EZH2 is involved in vulnerability to neuroinflammation and depression-like behaviors induced by chronic stress in different aged mice

Wei Wang, Xiaqing Qin, Rui Wang, Jingjing Xu, Huiran Wu, Arslan Khalid, Hong Jiang, Dexiang Liu, Fang Pan

A B S T R A C T

Cytokines Background: Microglial activation and pro-inflammatory cytokines expression is closely related to pathogenesis of depression. Aging is a known risk factor for neuroinflammation in the central nervous system and subsequent behavioral impairment. Enhancer of zeste homolog 2 (EZH2), a methyltransferase of histone H3 lysine 27 which regulates microglial activation, plays a crucial role in proinflammatory cytokines expression. However, whether the EZH2 is involved in susceptibility to depression in different ages remains elusive.
Methods: Young and aged C57BL/6 mice were exposed to chronic unpredictable mild stress for three weeks. Depression- and anxiety-like behaviors, spatial memory impairment, and the expression of pro-inflammatory cytokines, P-p65, EZH2, H3K27me3 and SOCS3 in the prefrontal cortex and hippocampus were measured using an established behavioral battery, ELISA, immunohistochemistry and western blotting techniques. Moreover, EPZ-6438, an inhibitor of EZH2, was utilized to detect the role of EZH2 in neuroinflammation and behavioral abnormalities.
Results: CUMS induced depression-like behaviors and spatial memory impairment, elevated levels of proinflammatory cytokines and P-p65, enhanced M1 microglia activation, and increased levels of EZH2, H3K27me3 and SOCS3 in the prefrontal cortex and hippocampus in young and aged mice. Both unstressed and stressed aged mice displayed attention-deficit behavioral outcomes, alteration of protein levels compared with young mice. However, inhibition of EZH2 could relieve most of behavioral and molecular alterations.
Limitations: A relative small sample size is a limitation.
Conclusions: EZH2 might be involved in susceptibility to neuroinflammation and depression-like behaviors in different aged mice.

Keywords:
CUMS
Depresion
Aging
EZH2
Microglia

1. Introduction

Depression is a mood disorder characterized by feelings of sadness, anhedonia (diminished capacity to experience pleasure), worthlessness, and thoughts of death (American Psychiatric Association, 2013). As a common psychiatric disorder, depression affects approximately 300 million people around the world (WHO, 2017). More importantly, the aged population has a higher prevalence of depression than young adults, and depression in aged patients normally lasts longer and has more severe symptoms (Diegelmann et al., 2016; Glaesmer et al., 2011). Previous studies suggested that cellular and molecular alterations, accompanied by neurobiological and neuropsychological changes in laterlife, contribute to progression of depression in a variety of ways (Naismith et al., 2012). It has been proposed that immune processes regulate changes in cognitive and emotional functions in later life (Alexopoulos and Morimoto, 2011). A recent epidemiologic study demonstrated that acute episodes of inflammation increase the risk of suicide in a proportional relationship (Lund-Sørensen et al., 2016). A recent meta-analysis study reported that it is likely that inflammation leads to depression in the elderly (Smith et al., 2018). Furthermore, preclinical studies on brains of aged rodents provided evidences of increased levels of several inflammatory cytokines, including tumor necrosis factor (TNF)-α, interleukin (IL)-1β, and IL-6 (Garg et al., 2016; Kuzumaki et al., 2010; Ye and Johnson, 1999).
Etiological factors of depression include abnormalities in the levels of neurotransmitters, neuroendocrine dysfunction, and lower level of brain-derived neurotrophic factor (BDNF) in the brain (Belmaker and Agam, 2008). Recently, studies have documented that neuroinflammation is association with dysregulated neurotransmission and abnormal neurocircuits on the one hand and increased risk for depression on the other. Microglia are the main regulator of this association (Jones and Thomsen, 2013; Yirmiya et al., 2015). Normally, microglia modulate neuronal circuitry and maintain neuronal function by phagocytosis and release of cytokines, chemokines, and prostaglandins, which are involved in pathological responses of impair and repair (Wolf et al., 2017). However, microglia in aged brains demonstrate a more inflammatory-related phenotype, with impaired regulation, which defines these cells as primed or sensitized (Barrientos et al., 2015). The primed microglia in aged brains result in amplified and extended chronic response to secondary maladies, including infection and psychological stress (Buchanan et al., 2008; d’Avila et al., 2018; von Leden et al., 2017).
Moreover, epigenetic changes, including histone modifications and DNA methylation, are considered to be involved in regulating cellular phenotype and reprogramming related to microglial plasticity in aging (Cheray and Joseph, 2018). Specifically, methylation of histone 3 on Lysine 27 is mediated by the polycomb repressive complex 2 (PRC2). PRC2 is composed of enhancer of zeste homolog 2 (EZH2), suppressor of zeste 12 homolog (SUZ12), and embryonic ectoderm development (EED). The complex is linked to silencing of gene expression (Cao et al., 2002). Previous studies have shown that EZH2 serves as a regulator of methylation in the neuroinflammatory process, and inhibition of EZH2 attenuates the expression of pro-inflammatory mediators in vivo and in vitro (Arifuzzaman et al., 2017; Chen et al., 2019; Yadav and Weng, 2017; Zhang et al., 2018). EPZ-6438 (tazemetostat) is a small molecule, acting as an efficient and selective inhibitor of EZH2. The molecule is a competitive inhibitor, selectively binding to S-adenosylmethionine (SAM), thus inhibiting trimethylation of histone 3 Lysine 27 (H3K27me3) (Knutson et al., 2013). Treatment with EPZ-6438 resulted in significant down-regulation of key inflammatory mediators, such as signal transducer and activator of transcription 1 (STAT1) and interferon regulatory factor 1 (IRF1). The increase in these inflammatory mediators is stimulated by lipopolysaccharide (LPS) in the microglial cells (Arifuzzaman et al., 2017). Moreover, it has been proposed that EZH2 directly targets and inhibits the expression of suppressor of cytokine signaling 3 (SOCS3), thus affecting activation of nuclear factor-κB (NF-κB) through one of the Toll-like receptor-associated pathways. Myeloid differentiation factor 88 (MyD88) leads to the activation of TNF receptor-associated factor 6 (TRAF6), which, in its turn activates NF-κB and the inflammatory response. SOCS3 is known to promote the proteosomal degradation of TRAF6 (Zhang et al., 2018). Studies have also shown that administration with 3-deazaneplanocin A (DZNep, an EZH2 inhibitor) could lead to a marked upregulation of SOCS3 expression during septic lung inflammation (Zhang et al., 2019). In particular, SOCS3 is involved in repressing the M1 proinflammatory phenotype and thus also the suppression of the inflammatory response in macrophage (Qin et al., 2012). Upregulation of SOCS3 could significantly inhibit the expression of proinflammatory cytokines and protect the nervous system from neuroinflammatory responses (Fan et al., 2018; Wu et al., 2016). Important to note that neuroinflammation is not only seen as a deleterious mechanism, but modulation of this over-inflammatory response could present interesting insights.
In the present study, we utilized chronic unpredictable mild stress (CUMS) to generate young and aged depression mice model, which were shown in a previous study to display chronic neuroinflammation in the stress-responsive zones of the brain (Farooq et al., 2012). The CUMS model has been shown to have a high degree of predictive, face, and construct validity and reliability (Willner, 2017), imitating psychiatric risk factors such as psychological, psychosocial, and physical stresses. The C57BL/6 mice were chosen to act as the depression model animal because they are more sensitive to CUMS than other mouse strains, such as KM, ICR, or BABL/c (Xu and Wang, 2016). Our hypothesis was that aging could accelerate CUMS-induced depression-like behaviors and cognitive decline, and increase proinflammatory cytokines (TNF-α, IL-1β, and IL-6), microglia M1 activation, and EZH2 expression in the prefrontal cortex and hippocampus. Secondly, the inhibitor of EZH2, which targets H3K27me3 and SOCS3, could restore behavioral impairment by reducing microglial activation and diminishing proinflammatory cytokines in mice of different ages. The aim of this study was to explore EZH2 involvement in alterations of depression-like behaviors and cognitive impairment, neuroinflammation, and epigenetic regulation in depression induced by chronic stress in different aged mice.

