Mouse neuroblastoma N2a cells (China National Infrastructure of Cell Line Resource, Beijing, China) were used and cultured in minimum essential medium (MEM; Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) supplemented with 10% fetal bovine serum (Gibco; Thermo Fisher Scientific, Inc.), 100 U/ml streptomycin and 100 U/ml penicillin (Gibco; Thermo Fisher Scientific, Inc.) in a 37°C humidified incubator with 95% air and 5% CO₂. The cells were resuspended and plated into fresh dishes prior to treatment. For stress hormone administration, following cell attachment, the medium was replaced by fresh MEM with corticosteroids (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany) at different final concentrations and incubated for an additional 24–48 h prior to collection. For HDAC2 silencing, N2a cells were transfected with a lentivirus-HDAC2-short hairpin (sh)RNA plasmid (LV-HDAC2-shRNA; Biowit Technologies, Ltd., Shenzhen, China), which encoded shRNA that targeted the mouse HDAC2 transcript. The HDAC2 expression levels were confirmed by western blot analysis.

MTT (Sigma-Aldrich; Merck KGaA) was used to assess cell viability as described by Wang et al (21). N2a cells were seeded into a 96-well plate at 1×104 cells/ml and 200 µl/well. The medium was refreshed following cell attachment, and the cells were exposed to corticosteroids at various concentrations (1–10 µM). Following a 24 h incubation, 20 µl MTT solution (5 mg/ml) was added to each well for an additional 4 h culture at 37°C. Cells without corticosteroid treatment were treated with culture medium under the same culture conditions and were used as the control group. The media were emptied, and 200 µl dimethylsulfoxide were added to each well to dissolve the crystals. The absorbance was assessed at 550 nm to determine the cell viability; four vice holes were used per dose.

The N2a cells were treated with a 2 µM solution of corticosteroids (cat. no. C2505; Sigma-Aldrich; Merck KGaA) for 48 h, and the morphology was analyzed and compared with the control cells. An inverted phase-contrast microscope (Olympus Corporation, Tokyo, Japan) and Image-Pro Plus software (version, 6.0; Media Cybernetics, Inc., Rockville, MD, USA) were used to evaluate morphological changes. The cell density and length of the neurite-like outgrowths were measured, and the cell size, including the number of primary dendrites, were determined as previously described (22). Then, >100 cells were counted per group at random at the same magnification (x40), and at least 3 cultures were used per group.

Corticosteroid-treated and control cells were fixed with 4% paraformaldehyde for 10 min at room temperature; the frozen brain sections (8 µm) were fixed for 20 min. The slides were subsequently incubated with 0.5% Triton followed by 3% hydrogen peroxide (H₂O₂) for 10 min each. The slides were then incubated with mouse anti-HDAC2 (1:200; cat. no. ab51832) and rabbit anti-PSD95 primary antibodies (1:200; cat. no. ab18258), which were both obtained from Abcam (Cambridge, MA, USA), overnight at 4°C. The following day, the slides were incubated with goat anti-mouse and goat anti-rabbit fluorescein-conjugated IgG secondary antibodies (1:100; cat. nos. ZB-0312 and ZB-0311, respectively; Zhongshan Golden Bridge Biotechnology Co., Ltd., Beijing, China) for 40 min at room temperature. The slides were examined via a laser scanning confocal microscope (Olympus Corporation) and analyzed using ImageJ 3.1 software (National Institutes of Health, Bethesda, MD, USA).

