and the surgery group. A similar pattern was observed for IL-6 protein (Fig 5). This indicated that RU486 substantially blunted the potentiating effects of CUS on surgery-induced pro-inflammatory processes in the hippocampus.
Chronic unpredictable stress upregulated surgery-induced microglial Iba-1 expression and reduced the mRNA levels of M2 phenotype marker Arg1 Compared to the naive controls, CUS failed to increase the expression of Iba-1 48 h post-stress. The levels of Iba-1 were significantly upregulated compared with the controls on postoperative days 1 (p = 0.037) and 3 (p = 0.002), and returned to baseline on day 7 (p = 0.209). Higher levels of Iba-1 were observed in the animals of CUS+surgery group compared with those of the surgery group on postoperative day 3 (p = 0.027). Pretreatment with RU486 blocked the effects of CUS on surgery-induced Iba-1 upregulation (Fig 6).
CUS failed to downregulate the levels of microglial M2 phenotype marker Arg1 in the stressed rats compared with the controls. The gene levels of Arg1 were significantly decreased following surgical challenge on postoperative day 1 (p < 0.001). CUS produced an additional reduction of Arg1 expression in the CUS+surgery group compared with the surgery group on postoperative days 1 and 3 (p < 0.001 and p < 0.001, respectively). Treatment with RU486 partially blunted the potentiating effects of CUS on surgery-induced downregulation of Arg1 expression in the hippocampus (Fig 7).
Chronic unpredictable stress and surgical trauma decreased the levels of CD200 mRNA in the hippocampus Compared to the controls, CD200 mRNA expression was downregulated in the CUS group at 48 and 96 h post-stress (p < 0.001 and p < 0.001, respectively). The levels of CD200 mRNA was significantly decreased on postoperative days 1 and 3 (p < 0.001 and p < 0.001, respectively), and improved on day 7 (p = 0.128). An additional significant decrease in hippocampal   RU486 attenuated the potentiating effects of CUS on surgery-induced downregulation of CD200 expression (Fig 8).

The combination of chronic unpredictable stress and surgical trauma reduced the levels of BDNF in the brain
Surgical trauma significantly decreased the levels of BDNF compared with the controls on postoperative day 1 (p < 0.001), while CUS alone failed to downregulate BDNF expression in the stressed rats 48 h post-stress ( Fig 9A). Furthermore, the lower levels of BDNF were  observed in the CUS+surgery group compared with the surgery group on postoperative day 1 (p = 0.002) (Fig 9A). Pretreatment of RU486 blunted these priming effects induced by CUS (Fig 9A).  Chronic unpredictable stress and surgical trauma increased the levels of plasma GCs ELISA analysis showed that exposure to CUS resulted in a significant increase in plasma GCs compared to the non-stressed rats 48 h (p < 0.001) after the last session of stress and the levels of GCs returned to baseline 96 h post-stress (Fig 10). Surgical trauma significantly increased the levels of plasma GCs on postoperative days 1 and 3 (p < 0.001 and p = 0.002, respectively) and returned to baseline on day 7 compared with the controls. CUS enhanced the increased levels of GCs on postoperative day 1 (p < 0.001) (Fig 10).

Discussion
Accumulating evidence has shown that both acute and chronic stress potentiated the sickness response to LPS administered 24 h post-stress [13][14][15]. In the present study, surgical trauma impaired spontaneous locomotor activities and increased the levels of anxiety in the adult rats. CUS exacerbated surgery-induced sickness behavior. There effects were blocked by pretreatment with RU486, indicating that GCs, at least partly, is involved in the stress-induced sensitization of behavior changes following surgical challenge. Importantly, these changes were accompanied by microglial activation and neuroinflammatory responses in the brain, which might mediate sickness behavior.
The severity of the surgery influences the magnitude of the immune response and has been shown to correlate with the degree of postoperative inflammation and sickness behavior [16,17]. Previous work has shown that the incidence of POCD in elderly patients after minor surgery (primarily laparoscopy) was significantly lower than those after cardiac and noncardiac major surgery suggesting that extent of surgery contributes to postoperative brain dysfunction [18]. A study by Hempenius et al also indicates that the severity of the surgical procedure is independent risk factors for postoperative delirium in elderly patients undergoing elective surgery [19]. Additionally, Rosczyk et al have demonstrated that locomotor activity is not depressed in both adult and aged mice following sham operation-minor abdominal surgery, which reveals that the decrease in locomotion is not due to minor surgery procedure [20]. It is likely that a more "major" surgery-partial hepatectomy would induce a state of neuroinflammation that could result in sickness behavior.