2. Materials and methods

2.1. Animals

Thirty young male C57BL/6 mice (2 months, 25-30 g) and thirty aged male C57BL/6 mice (12 months, 40-45 g) were purchased from the Beijing Vital River Laboratory Animal Technology Co., Ltd. The animals were housed in polypropylene cages (12 h light/12 h dark cycle, 23 ± 2°C) with water and food available ad libitum, except when stress was induced. Mice were housed single through the course of the experiments. All procedures were performed in accordance with Helsinki Declaration as revised in 1989, and were approved by the Animal Ethics Committee of Shandong University.

2.2. Grouping and treatment

Young and aged mice were randomly divided into six groups of ten mice per group: young control group (YC), young CUMS group (YS), young CUMS with EPZ-6438 group (YSE), aged control group (AC), aged CUMS group (AS) and aged CUMS with EPZ-6438 group (ASE). After adaptation for one week, mice in the stressed groups (YS, YSE, AS, ASE) received the CUMS procedure (see point 2.3 below) for 21 days. EPZ-6438 (1 mg/kg, i.p., Biochempartner, China) was dissolved in 0.1% dimethyl sulfoxide (DMSO) at a concentration of 0.1 mg/mL. This was administered once daily, for the 21 days of the CUMS procedure. Mice in the control and the CUMS alone groups were administrated daily with 0.1% DMSO to control for vehicle effects. The dosage of EPZ-6438 was based on a previous study (Arifuzzaman et al., 2017). All mice went in a sequential manner through the following tests: sucrose preference (SP), open field (OF), elevated plus maze (EPM), and Morris water maze (MWM; after training). Following the behavioral tests, animals were sacrifice for tissue preparation and in vitro analyses. The experimental procedure is shown in Figure 1.

2.3. The CUMS procedure

The CUMS procedure was slightly modified from the method previously described by Willner et al. (Willner et al., 1987). The following stressors were administrated in a semi-random order: white noise for 4 h (90 dB), food deprivation for 24 h, water deprivation for 24 h, wet bedding overnight, lights on overnight, 45°C heat stress for 5 min, cage shaking for 10 min. The stressed animals were exposed to various stressor for three weeks as described in Table 1. Control mice were under comparable handling without stress.

2.4. Behavioral tests

2.4.1. The sucrose preference test

The SPtest was used to operationally define anhedonia (Willner et al., 1987). Individually housed mice of both age groups were given two bottles, one with tap water and the other with 1% sucrose solution. The position of the two bottles was switched (left to right, right to left) each day to ensure that the mice did not develop a side preference. After 5 days of habituation, mice were deprived of water and food for 23 h, and were then given the two bottles freely for 2 hours. To prevent the possible effects of a side preference in drinking behavior, the position of the bottles in the cage was switched after one hour during the test. The bottles were weighed to calculate the liquid consumption. Sucrose preference (%) was expressed as = sucrose consumption / (sucrose consumption + water consumption), with depressive mice drinking less sucrose.

2.4.2. The open field test

The OF test was used to assess basic activities and anxiety-like behavior (Prut and Belzung, 2003). The open field is a white wooden open field box (40 × 40 × 40 cm). The arena surface was cleaned with 75% ethanol to remove odors left by former mice. The mice were individually placed in the center of the field and allowed to move freely for five minutes. Subsequently, the mice activities were recorded by a video tracking software (SMART 2.5, Panlab Harvard Apparatus, Spain). Horizontal locomotion (the number of sectors visited with at least three paws), rearing frequency (lifting two forepaws from the ground), and time spent in the central area and border area were analyzed to assess anxiety-like behavior, with anxious mice normally spending less time in the central area.

2.4.3. The elevated plus maze test

The EPM test was designed to test the animals’ anxiety-like behavior by examining the frequency of entries into the open arms of the maze and the duration of time spent in them (Walf and Frye, 2007). The device was elevated 50 cm above the floor, and it had two open arms and two closed arms with opaque walls that were connected by a square platform. The mice were placed in the center of the platform with their head facing an open arm, and allowed to explore the maze for five minutes. The device was cleaned with 75% ethanol before each trial. Number of entries into and the time spent in the open arms and closed arms were recorded with SMART, with anxious mice normally spending less time and registering fewer entries into the open arms.

2.4.4. The Morris water maze test

The MWM test was performed to assess spatial learning and memory function (Illouz et al., 2016; Morris, 1984). A water pool (120 cm in diameter) was filled with water and a platform (13 cm in diameter) was hidden 1 cm below the surface. The water was made opaque white with titanium dioxide. Mice were trained to escape from the water by swimming from a semi-random set of starting positions, and were allowed to stay on the platform for 15 s when reaching it. Subjects that were unable to find the platform during trials were placed on the platform by an experimenter and were left there for 15 s. All mice underwent five days of training, four sessions per day. A 60 s probe trial was performed on the day following the training period, with the platform removed. Trajectories of the mice were recorded and analyzed by SMART. The pool was divided in the analyzing software into four quadrants, and the quadrant in which the platform was previously located was defined as the target quadrant. Entries and time spent in the target quadrant were measured by the software. The mice with spatial impairment may spending less time and show fewer entries into the target quadrant.

2.5. Enzyme-linked immunosorbent assay (ELISA)

After the behavioral tests, the mice were sacrificed and the prefrontal cortex (PFC) and hippocampus were rapidly dissected for the biochemical tests. The tissues were not perfused. Dissected tissues were weighed and then homogenized completely in phosphate-buffered solution (PBS) with phenylmethylsulfonyl fluoride (PMSF, Beyotime Biotechnology, China). After centrifugation (10000 × g, 5 min, 4°C) the supernatants were collected immediately and stored at -80°C pending evaluation. Concentrations of TNF-α, IL-1β, and IL-6 were measured, using the ELISA kit, according to the manufacturer’s instruction (Tianjin Anoric Bio-technology, Co., Ltd., China). Total protein was determined by the bicinchoninic acid (BCA) assay kit (Beyotime Biotechnology, China). The levels of the cytokines were expressed in pg/mg total protein.

