Adult male C57BL/6J mice (age, 8-10 weeks; weight, 23-25 g) were provided by Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). The animals were maintained at temperature of 23±11°C with 50±10% relative humidity in specific pathogen-free conditions and had access to food and water ad libitum. All procedures were performed in accordance with the National Institute of Health Guide for the Care and Use of Laboratory Animals (publication no. 80-23) revised 1996, and were approved by the Animal Care and Use Committee for experimental animals of Tongji Medical College (Wuhan, China).

Focal cerebral ischemia was induced by MCAO as previously reported (27). The mice were anesthetized by intraperitoneal injection of ketamine (100 mg/kg) and xylazine (8 mg/kg). Under an operating microscope, the right external carotid artery was incised and a 6-0 silicone-coated nylon monofilament with a rounded tip was inserted into the internal carotid artery until it reached the bifurcation of the MCA for occlusion for 60 min, followed by reperfusion. The sham-operated mice underwent an identical surgical procedure without ischemia. A laser-Doppler probe (Periflux system 5000; Perimed AB, Stockholm, Sweden) was placed on the skull (5 mm lateral and 2 mm posterior to the bregma) to monitor the regional cerebral blood flow (rCBF) during MCAO. Mice with <25% rCBF reduction following occlusion and ~80% rCBF increase upon reperfusion were included in further analyses.

During the MCAO, the rectal temperature was maintained at 37±0.5°C. Immediately following surgery and on a daily basis thereafter, the MCAO mice in all housing groups received a subcutaneous injection of normal saline (NS; 1% v/w).

Each male mouse was randomly allocated to housing groups in standard clear plastic cages (27×17×12.5 cm) either individually (socially isolated, SI) or with an ovariectomized female mouse (pair-housed, PH) immediately following ischemia for 14 continuous days.

Experiment 1: The mice were randomly divided into six groups (Fig. 1A): Group 1) Sham operated SI mice (sham + SI; n=12 for behavioral assessment); group 2) sham operated PH mice (sham + PH; n=12 for behavioral assessment); group 3) post-stroke SI mice [MCAO + SI; n=12 for behavioral assessment; n=6 for reverse transcription-quantitative polymerase chain reaction (RT-qPCR) analysis; n=6 for western blotting (WB); and n=6 for immunofluorescence (IF)]; group 4) post-stroke PH mice (MCAO + PH; n=12 for behavioral assessment; n=6 for RT-qPCR; n=6 for WB; and n=6 for IF); group 5) post-stroke SI mice treated with recombinant mouse (r) TGF-β (MCAO + SI + rTGF-β; n=12 for behavioral assessment and n=6 for IF); and group 6) post-stroke SI mice treated with PBS (MCAO + SI + PBS; n=12 for behavioral assessment and n=6 for IF). Immediately following surgery, a cannula was inserted into the left lateral ventricle (0.1 mm posteriorly, 0.9 mm lateral to the midline and 3.1 mm deep from the bregma) in groups 5 and 6 using a stereotaxic instrument (RWD Life Science Co., Ltd., Shenzhen, China) for intracerebroventricular (i.c.v.) injection (28). rTGF-β (2.5 µg/kg; Sino Biological, Inc., Beijing, China) or sterile PBS was administrated i.c.v. into the left lateral ventricle of the SI mice using a sterile 26-G Hamilton microsyringe (80330; Hamilton Company, Reno, NV, USA) every 24 h for 7 continuous days commencing at 7 days post-ischemia (dpi).

Experiment 2: The mice were randomly divided into five groups (Fig. 1B): Group 1) Post-stroke SI mice treated with hyperforin (MCAO + SI + hyperforin; n=6 for RT-qPCR and n=6 for WB); group 2) post-stroke SI mice treated with NS (MCAO + SI + NS; n=6 for RT-qPCR and n=6 for WB); group 3) post-stroke SI mice treated with NS plus PBS (MCAO + SI + NS + PBS; n=12 for functional assays and n=6 for IF); group 4) post-stroke SI mice treated with hyperforin plus TGF-β-neutralizing antibody (MCAO + SI + hyperforin + anti-TGF-β; n=12 for functional assays and n=6 for IF); and group 5) post-stroke SI mice treated with hyperforin plus PBS (MCAO + SI + hyperforin + PBS; n=12 for functional assays and n=6 for IF). Hyperforin (Cayman Chemical Company, Ann Arbor, MI, USA) was dissolved in ethanol at 20 µg/µl. The stock solution was diluted with NS to 0.5 µg/µl and intranasally administered to alternating nostrils with a 2 min interval between each application, every 24 h for 7 days (starting at 7 dpi). Drops (2 µl) were administered to the surface of each nare, allowing the mice to inhale each drop into the nasal cavity (29). A total of 10 µl treatment was delivered over 5 min. The animals in groups 2 and 3 were treated accordingly using NS. TGF-β-neutralizing antibody (1 µg; Abcam, Cambridge, UK) was dissolved in PBS and i.c.v. injected into the left lateral ventricle 30 min prior to the application of hyperforin.

The mice were injected intraperitoneally twice with S-phase marker BrdU (50 µg/g body weight in NS; Sigma-Aldrich; Merck KGaA, Darmstadt, Germany) with 8 h between injections at 13 dpi. The mice were anesthetized and transcardially perfused 1 day following the final injection to analyze BrdU-labeling of the recently proliferated cells by immunofluorescence staining and observed under a fluorescence microscope (BX51; Olympus Corporation, Tokyo, Japan).