Pro-inflammatory cytokines inhibit hippocampal neuronal functions, including long-term potentiation (LTP) and dendritic branching, which are involved in memory formation and maintenance [21]. Stressors can induce two different types of inflammatory responses in the brain. The first is a rapid, short duration (several hours) increase in inflammatory mediators [22]. Brain levels of pro-inflammatory cytokines are elevated immediately after a moderate duration (2 h) of stress and persist for 4-6 h [23]. The second is a slower developing, longer lasting (days) sensitization (or ''priming") of neuroinflammatory responses to subsequently occurring infectious/pathogenic stimuli or stressors (i.e. delayed challenge) [13]. Acute and chronic stress has been found to sensitize neuroinflammatory responses to both peripheral and central immunologic challenges [13,24,25]. CUS alone failed to modulate the expression of pro-inflammatory cytokines 48 h post-stress. However, consistent with a growing body of evidence, this study demonstrated the priming effects of CUS in neuroinflammatory processes [26]. CUS amplified surgery-induced neuroinflammatory responses in the brain. Additionally, Barrientos et al demonstrated a single intracisternal administration of IL-1 receptor antagonist (IL-1RA) at the time of surgery was sufficient to block both the behavioral deficit and the neuroinflammatory response. Injecting the same dose of IL-1RA peripherally failed to have a protective effect [27]. These data provided strong support for the specific role of central, not peripheral, IL-1β in behavior deficits.
Microglia are primary immune cells and major source of pro-inflammatory cytokines in the central nervous system (CNS) [28,29]. Espinosa-Oliva et al found that microglia were neuroimmune substrates of peripheral stress-induced GCs action, which in turn ''prime" microglia to over-respond to subsequent challenge in the aged brain [24]. Notably, hippocampal microglia isolated 24 h post-stress also exhibited a potentiated neuroinflammatory response to LPS ex vivo, suggesting that central microglia instead of peripheral immunologic substrate(s) were sensitized or primed. In this study, CUS sensitized microglia to pro-inflammatory immunophenotype. When further stimulated in this state through a peripheral surgical challenge, microglia overexpressed pro-inflammatory cytokines (e.g., IL-1β and IL-6). These data suggest that stress-induced microglial priming contributes to neuroinflammatory responses to subsequent surgical challenge.
On the basis of gene expression profiles, activated microglia may be categorized into two opposite types: M1 phenotype and M2 phenotype [30]. M1 represents a detrimental state of microglia, characterized by high expression of pro-inflammatory cytokines (e.g. IL-1β, IL-6 and TNF-α). Conversely, M2 phenotype may reverse the neuron loss, repair neural networks and enhance the production of anti-inflammatory cytokines (e.g. IL-10) and neurotrophic mediators (e.g. BDNF) [30]. M1 and M2 polarization states of microglia play an important role in controlling the balance between pro-inflammatory and anti-inflammatory conditions. This study demonstrated that CUS exaggerated surgery-induced neuroinflammatory responses and exacerbated sickness behavior by inhibiting the function of M2-polarized microglia. CUS mediated the shift of the neuroimmune microenvironment toward a microglial M1 phenotype.
CD200 expressed on the surface of neurons is thought to maintain microglial cells in a quiescent state through interactions with its receptor CD200R [31]. The CNS parenchyma shows high levels of CD200 as well as very low levels of basal MHC-II and Iba-1, indicating the quiescent immunophenotype of the normal brain microenvironment [32]. The present study demonstrated that CUS downregulated hippocampal CD200 expression at 48 and 96 h post-stress. The effect of CUS on CD200 suggests that stress-induced sensitization of microglia may be mediated in part by attenuation of neuronal control of microglia.
GCs appear to play a pivotal role in CUS-induced potentiation of neuroinflammatory processes [9,33]. GCs modulate innate immune signaling pathways (e.g., Toll-like receptors and inflammasome formation) that are pivotal to generating a pro-inflammatory immune response [34]. In this study, treatment with GCs receptor antagonist RU486 blocked CUS-induced priming of hippocampus pro-inflammatory responses and microglial activation following surgical procedure. These data suggest that elevated GCs may mediate CUS-induced sensitization of neuroinflammatory processes triggered by surgical trauma.