2.6. Immunohistochemistry (IHC)

Mice were deeply anesthetized with pentobarbital sodium (50 mg/ kg, i.p.) and transcardially perfused with cold normal saline and then cold 4% paraformaldehyde dissolved in PBS. Their brains were then dissected out and further fixed in 4% paraformaldehyde for 24 h at 4°C. Tissues were dehydrated, cleared and embedded in paraffin. The blocks with the tissues were sectioned serially at 5 μm, using a microtome, and then mounted on slides. The sections were deparaffinated, rehydrated, and then immersed in sodium citrate (PH = 6) for antigen retrieval, using a microwave oven. The sections were then treated with 3% H2O2 for 10 min to eliminate endogenous peroxidase. After washing three times with PBS, the sections were blocked with 5% bovine serum albumin (BSA) in PBS for 30 min and then incubated with primary antiionized calcium binding adaptor molecule 1 (IBA-1) antibody (ab5076, 1:400, Abcam, USA) or anti-inducible nitric oxide synthase (iNOS) antibody (ab15323, 1:400, Abcam, USA) overnight at 4°C. After three PBS washes, these sections were incubated with biotinylated-conjugated rabbit anti-goat IgG secondary antibody (SA1022, Boster, China) for 30 min at 37°C. After three PBS washes, sections were treated with Streptavidin-Biotin Complex for 30 min at 37°C and then washed three times with PBS. A diaminobezidine (DAB) kit (AR1022, Boster, China) was used for visualization, then sections were stained with haematoxylin and covered with cover slips. Images were captured using camera connected to an OLYMPUS microscope (BX53, Japan). The number of positive cells was calculated using ImageJ software, version 1.50i (NIH, USA). Results are expressed as the number of positive cells per mm2.

2.7. Western blotting

Twenty-four hours after behavioral tests, the PFC and hippocampus were carefully dissected from the mice for western blot assays. The tissues were not perfused. The samples were homogenized with radioimmunoprecipitation assay (RIPA) buffer containing PMSF (10 µL/mL, Beyotime Biotechnology, China). The homogenates were centrifuged at 14,000 × g for 10 min at 4°C, and the supernatants were collected and aliquoted for further analysis. The concentration of total protein was determined using a BCA protein assay kit (Beyotime Biotechnology, China). Protein samples (30 μg) were separated using sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), and then electrophoretically transferred onto polyvinylidene fluoride (PVDF) membranes (Bio-Rad, USA). Membranes were blocked with 5% no fat milk in tris-buffered saline with Tween 20 detergent (TBST) for 1 h and then incubated overnight at 4°C with primary antibodies against EZH2 (A16846, 1:1,000, Abclonal, China), H3K27me3 (BS7237, 1:1,000, Bioworld, China), NF-κB p65 (ab32536, 1:1,000, Abcam, USA), Phosphorylated NF-κB p65 (S536) (AP0124, 1:1,000, Abclonal, China) or SOCS3 (A0769, 1:1,000, Abclonal, China) and an internal control GAPDH (AF1186, 1:5,000, Beyotime Biotechnology, China). After three TBST washes, the membranes were incubated with a horseradish peroxidase-conjugated secondary antibody, sheep anti-rabbit IgG (1:10,000, Beyotime Biotechnology, China) for 1 h at room temperature. After washing three times with TBST, the membranes were incubated with chemiluminescence substrates (Millipore Corp, USA) and exposed inside a luminescent image workstation (Tanon Science & Technology Co., Ltd, China). The grey value of the bands was quantified using ImageJ.

2.8. Statistical analysis

Data were analyzed with SPSS Statistics 23.0 (IBM Corp., USA). In most cases, statistical significance was determined by two-way ANOVA with Bonferroni corrected post hoc test or by independent t-test. For the MWM test, the average escape latency was evaluated by three-way or one-way repeated-measures ANOVA. Differences were considered statistically significant if P < 0.05. All the quantitative data are presented as mean ± standard error of the mean (SEM).

3. Results

3.1. The effect of CUMS and inhibition of EZH2 on depression- and anxietylike behaviors and spatial learning and memory

3.1.1. Depression- and anxiety-like behaviors

As shown in Fig. 2A, in terms of sucrose consumption, two-way ANOVA analysis indicated a significant effect for CUMS [F (1, 36) = 24.15, P < 0.001] and age [F (1, 36) = 22.52, P < 0.001]. No effect was observed for exposure-age interaction in sucrose consumption. Age was a factor in the dramatic decrease in preference of sucrose in the SP test. CUMS with EPZ-6438 groups (young and aged) had consumed significantly more sucrose than CUMS groups, respectively.
Fig. 2B-F shows the results in the OF test. Two-way ANOVA analysis showed a significant effect for CUMS [F (1, 36) = 14.1, P < 0.001, Fig. 2B, F (1, 36) = 12.66, P < 0.01, Fig. 2C, and F (1, 36) = 8.92, P < 0.01, Fig. 2D] and age [F (1, 36) = 14.1, P < 0.001, F (1, 36) = 11.39, P < 0.01, and F (1, 36) = 12.27, P < 0.01] on the number of crossing, total distance, and rearing number, respectively. No effect was observed for the interaction between exposure and age in the crossing number, total distance and rearing number. Age was a factor contributing to a dramatic decrease in the crossing number, total distance and rearing number in the OF test. Young CUMS with EPZ-6438 group had significantly more crossing number and total distance when compared to the young CUMS group, and aged CUMS with EPZ-6438 group had significantly more rearing number than aged CUMS group.
Two-way ANOVA analysis showed a significant effect for CUMS [F (1, 36) = 21.92, P < 0.001, Fig. 2E and F (1, 36) = 21.92, P < 0.001, Fig. 2F] and age [F (1, 36) = 12.25, P < 0.01 and F (1, 36) = 12.25, P < 0.01] on central time and border time in the OF test, respectively. No effect was observed for exposure-age interaction in central time and border time. Age was a factor in the dramatic decrease in the time in the central area and increase in the time in the border area in the OF test. CUMS with EPZ-6438 groups (young and aged) spent significantly more time in the central and less time in the border area than CUMS groups, respectively.
Fig. 2G-H show the results of the EPM test. In terms of time spent in the open arms and entry into them, two-way ANOVA analysis indicated a significant effect for CUMS [F (1, 36) = 20.64, P < 0.001, Fig. 2G and F (1, 36) = 19.8, P < 0.001, Fig. 2H] and age [F (1, 36) = 8.41, P < 0.01 and F (1, 36) = 4.53, P < 0.05], respectively. No effect was observed for exposure-age interaction. Age was a factor in the dramatic decrease in the time spent in the open arms and entries into them in the EPM test. CUMS with EPZ-6438 groups (young and aged) spent significantly more time and had a higher number of entries into the open arms than the CUMS groups, respectively.