To obtain the total protein, N2a cells maintained in plates or hippocampal tissue were lysed with a mixture of lysis buffer (25 mM HEPES, pH 7.4, 150 mM NaCl, 1 mM EDTA and 0.5% Triton X-100) and protease inhibitors on ice for 30 min. The lysates were clarified via centrifugation at 13,000 × g for 30 min at 4°C. The supernatant was discarded prior to protein analysis and western blotting. A total of 40 µg total proteins from each sample were separated by 10% SDS-PAGE, and transferred to a polyvinylidene difluoride membrane (0.45 µm). The membrane was incubated with 5% fat-free milk to block the blots for 1 h at room temperature prior to incubation with primary antibodies [HDAC2, 1:1,000, cat. no. ab51832; PSD-95, 1:1,000, cat. no. ab18258; acetyl-H3K9, 1:500, cat. no. ab10812; H3, 1:500, cat. no. ab1791; acetyl-H4K8, 1:500, cat. no. ab15823; H4, 1:500, cat. no. ab10158; all obtained from Abcam; Akt, 1:1,000, cat. no. 4691; p-Akt (Ser 473), 1:1,000, cat. no. 4058; PI3K, 1:500, cat. no. 4292; and p-PI3K-p85, 1:500, cat. no. 4228; all obtained from Cell Signaling Technology, Inc., Danvers, MA, USA] overnight at 4°C. The blots were probed with horseradish peroxidase-conjugated GAPDH (1:15,000; KC-5G5; Kangcheng Technology Co., Ltd.), and goat anti-mouse and goat anti-rabbit horseradish peroxidase-conjugated IgG secondary antibodies (1:15,000; cat. nos. ZB-2305 and ZB-2301, respectively; Zhongshan Golden Bridge Biotechnology Co., Ltd.) the following day for 1 h at room temperature and the protein expression levels were visualized via X-ray films and quantified using ImageJ 3.1 software. GAPDH, total H3, total H4, total AKT, and total PI3K were used as loading controls for HDAC2 and PSD-95, acetyl-H3K9, acetyl-H4K8, p-Akt and p-PI3K, respectively.

In total, 20 female C57BL/6J mice (12-months-old; weight, 20–30 g) were used because older mice are more sensitive to stress (23) and mice were divided into 2 groups randomly (n=10 per group). The mice were maintained in a temperature- and humidity-controlled room on a 12 h light-dark cycle and had free access to food and water. The use of animals was approved by the Animal Care and Use Committee of The Institute of Laboratory Animal Science of Peking Union Medical College (Beijing, China; ILAS-GC-2012-004). Chronic restraint stress was performed and modified according to Yang et al (24). Briefly, the stressed mice were restrained in cylinder tubes to limit their autonomous actions for 7 h/day (from 11:00 p.m. to 6:00 a.m.) for 28 consecutive days in a behavior science laboratory. The control mice remained in their home cages without stress exposure.

Fresh serum was obtained immediately following euthanasia at 8:30-9:30 a.m., and corticosteroids were measured with a corticosteroid ELISA kit (ADI-900-097; Enzo Life Sciences, Inc., Farmingdale, NY, USA) according to the instruction manual.

The MWM procedure was similar to previous descriptions (25) and was conducted for 7 days following the completion of the stress protocol. Briefly, the mice were trained twice per day on days 1–6 during the learning period and examined via a probe test on day 7. The platform was hidden starting on day 2 and was absent on day 7. Each swimming cycle was 60 sec. The mice were considered to have located the platform when they remained on it for at least 5 sec. The automatic tracking system Noldus Ethovision XT (Noldus Information Technology BV, Wageningen, Netherlands) was used to monitor and analyze the data.

The data were analyzed using SPSS software (version, 13.0; SPSS, Inc., Chicago, IL, USA). The data are expressed as the means ± standard error of the mean, and P<0.05 was considered to indicate a statistically significant difference. T-tests were used to evaluate the differences between two groups, and one-way analyses of variance (ANOVA) followed by Fisher's least significant difference test were performed to evaluate the differences among multiple groups. Repeated measures and multivariate ANOVAs were used for the MWM.