Human studies provide direct evidence of cognitive impairment following high levels of GCs. A 4-day treatment with dexamethasone produced a selective impairment of declarative memory [35]. Cushing's syndrome is associated with cognitive impairment and hippocampal atrophy that is reversible following the lowering of cortisol levels after treatment [36]. Consistent with previous research, this study demonstrated that surgical trauma significantly increased the levels of serum GCs, which related to sickness behavior [37]. Thus, a body of knowledge suggests that GCs, at least partly, involves behavior deficits in a variety of settings.
Increased levels of GCs are almost universally considered to be anti-inflammatory [38]. However, the results in this study appeared contradictory. Stress-induced GCs failed to suppress neuroinflammatory responses following surgical procedure. Of particular relevance to the present study, the timing of stress exposure relative to an immune challenge was a critical parameter in determining the outcome. Frank et al found that GCs treatment prior to LPS potentiated the neuroinflammatory response, whereas GCs treatment after LPS blunted neuroinflammatory responses to LPS. Notably, Barnum et al showed that CUS as well as a chronic psychological stress blunted the neuroinflammatory response to LPS [39]. It is important to note that LPS was administered two weeks after the last stress. Whereas consistent with other studies, partial hepatectomy was performed 24 h post-stress in this study [7]. The timing of immunologic challenge relative to stressor offset may account for these discrepant findings.
GCs bind two different receptors: high-affinity mineralocorticoid receptor (MR) and the lower-affinity GR in the CNS. MR is heavily occupied basally and becomes saturated by GCs levels in the mild stress range, whereas GR is heavily occupied only after major stressors [8]. Because MR and GR signaling can have different transcriptional effects, basal and high-stress GCs levels can have divergent, even opposite effects [40]. Pretreatment with RU486 blocked CUS-induced microglia activation and neuroinflammatory responses in the brain, which indicated that GCs and GR involved in the deleterious effects of CUS.
BDNF is an important regulator of synaptic transmission and LTP in the brain, which is associated with learning and memory formation [41,42]. Administered neural injections of a BDNF antibody exacerbated cognitive deficits assessed by a Morris water maze [43]. Contrarily, exogenous BDNF improved the cognitive performance [44]. A growing body of evidence suggests that severe stress can suppress BDNF signaling, impair synaptic activity and increase susceptibility to affective disorders, resulting in neuronal atrophy and cognitive impairment [45]. Several evidence indicate that chronic stress and low level of BDNF are the major components of sickness behavior [46]. In this study, surgical trauma in combination with CUS inhibited BDNF expression in the brain, which was accompanied with sickness behavior. Surgery-induced behavioral deficits were mediated, in part, by downregulation of hippocampal BDNF expression. Stress affects neuroimmune system functions both directly and indirectly. Previous research has indicated that chronic stress induces inflammatory responses, cognitive impairments and regulates microglial activity in the brain region [47,48]. However, in this study, CUS alone failed to exacerbate sickness behavior, modulate the expression of pro-inflammatory cytokines and alter microglial Iba-1 and Arg 1 expression 48 h post-stress. Contradictory results may be considered to be due to different treatment protocols, such as animals (adult vs aged rats), stress variables [e.g., duration (14 vs 40 days) and frequency] and time point of evaluation (48 vs 2 hours) [7]. Furthermore, Bian et al provides evidence that CUS-induced impairments in spatial learning and memory and CUS-induced changes in glia cells could be reversible [49].
This study provides converging evidence that CUS reinforces the effect of surgical trauma challenges in a synergistic manner, creating an exaggerated sickness behavior and neuroinflammatory responses and inhibiting BDNF expression in this rodent model. Stress-induced microglial activation contributes to the sensitization of pro-inflammatory responses. GCs play a pivotal role in enhancing stress-induced neuroinflammatory responses by modulation of microglia functions.
Supporting information S1 Checklist. NC3Rs ARRIVE guidelines checklist. (PDF) S1 Dataset. Relevant data underlying the findings described in manuscript.

Author Contributions
Data curation: Na Wang.