3.1.2. Spatial learning and memory

As shown in Fig. 3A, in terms of escape latency, three-way repeatedmeasures ANOVA analysis indicated a significant effect for CUMS [F (1, 36) = 8.75, P < 0.01] and age [F (1, 36) = 23.22, P < 0.001]. No effect was observed for the interaction between CUMS, age, and days. No effect was observed for the interactions of CUMS and age, CUMS and days, or age and days. One-way repeated measures ANOVA analysis showed that the young and aged CUMS with EPZ-6438 groups had significantly shorter latency compared to the young and aged CUMS groups [F (1, 18) = 10.36, P < 0.01 and F (1, 18) = 5.61, P < 0.05, respectively].
Fig. 3B shows representative trajectories in different groups on the test day in the MWM test. Two-way ANOVA analysis indicated a significant effect for CUMS [F (1, 36) = 22.84, P < 0.001, Fig. 3C and F (1, 36) = 18.77, P < 0.001, Fig. 3D] and age [F(1, 36) = 4.21, P < 0.05 and F (1, 36) = 4.89, P < 0.05] on the time and entries in the target quadrant, respectively. No effect was observed for exposure-age interaction on the time and entries in the target quadrant. Age was a factor contributing to a dramatic decrease in the time and entries in the target quadrant in the MWM test. CUMS with EPZ-6438 groups (young and aged) spent significantly more time and had a higher number of entries in the target quadrant than CUMS groups, respectively.

3.2. The effect of CUMS and inhibition of EZH2 on the level of proinflammatory cytokines and activation of NF-κB in the PFC and hippocampus

Two-way ANOVA analysis suggested a significant effect for CUMS [F (1, 12) = 29.28, P < 0.001, Fig. 4A and F (1, 12) = 18.9, P < 0.001, Fig. 4B] and age [F (1, 12) = 20, P < 0.001 and F (1, 12) = 39.19, P < 0.001] on the level of TNF-α in the PFC and hippocampus, respectively. No effect was observed for exposure-age interaction on the level of TNFα in the PFC or hippocampus. Age was a factor contributing to a dramatic increase in the level of TNF-α in both the PFC and hippocampus. The CUMS with EPZ-6438 groups had lower level of TNF-α in the PFC than CUMS groups in the young and aged, respectively. The aged CUMS with EPZ-6438 group also had lower level of TNF-α in the hippocampus when compared to the aged CUMS group.
Two-way ANOVA analysis indicated a significant effect for CUMS [F (1, 12) = 34.88, P < 0.001, Fig. 4C and F (1, 12) = 35.49, P < 0.001, Fig. 4D] and age [F (1, 12) = 10.92, P < 0.01 and F (1, 12) = 27.8, P < 0.001] on the level of IL-1β in the PFC and hippocampus, respectively. No effect was observed for exposure-age interaction on the level of IL1β in the PFC or hippocampus. Age was a factor contributing to a dramatic increase in the level of IL-1β in both the PFC and hippocampus. The young CUMS with EPZ-6438 group had lower level of IL1β in the PFC when compared to the young CUMS group. The CUMS with EPZ-6438 groups had lower level of IL-1β in the hippocampus than CUMS groups in the young and aged, respectively.
Two-way ANOVA analysis showed a significant effect for CUMS [F (1, 12) = 74.73, P < 0.001, Fig. 4E and F (1, 12) = 19.17, P < 0.001, Fig. 4F] and age [F (1, 12) = 22.67, P < 0.001 and F (1, 12) = 23.76, P < 0.001] on the level of IL-6 in the PFC and hippocampus, respectively. No effect was observed for exposure-age interaction on the level of IL-6 in the PFC or hippocampus. Age was a factor contributing to a dramatic increase in the level of IL-6 in both the PFC and hippocampus. The CUMS with EPZ-6438 groups had lower level of IL-6 in both the PFC and hippocampus than CUMS groups in the young and aged, respectively.
Two-way ANOVA analysis indicated a significant effect for CUMS [F (1, 16) = 53.36, P < 0.001, Fig. 4G and F (1, 16) = 42.1, P < 0.001, Fig. 4H] and age [F (1, 16) = 10.65, P < 0.01 and F (1, 16) = 25.3, P < 0.001] on the level of phosphorylated NF-κB p65 (P-p65) in the PFC and hippocampus, respectively. There was a significant effect for CUMS-age interaction [F (1, 16) = 5.22, P < 0.05] in the level of P-p65 in the PFC.
Age was a factor contributing to a dramatic increase in the level of Pp65 in both the PFC and hippocampus. The CUMS with EPZ-6438 groups had lower level of P-p65 in both the PFC and hippocampus than CUMS groups in the young and aged, respectively.

3.3. The effect of CUMS and inhibition of EZH2 on M1 microglial activation in the PFC, and the dentante gyrus (DG) and cornu ammonis 1 (CA1) sub-region of the hippocampus

Fig. 5A-B shows representative images of the expression of IBA-1 and iNOS in the PFC, and the sub-regions of the hippocampus -DG and -CA1. Two-way ANOVA analysis showed a significant effect for CUMS [F (1, 12) = 30.15, P < 0.001, Fig. 5C, F (1, 12) = 39.16, P < 0.001, Fig. 5D, and F (1, 12) = 14.59, P < 0.01, Fig. 5E] and age [F (1, 12) = 13.37, P < 0.01, F (1, 12) = 9.09, P < 0.01, and F (1, 12) = 35.47, P < 0.001] on the number of IBA-1-positive cells in the PFC, DG, and CA1, respectively. No effect was observed for the interaction between exposure and age in the number of IBA-1-positive cells in the PFC, DG or CA1 regions. Age was a factor contributing to a dramatic increase in the number of IBA-1-positive cells in the PFC, DG and CA1. CUMS with EPZ-6438 groups had fewer IBA-1-positive cells in the PFC, DG and CA1 when compared to the CUMS groups in both the young and aged groups.
Two-way ANOVA analysis showed a significant effect for CUMS [F (1, 12) = 34, P < 0.001, Fig. 5F, F (1, 12) = 22.04, P < 0.001, Fig. 5G, and F (1, 12) = 12.89, P < 0.01, Fig. 5H] and age [F (1, 12) = 12.4, P < 0.01, F (1, 12) = 21, P < 0.001, and F (1, 12) = 8.25, P < 0.01] on the number of iNOS-positive cells in the PFC, DG, and CA1, respectively. No effect was observed for the interaction between exposure and age in the number of iNOS-positive cells in the PFC, DG and CA1. Age was a factor contributing to a dramatic increase in the number of iNOS-positive cells in the PFC, DG and CA1. CUMS with EPZ-6438 groups had fewer iNOSpositive cells in the PFC, DG and CA1 when compared to the CUMS groups in both the young and aged groups.