Corticosteroids were used as a stressor agent to mimic stress conditions in vitro. To determine a suitable corticosteroid concentration, the MTT assay was used to evaluate N2a cell viability. Cells were incubated with corticosteroids at a series of concentrations (1, 2, 5 and 10 µM) for 24 h. The absorbance at 550 nm was measured following incubation, and the optical density (OD) value in the defined conditions reflected the number of viable cells present. The OD value exhibited a decreasing trend when the corticosteroid concentration was >1 µM, which indicated that the cells were injured. When the concentration increased to 5 µM, the cell viability significantly decreased, which indicated substantial cell death. Taken together, the stress hormone corticosteroid exhibited debilitative effects on N2a cells, and at a higher dose, it was associated with cell lethality (P<0.05; Fig. 1). Therefore, 2 µM was chosen as a suitable dose to induce injuries without excessive cell apoptosis in the present study.

In the current study, the authors chose a continuous 48 h treatment of 2 µM corticosteroids for the experimental conditions to mimic chronic stress. These conditions induced significant changes in cell morphology and density (P<0.05; Fig. 2A-C). The cell number significantly decreased in these experimental conditions and resulted in a lower cell density, which reflected an effect of corticosteroid growth inhibition on the N2a cells (P<0.05; Fig. 2B). In addition to the reduced cell density, the N2a cells treated with corticosteroid had shorter neurites, and the longest neurite length of the cells significantly decreased after corticosteroid treatment compared with the control cells (P<0.05; Fig. 2C). There were no significant changes in the average neurite number per cell (P>0.05; Fig. 2D). The decreased cell number and neurite outgrowth resulted in a loosened cellular attachment, which reflected a weakened connection and communication among cells. In addition to cell growth inhibition, cell function was also injured after corticosteroid incubation. Postsynaptic density protein 95 (PSD95), a scaffold protein, is a well-established synapse-associated marker that is crucial for synaptic plasticity (26). PSD95 is located in the postsynaptic density domain, and its expression is important for synaptic function. The PSD95 protein was located at the neurite outgrowth of the N2a cells (Fig. 2G). During the experimental conditions, the PSD95 expression level decreased as detected by western blot analysis (P<0.05; Fig. 2E and F). Furthermore, immunofluorescence staining indicated decreased PSD95 expression following corticosteroid incubation (P<0.05; Fig. 2G and H). The downregulated PSD95 verified the functional impairment of the N2a cells.

The acetylation of histone residues, including H3K9, H3K14, H4K8 and H4K12, enhances gene expression related to learning and memory (27). In previous studies of the authors, decreased acetylation levels of histone H3K9 and H4K8 were identified in a mouse model of AD, which is characterized by cognitive impairment (unpublished data). Glucocorticoids can also cause cognitive impairments in mice via a reduction in histone acetylation (28). In the present study, the acetylation levels of H3K9 were significantly decreased following corticosteroid treatment (P<0.05; Fig. 3E and F), consistent with abnormal epigenetic regulation involved in stress, even though H4K8 acetylation remained stable (P>0.05; Fig. 3G-I). Histone deacetylation leads to more compact chromosomes and reduces gene transcription, which thus reduces proteins crucial for learning and memory. Histone deacetylation is catalyzed by histone deacetylases (HDACs), and the protein levels indicated both dose-dependent and time-dependent augmentation of HDAC2 (P<0.05; Fig. 3A-D). HDAC2 was prominently increased following 48 h of corticosteroid treatment at the same dose of 2 µM and exhibited a persistent increase with an increase in dose, when compared with the control group. The levels of HDAC2 were determined using immunofluorescence staining. HDAC2 was expressed in the nucleus (Fig. 3G). At 48 h of corticosteroid incubation at a dose of 2 µM, there was an augmentation of HDAC2 compared with the control cells (Fig. 4G and H). This corresponding upregulation of HDAC2 and cell injuries suggested that HDAC2 may be a negative mediator in corticosteroid-induced neural damage.