3.4. The effect of CUMS and inhibition of EZH2 on the expression of EZH2, H3K27me3 and SOCS3 in the PFC and hippocampus

Fig. 6A-B shows representative images of EZH2 expression and the level of H3K27me3 in the PFC and hippocampus. Two-way ANOVA analysis showed a significant effect for CUMS [F (1, 16) = 39.23, P < 0.001, Fig. 6C and F (1, 16) = 57.25, P < 0.001, Fig. 6D] and age [F (1, 16) = 40.84, P < 0.001 and F (1, 16) = 56.54, P < 0.001] on the level of EZH2 in the PFC and hippocampus, respectively. No effect was observed for the exposure-age interaction on the level of EZH2 in the PFC and hippocampus. Age was a factor contributing to the dramatic increase in the level of EZH2 in the PFC and hippocampus. CUMS with EPZ-6438 groups had lower level of EZH2 in both the PFC and hippocampus when compared to the CUMS groups in the young and aged, respectively.
Two-way ANOVA analysis indicated a significant effect for CUMS [F (1, 16) = 21.55, P < 0.001, Fig. 6E and F (1, 16) = 45.37, P < 0.001, Fig. 6F] and age [F (1, 16) = 43.36, P < 0.001 and F (1, 16) = 54.09, P < 0.001] on the level of H3K27me3 in the PFC and hippocampus, respectively. No effect was observed for the exposure-age interaction on the level of H3K27me3 in the PFC and hippocampus. Age was a factor contributing to the dramatic increase in the level of H3K27me3 in the PFC and hippocampus. CUMS with EPZ-6438 groups had lower level of H3K27m3 in both the PFC and hippocampus when compared to the CUMS groups in the young and aged, respectively.
Two-way ANOVA analysis showed a significant effect for CUMS [F (1, 16) = 31.48, P < 0.001, Fig. 6G and F (1, 16) = 23.31, P < 0.001, Fig. 6H] and age [F (1, 16) = 36.16, P < 0.001 and F (1, 16) = 23.31, P < 0.001] on the level of SOCS3 in the PFC and hippocampus, respectively. No effect was observed for the exposure-age interaction on the level of SOCS3 in the PFC and hippocampus. Age was a factor contributing to the dramatic decrease in the level of SOCS3 in both the PFC and hippocampus. CUMS with EPZ-6438 groups had higher level of SOCS3 in both the PFC and hippocampus when compared to the CUMS groups in the young and aged, respectively.

4. Discussion

Aging and chronic stress are the two most important risk factors for depression due to chronic neuroinflammatory microenvironment in the brain. The present study demonstrates that three weeks of CUMS was able to induce depression-like behaviors in both young and aged mice. The behavioral changes were accompanied by a significant increase in proinflammatory cytokines and NF-κB, activation of microglia, higher expression of EZH2 and SOCS3, and increase in H3K27me3 in the PFC and hippocampus in stressed young and aged animals. However, treatment with EPZ-6438 was able to attenuate behavioral alterations, proinflammatory cytokines and P-p65 expression, activation of microglia, expression of EZH2, and the rate of H3K27me3, but increased the activation of SOCS3 in both young and aged animals.
Mounting evidence suggests that aging in rodents increases their susceptibility to developing anhedonia (Herrera-Pérez et al., 2008; Malatynska et al., 2012). Furthermore, previous studies have shown that infection could induce severe or prolonged depression-like behaviors in aged animals (Godbout et al., 2008; Kelley et al., 2013), which supports the conclusion that aging can exacerbate the effects of stress or infection on behavioral functions. Our results indicate that three weeks of CUMS induced more severe depression- and anxiety-like behaviors in aged mice than in young ones, suggesting that aged mice are more vulnerable to chronic stress. Interestingly, the inhibitor of EZH2 could attenuate behavioral deficits in mice of both age groups, indicating that EZH2 is an underlying regulator of depressive symptoms.
Cytokines in the brain are strongly associated with the pathophysiology of depression (Kim et al., 2016; Miller and Raison, 2016). For instance, TNF-α directly or indirectly affects the hypothalamic-pituitary-adrenocortical (HPA) axis, monoamine neurotransmitters, and genetic polymorphisms that are involved in the pathogenesis of depressive disorders (Ma et al., 2016). Patients with depression have higher levels of serum IL-1β, and targeting IL-1β can be useful in the treatment of subgroups of patients with depression (Ellul et al., 2016). Sukoff Rizzo and colleagues found that blocking IL-6 trans-signaling prevents depression-like behaviors in rodents (Sukoff Rizzo et al., 2012). More importantly, previous studies have found that acute systemic inflammation caused severe and prolonged neuroinflammation and behavioral deficits in aged mice (d’Avila et al., 2018; Godbout et al., 2005). Likewise, excessive and sustained elevation of IL1β, caused by peripheral infection, was responsible for memory impairment in aged rodents (Barrientos et al., 2009). In the present study, CUMS induced a significant increase in the levels of TNF-α, IL-1β, and IL-6 in the PFC and hippocampus in both young and aged mice, but the cytokines levels in aged mice were higher. Our results indicate that higher levels of proinflammatory cytokines in the brain, induced by CUMS, are associated with vulnerability to depression in aged mice. Specifically, our results show that not all the proinflammatory cytokines increased significantly in the PFC and hippocampus after CUMS, in agreement with our previous study in which we showed that proinflammatory cytokines have region-specific expression in brain under stress (Wang et al., 2018).
Previous studies have suggested that the neuroinflammatory state is primarily determined by microglia regulation (Wolf et al., 2017), and that aged microglia in a sensitive state might release unbalanced proinflammatory cytokines on a second assault (Cornejo and von Bernhardi, 2016). In our study, CUMS resulted in a higher number of IBA-1- and iNOS-positive cells in the PFC and hippocampus in aged stressed mice, when compared with young stressed mice. Our results suggest that aging can accelerate the rate of microglia transformation into the proinflammatory state, which supports the conclusion that primed microglia in the aged amplify their M1 activation under stress, leading to a harsh proinflammatory environment in the PFC and hippocampus.
Studies recently have shown that histone modification is one of critical modifier in regulating neuroinflammation and microglial activation. Nerve injury drastically increases the levels of EZH2 and H3K27me3, and the number of microglia with EZH2 expression in the spinal dorsal horn in rodents (Yadav and Weng, 2017). Ischemic/reperfusion injury upregulates the level of EZH2 in microglia in vivo (Chen et al., 2019). LPS can induce a higher level of EZH2 mRNA in primary and BV2 microglial cell lines (Arifuzzaman et al., 2017). Meanwhile, EPZ-6438 or DZNep, inhibitors of EZH2 could inhibit inflammatory-related gene expression and microglial activation (Arifuzzaman et al., 2017; Chen et al., 2019; Yadav and Weng, 2017). In addition, previous studies have found elevation of H3K27me3 in the midbrain of aged mice compared to young mice, which indicates the occurrence of epigenetic alterations with age (Tang et al., 2014). More importantly, Hai-Liang and colleagues found that downregulation of EZH2 could attenuate anxiety-like behaviors associated with the loss of miR-137 in mice, and indicated that EZH2 might be a potential therapeutic target for depressive and anxiety phenotypes (Yan et al., 2019). Consistent with those studies, we found that stress induced higher levels of EZH2 and H3K27me3 in the PFC and hippocampus in both young and aged mice. More importantly, both stressed and unstressed aged mice had higher levels of EZH2 and H3K27me3 in the PFC and hippocampus than stressed and unstressed young mice, and inhibition of EZH2 could downregulate the activation of microglia and level of proinflammatory cytokines in the PFC and hippocampus in young and aged mice induced by stress. These results indicate that EZH2 plays crucial roles in both the pathogenesis of CUMS-induced depression and age-related vulnerability to depression.
It is well known that SOCS3, an anti-inflammatory mediator, has a broad range of effects, such as down-regulating the NF-κB signaling pathway, antagonizing cAMP-mediated signaling, and enhancing Ras/ mitogen-activated protein kinase (MAPK) signaling pathway (Yin et al., 2015). Studies have demonstrated that elevated SOCS3 can suppress the expression of iNOS and proinflammatory cytokines in LPS-stimulated BV2 microglial cells (Chakrabarti et al., 2018) and the activation of microglia and release of IL-1β and TNF-α in vivo (Zhang et al., 2017). A recent study has documented the involvement of SOCS3 in the polarization of M1 proinflammatory phenotype macrophage (Qin et al., 2012). In our study, CUMS was able to increase the expression of SOCS3 in the PFC and hippocampus, which suggests that CUMS can induce expression of SOCS3. However, slightly activating SOCS3 is not enough to counter the proinflammatory mediators. Moreover, previous studies have found that SOCS3 is significantly lower in the cortex of aged mice with traumatic brain injury (TBI) when compared with young TBI mice (Kumar et al., 2013). RNA sequencing revealed that SOCS3 mRNA in microglial cells sorted from the brain decreases significantly with age (O’Neil et al., 2018). In the present study, aged mice had lower induction of SOCS3 compared to young mice, which is consistent with a previous conclusion stating that the ability to suppress neuroinflammation decreases with age, and this may be one reason for the higher level of neuroinflammation found in aged mice at baseline or following assault. In a previous study, inhibition of EZH2 was shown to cause upregulation of SOCS3 and reduce inflammation in mice with septic lung injury (Zhang et al., 2019). In the present study, inhibition of EZH2 upregulated SOCS3 and downregulated P-p65, TNF-α, IL-1β, and IL-6 in the PFC and hippocampus. Our results suggest that inhibition of EZH2 reduces neuroinflammation and depression-like behaviors by removing the inhibition EZH2 exerts on SOCS3 at the epigenetic level (Fig. 7). Our results support the conclusion in a previous report where it was suggested that EZH2-mediated H3K27me3 directly targets SOCS3 and inhibits its transcription (Zhang et al., 2018).

5. Limitations

The present study has limitations. First, based on the ethics approval we received, young and aged groups treated with EPZ-6438 alone could not be included in the current study. Second, due to the small sample size, we were unable to analyze the ratio of resilience and vulnerability, although such analysis could provide more evidence for our conclusions (Rutter, 2006). Lastly, a previous study has shown that inhibition of EZH2 induces expression of M2-related genes, such as arginase-1 and CD206 (Edward et al., 2014). Further research would need to explore the role of EZH2 in the M2 microglial activation and M2-related genes expression in the rodent depression model.

6. Conclusion

Present study demonstrated that CUMS induced depression-like behaviors, elevated P-p65 and proinflammatory cytokines, activated microglia, and increased the level of EZH2, H3K27me3, and SOCS3. Additionally, aging exacerbated behavioral deficits, cytokines release, activation of microglia, elevated P-p65, EZH2, and H3K27me3, and repressed SOCS3. Moreover, EPZ-6438, the inhibitor of EZH2, attenuated these behavioral deficits and molecular alterations in different aged mice. The study concludes that EZH2 is involved in the susceptibility to neuroinflammation and depression-like behaviors in the prefrontal cortex and hippocampus in different aged mice.

References

Alexopoulos, G.S., Morimoto, S.S., 2011. The inflammation hypothesis in geriatric depression. International Journal of Geriatric Psychiatry 26, 1109–1118. https://doi. org/10.1002/gps.2672.
American PsychiatricAssociation, 2013. Diagnostic and Statistical Manual of Mental Disorders (5th ed.).
Arifuzzaman, S., Das, A., Kim, S.H., Yoon, T., Lee, Y.S., Jung, K.H., Chai, Y.G., 2017. Selective inhibition of EZH2 by a small molecule inhibitor regulates microglial gene expression essential for inflammation. Biochemical Pharmacology 137, 61–80. https://doi.org/10.1016/j.bcp.2017.04.016.
Barrientos, R.M., Frank, M.G., Hein, A.M., Higgins, E.A., Watkins, L.R., Rudy, J.W., Maier, S.F., 2009. Time course of hippocampal IL-1 β and memory consolidation impairments in aging rats following peripheral infection. Brain, Behavior, and Immunity 23, 46–54. https://doi.org/10.1016/j.bbi.2008.07.002.
Barrientos, R.M., Kitt, M.M., Watkins, L.R., Maier, S.F., 2015. Neuroinflammation in the normal aging hippocampus. Neuroscience 309, 84–99. https://doi.org/10.1016/j. neuroscience.2015.03.007.
Belmaker, R.H., Agam, G., 2008. Mechanisms of disease: major depressive disorder. N.Engl. J. Med. 358, 55–68. https://doi.org/10.1056/NEJMra073096.
Buchanan, J.B., Sparkman, N.L., Chen, J., Johnson, R.W., 2008. Cognitive and neuroinflammatory consequences of mild repeated stress are exacerbated in aged mice. Psychoneuroendocrinology 33, 755–765. https://doi.org/10.1016/j.psyneuen.2008.02.013.
Cao, R., Wang, L.J., Wang, H.B., Xia, L., Erdjument-Bromage, H., Tempst, P., Jones, R.S., Zhang, Y., 2002. Role of histone H3 lysine 27 methylation in polycomb-group silencing. Science 298, 1039–1043. https://doi.org/10.1126/science.1076997.
Chakrabarti, S., Jana, M., Roy, A., Pahan, K., 2018. Upregulation of suppressor of cytokine signaling 3 in microglia by cinnamic acid. Curr Alzheimer Res 15, 894–904. https://doi.org/10.2174/1567205015666180507104755.
Chen, J., Zhang, M., Zhang, X., Fan, L., Liu, P., Yu, L., Cao, X., Qiu, S., Xu, Y., 2019. EZH2 inhibitor DZNep modulates microglial activation and protects against ischaemic brain injury after experimental stroke. European Journal of Pharmacology 857, 172452.

https://doi.org/10.1016/j.ejphar.2019.172452.