Based on previous results, the authors hypothesized that HDAC2 mediates cell injuries following incubation with corticosteroids. To test this hypothesis, further studies were performed. To knockdown HDAC2, cells were transfected with LV-shRNA-mHdac2. Lentivirus vectors that expressed Hdac2 shRNA were added to the media, with the multiplicity of infection = 4.6IU/cell. Polybrene was used as the transfection enhancer. The vectors also carried the gene encoding green fluorescent protein, and successfully transfected cells were recognized via green fluorescence. Scrambled shRNA was used as a control. Green fluorescence was present after transfection, and the efficiency was >50% (Fig. 4A). Western blot analysis indicated a significant reduction in HDAC2 expression following transfection (P<0.05; Fig. 4C and D). HDAC2 silencing was maintained during corticosteroid treatment, as presented in Fig. 4C and D. The transfected cells did not exhibit an increase in HDAC2 following corticosteroid treatment compared with the control cells.

Cell morphology was assessed again following transfection. Consistent with the authors' hypothesis, knockdown of HDAC2 restored cell density in the stressed cells (Fig. 4B). In addition, the cell neurite outgrowth length was partially restored (Fig. 4B). These morphological improvements in the cells provided an indicator of improved cell viability. In addition, PSD95 expression was partially restored, which was evaluated by western blot analysis (Fig. 4E and F). HDAC2 knockdown was followed by an improvement of cell function, and the cell recovery mechanism was subsequently explored. As previously discussed, the acetylation of histones, such as histone H3K9 acetylation, was reduced. Following corticosteroid treatment, transfected cells exhibited an augmentation of acetylated H3K9 compared with the control cells, as demonstrated by western blot analysis (P<0.05; Fig. 5A and B), which may explain, in part, the restoration of protein PSD95 expression.

Next, the phosphorylation status of the PI3K/AKT signaling cascade was assessed. HDAC2 knockdown improved the protective phosphorylation of this signaling pathway (P<0.05; Fig. 5C-F). The phosphorylation of the PI3K subunit p85 is an activated form that phosphorylates downstream AKT, and the activated PI3K/AKT cascade was beneficial for neurons (29). The phosphorylation levels of p85 and serine 473 of AKT decreased dramatically during corticosteroid incubation; however, this trend could be partially reversed by HDAC2 knockdown.

The chronic restraint procedure is a reliable model of chronic stress and has been widely used (30,31). Following stress, the HPA axis is activated, and the main stress hormone glucocorticoid is overexpressed, which exerts negative effects on cognition. On average, an increased blood corticosteroid level was identified in the stressed mice compared with the control mice (P<0.05; Fig. 6B), which indicates the involvement of stress hormones during stress and is consistent with previous studies (32). The stressed mice simultaneously exhibited lower body weights (P<0.05; Fig. 6A), which indicates disturbed living states. The MWM test is a classical behavior detection method that was used to evaluate cognitive performance and identify cognitive impairment in the chronic stressed mice (Fig. 6C and D). In the learning period, there was a significant difference in the escape latency between groups starting on day 3 (P<0.05; Fig. 6C). Specifically, the stressed mice took a longer time to find the platform compared with the wild-type control mice starting on day 3 in the learning period. The weakened learning abilities of the stressed mice lasted for four days to the end of this test period. During the exploratory period, the frequency of crossing the platform zone in the stressed group was also significantly less than the control group (P<0.05; Fig. 6D), and the stressed mice exhibited a deteriorated ability to remember the exact platform position. These results confirmed that chronic stress could cause cognitive decline via glucocorticoid upregulation, which is consistent with previous reports (33,34).

Following the MWM trials, both the stressed and control mice were euthanized. Proteins from the hippocampus and cortex related to cognition were independently extracted. Following the identification of cognitive decline during stress, HDAC2 was assessed via western blotting for its specific role in blocking cognition. Interestingly, there was a corresponding change in HDAC2 during chronic stress. The hippocampal HDAC2 protein levels were significantly increased compared with the wild-type control mice (P<0.05; Fig. 6E and F); however, cortical HDAC2 was unchanged (P>0.05, Fig. 6G and H). The hippocampus is the predominate region related to cognition, and the upregulation of hippocampal HDAC2 following stress confirmed that HDAC2 is involved in chronic stress-induced cognitive impairment in vivo.