Cheray, M., Joseph, B., 2018. Epigenetics control microglia plasticity. Front. Cell. Neurosci. 12
Cornejo, F., von Bernhardi, R., 2016. Age-dependent changes in the activation and regulation of microglia. In: VonBernhardi, R. (Ed.), Glial cells in health and disease of the CNS. Springer International Publishing Ag, Cham, pp. 205–226.
d’Avila, J.C., Siqueira, L.D., Mazeraud, A., Azevedo, E.P., Foguel, D., Castro-Faria-Neto, H.C., Sharshar, T., Chrétien, F., Bozza, F.A., 2018. Age-related cognitive impairment is associated with long-term neuroinflammation and oxidative stress in a mouse model of episodic systemic inflammation. J Neuroinflammation 15, 28. https://doi. org/10.1186/s12974-018-1059-y.
Diegelmann, M., Schilling, O.K., Wahl, H.-W., 2016. Feeling blue at the end of life: trajectories of depressive symptoms from a distance-to-death perspective. Psychology and Aging 31, 672–686. https://doi.org/10.1037/pag0000114.
Edward, Koellhoffer, Rodney, Ritzel, Jeremy, Grenier, Rodney, Ritzel, Louise, McCullough, 2014. Tilting the balance of M1/M2 microglial phenotypes: inhibition of Ezh2 leads to decreased M1 and enhanced M2 phenotype. Journal of Neuroimmunology 275 (1–2), 157. https://doi.org/10.1016/j.jneuroim.2014.08. 421.
Ellul, P., Boyer, L., Groc, L., Leboyer, M., Fond, G., 2016. Interleukin-1 β-targeted treatment strategies in inflammatory depression: toward personalized care. Acta Psychiatrica Scandinavica 134, 469–484. https://doi.org/10.1111/acps.12656.
Fan, Y.-X., Qian, C., Liu, B., Wang, C., Liu, H., Pan, X., Teng, P., Hu, L., Zhang, G., Han, Y., Yang, M., Wu, X.-F., Liu, W.-T., 2018. Induction of suppressor of cytokine signaling 3 via HSF-1-HSP70-TLR4 axis attenuates neuroinflammation and ameliorates postoperative pain. Brain, Behavior, and Immunity 68, 111–122. https://doi.org/10.1016/j.bbi.2017.10.006.
Farooq, R.K., Isingrini, E., Tanti, A., Le Guisquet, A.-M., Arlicot, N., Minier, F., Leman, S., Chalon, S., Belzung, C., Camus, V., 2012. Is unpredictable chronic mild stress (UCMS) a reliable model to study depression-induced neuroinflammation? Behavioural Brain Research 231, 130–137. https://doi.org/10.1016/j.bbr.2012.03.020.
Garg, G., Singh, S., Singh, A.K., Rizvi, S.I., 2016. Antiaging effect of metformin on brain in naturally aged and accelerated senescence model of rat. Rejuvenation Research 20, 173–182. https://doi.org/10.1089/rej.2016.1883.
Glaesmer, H., Riedel-Heller, S., Braehler, E., Spangenberg, L., Luppa, M., 2011. Age- and gender-specific prevalence and risk factors for depressive symptoms in the elderly: a population-based study. International Psychogeriatrics 23, 1294–1300. https://doi. org/10.1017/S1041610211000780.
Godbout, J.P., Chen, J., Abraham, J., Richwine, A.F., Berg, B.M., Kelley, K.W., Johnson, R.W., 2005. Exaggerated neuroinflammation and sickness behavior in aged mice following activation of the peripheral innate immune system. The FASEB Journal 19, 1329–1331. https://doi.org/10.1096/fj.05-3776fje.
Godbout, J.P., Moreau, M., Lestage, J., Chen, J., Sparkman, N.L., O’Connor, J., Castanon, N., Kelley, K.W., Dantzer, R., Johnson, R.W., 2008. Aging exacerbates depressive-like behavior in mice in response to activation of the peripheral innate immune system. Neuropsychopharmacology 33, 2341–2351.
Herrera-Pérez, J.J., Martínez-Mota, L., Fernández-Guasti, A., 2008. Aging increases the susceptibility to develop anhedonia in male rats. Progress in NeuroPsychopharmacology and Biological Psychiatry 32, 1798–1803. https://doi.org/10.1016/j.pnpbp.2008.07.020.
Illouz, T., Madar, R., Louzon, Y., Griffioen, K.J., Okun, E., 2016. Unraveling cognitive traits using the Morris water maze unbiased strategy classification (MUST-C) algorithm. Brain, Behavior, and Immunity 52, 132–144. https://doi.org/10.1016/j.bbi. 2015.10.013.
Jones, K.A., Thomsen, C., 2013. The role of the innate immune system in psychiatric disorders. Mol. Cell. Neurosci. 53, 52–62. https://doi.org/10.1016/j.mcn.2012.10. 002.
Kelley, K.W., O’Connor, J.C., Lawson, M.A., Dantzer, R., Rodriguez-Zas, S.L., McCusker, R.H., 2013. Aging leads to prolonged duration of inflammation-induced depressionlike behavior caused by Bacillus Calmette-Guérin. Brain Behav Immun 32, 63–69. https://doi.org/10.1016/j.bbi.2013.02.003.
Kim, Y.-K., Na, K.-S., Myint, A.-M., Leonard, B.E., 2016. The role of pro-inflammatory cytokines in neuroinflammation, neurogenesis and the neuroendocrine system in major depression. Progress in Neuro-Psychopharmacology and Biological Psychiatry 64, 277–284. https://doi.org/10.1016/j.pnpbp.2015.06.008.
Knutson, S.K., Warholic, N.M., Wigle, T.J., Klaus, C.R., Allain, C.J., Raimondi, A., Porter Scott, M., Chesworth, R., Moyer, M.P., Copeland, R.A., Richon, V.M., Pollock, R.M., Kuntz, K.W., Keilhack, H., 2013. Durable tumor regression in genetically altered malignant rhabdoid tumors by inhibition of methyltransferase EZH2. Proc Natl Acad Sci U S A 110, 7922–7927. https://doi.org/10.1073/pnas.1303800110.
Kumar, A., Stoica, B.A., Sabirzhanov, B., Burns, M.P., Faden, A.I., Loane, D.J., 2013. Traumatic brain injury in aged animals increases lesion size andchronically alters microglial/macrophage classical and alternative activation states. Neurobiology of Aging 34, 1397–1411. https://doi.org/10.1016/j.neurobiolaging.2012.11.013.
Kuzumaki, N., Ikegami, D., Imai, S., Narita, Michiko, Tamura, R., Yajima, M., Suzuki, A., Miyashita, K., Niikura, K., Takeshima, H., Ando, T., Ushijima, T., Suzuki, T., Narita, Minoru, 2010. Enhanced IL-1beta production in response to the activation of hippocampal glial cells impairs neurogenesis in aged mice. Synapse 64, 721–728. https://doi.org/10.1002/syn.20800.
Lund-Sørensen, H., Benros, M.E., Madsen, T., Sørensen, H.J., Eaton, W.W., Postolache, T.T., Nordentoft, M., Erlangsen, A., 2016. A nationwide cohort study of the association between hospitalization with infection and risk of death by suicide. JAMA Psychiatry 73, 912–919. https://doi.org/10.1001/jamapsychiatry.2016.1594.
Ma, K., Zhang, H., Baloch, Z., 2016. Pathogenetic and therapeutic applications of tumor necrosis factor-α (TNF-α) in major depressive disorder: a systematic review. Int J Mol Sci 17. https://doi.org/10.3390/ijms17050733.
Malatynska, E., Steinbusch, H.W.M., Redkozubova, O., Bolkunov, A., Kubatiev, A., Yeritsyan, N.B., Vignisse, J., Bachurin, S., Strekalova, T., 2012. Anhedonic-like traits and lack of affective deficits in 18-month-old C57BL/6 mice: implications for modeling elderly depression. Experimental Gerontology 47, 552–564. https://doi.org/10.1016/j.exger.2012.04.010.
Miller, A.H., Raison, C.L., 2016. The role of inflammation in depression: from evolutionary imperative to modern treatment target. Nature Reviews Immunology 16, 22–34. https://doi.org/10.1038/nri.2015.5.
Morris, R., 1984. Developments of a water-maze procedure for studying spatial learning in the rat. Journal of Neuroscience Methods 11, 47–60. https://doi.org/10.1016/ 0165-0270(84)90007-4.
Naismith, S.L., Norrie, L.M., Mowszowski, L., Hickie, I.B., 2012. The neurobiology of depression in later-life: clinical, neuropsychological, neuroimaging and pathophysiological features. Progress in Neurobiology 98, 99–143. https://doi.org/10.1016/j. pneurobio.2012.05.009.
O’Neil, S.M., Witcher, K.G., McKim, D.B., Godbout, J.P., 2018. Forced turnover of aged microglia induces an intermediate phenotype but does not rebalance CNS environmental cues driving priming to immune challenge. Acta Neuropathologica Communications 6, 129. https://doi.org/10.1186/s40478-018-0636-8.
Prut, L., Belzung, C., 2003. The open field as a paradigm to measure the effects of drugs on anxiety-like behaviors: a review. Eur J Pharmacol 463, 3–33. https://doi.org/10.1016/S0014-2999(03)01272-X.
Qin, H., Holdbrooks, A.T., Liu, Y., Reynolds, S.L., Yanagisawa, L.L., Benveniste, E.N., 2012. SOCS3 deficiency promotes M1 macrophage polarization and inflammation. J Immunol 189, 3439–3448. https://doi.org/10.4049/jimmunol.1201168.
Rutter, M., 2006. Implications of resilience concepts for scientific understanding. Ann N Y Acad Sci 1094, 1–12. https://doi.org/10.1196/annals.1376.002.
Smith, K.J., Au, B., Ollis, L., Schmitz, N., 2018. The association between C-reactive protein, Interleukin-6 and depression among older adults in the community: a systematic review and meta-analysis. Experimental Gerontology 102, 109–132. https:// doi.org/10.1016/j.exger.2017.12.005.
Sukoff Rizzo, S.J., Neal, S.J., Hughes, Z.A., Beyna, M., Rosenzweig-Lipson, S., Moss, S.J., Brandon, N.J., 2012. Evidence for sustained elevation of IL-6 in the CNS as a key contributor of depressive-like phenotypes. Transl Psychiatry 2, e199. https://doi.org/ 10.1038/tp.2012.120.
Tang, Y., Li, T., Li, J., Yang, J., Liu, H., Zhang, X.J., Le, W., 2014. Jmjd3 is essential for the epigenetic modulation of microglia phenotypes in the immune pathogenesis of Parkinson’s disease. Cell Death Differ 21, 369–380. https://doi.org/10.1038/cdd. 2013.159.
von Leden, R.E., Khayrullina, G., Moritz, K.E., Byrnes, K.R., 2017. Age exacerbates microglial activation, oxidative stress, inflammatory and NOX2 gene expression, and delays functional recovery in a middle-aged rodent model of spinal cord injury. J Neuroinflammation 14. https://doi.org/10.1186/s12974-017-0933-3.
Walf, A.A., Frye, C.A., 2007. The use of the elevated plus maze as an assay of anxietyrelated behavior in rodents. Nature protocols 2, 322. https://doi.org/10.1038/nprot. 2007.44.
Wang, R., Wang, W., Xu, J., Liu, D., Jiang, H., Pan, F., 2018. Dynamic effects of early adolescent stress on depressive-like behaviors and expression of cytokines and JMJD3 in the prefrontal cortex and hippocampus of rats. Front Psychiatry 9, 471. https:// doi.org/10.3389/fpsyt.2018.00471.
WHO, 2017. Depression and other common mental disorders: global health estimates.
WHO Document Production Services.
Willner, P., 2017. The chronic mild stress (CMS) model of depression: history, evaluation and usage. Neurobiology of Stress 6, 78–93. https://doi.org/10.1016/j.ynstr.2016.08.002.
Willner, P., Towell, A., Sampson, D., Sophokleous, S., Muscat, R., 1987. Reduction of sucrose preference by chronic unpredictable mild stress, and its restoration by a tricyclic antidepressant. Psychopharmacology 93, 358–364. https://doi.org/10.1007/BF00187257.
Wolf, S.A., Boddeke, H.w.g.m., Kettenmann, H., 2017. Microglia in physiology and disease. Annu. Rev. Physiol. 79, 619–643. https://doi.org/10.1146/annurev-physiol022516-034406.
Wu, L.-H., Lin, C., Lin, H.-Y., Liu, Y.-S., Wu, C.Y.-J., Tsai, C.-F., Chang, P.-C., Yeh, W.-L., Lu, D.-Y., 2016. Naringenin suppresses neuroinflammatory responses through inducing suppressor of cytokine signaling 3 expression. Mol. Neurobiol. 53, 1080–1091. https://doi.org/10.1007/s12035-014-9042-9.
Xu, L., Wang, K., 2016. Screening of mouse strain susceptive to chronic stress depression and preliminary research on its mechanism. Laboratory Animal and Comparative Medicine 36, 257–262. https://doi.org/10.3969/j.issn.1674-5817.2016.04.003.
Yadav, R., Weng, H.-R., 2017. EZH2 regulates spinal neuroinflammation in rats with neuropathic pain. Neuroscience 349, 106–117. https://doi.org/10.1016/j. neuroscience.2017.02.041.
Yan, H.-L., Sun, X.-W., Wang, Z.-M., Liu, P.-P., Mi, T.-W., Liu, C., Wang, Y.-Y., He, X.-C., Du, H.-Z., Liu, C.-M., Teng, Z.-Q., 2019. MiR-137 deficiency causes anxiety-like behaviors in mice. Front. Mol. Neurosci. 12. https://doi.org/10.3389/fnmol.2019. 00260.
Ye, S.-M., Johnson, R.W., 1999. Increased interleukin-6 expression by microglia from brain of aged mice. Journal of Neuroimmunology 93, 139–148. https://doi.org/10.
1016/S0165-5728(98)00217-3.
Yin, Y., Liu, W., Dai, Y., 2015. SOCS3 and its role in associated diseases. Human Immunology 76, 775–780. https://doi.org/10.1016/j.humimm.2015.09.037.
Yirmiya, R., Rimmerman, N., Reshef, R., 2015. Depression as a microglial disease. Trends in Neurosciences 38, 637–658. https://doi.org/10.1016/j.tins.2015.08.001.
Zhang, Q., Sun, H., Zhuang, S., Liu, N., Bao, X., Liu, X., Ren, H., Lv, D., Li, Z., Bai, J., Zhou, X., Tang, L., 2019. Novel pharmacological inhibition of EZH2 attenuates septic shock by altering innate inflammatory responses to sepsis. International Immunopharmacology 76, 105899. https://doi.org/10.1016/j.intimp.2019.105899.
Zhang, X., Wang, Y., Yuan, J., Li, N., Pei, S., Xu, J., Luo, X., Mao, C., Liu, J., Yu, T., Gan, S., Zheng, Q., Liang, Y., Guo, W., Qiu, J., Constantin, G., Jin, J., Qin, J., Xiao, Y., 2018. Macrophage/microglial Ezh2 facilitates autoimmune inflammation through inhibition of Socs3. Journal of Experimental Medicine 215, 1365–1382. https://doi.org/10. 1084/jem.20171417.
Zhang, Y., Tao, G.-J., Hu, L., Qu, J., Han, Y., Zhang, G., Qian, Y., Jiang, C.-Y., Liu, W.-T., 2017. Lidocaine alleviates morphine tolerance via AMPK-SOCS3-dependent neuroinflammation suppression in the spinal cord. J Neuroinflammation 14. https://doi. org/10.1186/s12974-017-0983-6.