Depression-type behavior in the mice (n = 12/group) was assessed using the FST at 14 dpi. The mice were placed into an opaque cylinder tank (diameter, 24 cm; depth, 53 cm), containing 30 cm³ of 25°C water and the duration of immobility was scored over 6 min. Immobility was assessed during the final 4 min as previously described (10). Quantification of floating, vs. swim time was analyzed using Observer software (version 5; Exeter Software, Setauket, NY, USA).

To examine depressive-type behavior, the SCT was performed at 14 dpi. Two identical 10-ml vials were placed on a custom-made wire cage top. At 3 days prior to MCAO, the individually-housed mice were provided with two 10-ml vials of water for 12 h and then were returned to their original housing condition and provided with their normal drinking water and cage tops. After 12 h, the mice were provided with two different 10-ml vials of 3% sucrose for 12 h. Following the completion of habituation, all mice were returned to their original housing condition and provided with their normal drinking water and cage tops. The mice were deprived of water for 12 h prior to the test and were then given access to 3% sucrose solution over 12 h. The volume of 3% sucrose consumed was recorded.

Anxiety-type behavior in the mice (n = 12/group) was assessed using the OFT at 14 dpi. The mice were placed in an open field apparatus (40×40×37.5 cm) and allowed to explore for 20 min. Locomotor activity was quantified using a photobeam activity system (FlexField; San Diego Instruments, San Diego, CA, USA). The relative level of activity occurring in the periphery, vs. center squares was recorded.

Spatial reference memory was assessed using the continuous variant of the Y-maze spontaneous alternation procedure. The Y-maze task was used with three identical open arms (50×16×32 cm). The mice were placed in the center of the maze and allowed to explore the three maze arms freely for 8 min. The mice made correct alternations when they sequentially visited the three arms without repeated entry into a previously visited arm. The total arm entries during the session were counted. Spontaneous alternation was calculated as follows: Spontaneous alternation (%) = [alternations/(total arm entries - 2)] × 100.

To assess recognition memory, the mice were placed in a chamber (35×25×35 cm) made of gray plastic for 30 min 2 days prior to the NORT. During the training sessions, two identical objects were placed symmetrically in the center of the chamber and the mouse was allowed to explore for 10 min. The time spent exploring each object was recorded. During the retention tests, the mice were placed in the chamber, in which one of the familiar objects was replaced with a novel object, and the mice were allowed to explore freely for 5 min. The mice were considered to be exploring an object when they faced the object at a distance of ≤1 cm. A discrimination index was determined as the ratio of the time spent exploring any one object (training session) or the novel object (retention session) over the total time spent exploring both objects.

To assess the contextual memory impairment, a step-through passive avoidance apparatus containing light (14×10×25 cm) and dark (25×25×25 cm) compartments equipped with an electric grid floor was used. The mice were allowed to adjust to the apparatus the day preceding the trials. For the training trials, a mouse was placed in the light compartment. When the mouse crossed to the dark compartment with all four paws, the door closed and the mouse was punished by intermittent electric shocks (2 sec; intensity, 0.8 mA). Following 30 sec, the mouse was removed from the apparatus. The retention tests were conducted 24 h following the training session (without electric shock) and observed for ≤300 sec.

Brain tissues of the ischemic hemisphere was lysed in RIPA lysis buffer containing protease and phosphatase inhibitors (Nanjing KeyGen Biotech Co., Ltd., Nanjing, China). An equal quantity of protein (30-50 µg/lane) was resolved on a 10% SDS-PAGE gel (Beyotime Institute of Biotechnology, Shanghai, China) and blotted onto a PVDF membrane (Millipore; Merck KGaA) using an electrophoresis apparatus (Bio-Rad Laboratories, Inc., Hercules, CA, USA). The membranes were blocked with 5% non-fat milk/TBST to reduce nonspecific binding and subsequently incubated with specific rabbit anti-TGF-β monoclonal antibody (Abcam; cat. no. ab64715) at 1:1,000 dilution and anti-β-actin (1:2,000; Santa Cruz Biotechnology, Inc., Dallas, TX, USA; cat. no. sc-10731) overnight at 4°C. Following extensive washing with TBST, the membranes were incubated with an appropriate peroxidase-conjugated secondary antibody (1:3,000; Wuhan Sanying Biotechnology, Wuhan, China; cat. no. LP1001b) for 2 h at room temperature. Following three washes with TBST, chemiluminescent signals were visualized using electrochemiluminescence western blotting detection reagents (Millipore; Merck KGaA) and bands were captured using an UVP gel documentation system (UVP, LLC, Phoenix, AZ, USA). The band intensity was quantified using ImageJ software version 1.41 (National Institutes of Health, Bethesda, MD, USA).

The mice were anesthetized and transcardially perfused with NS and 4% paraformaldehyde in PBS (pH 7.4). Following blocking in normal goat serum (Sigma-Aldrich; Merck KGaA) to reduce nonspecific binding, paraffin-embedded brain sections (thickness, 4-µm) were incubated overnight at 4°C with diluted monoclonal primary antibodies (Cell Signaling Technology, Inc., Danvers, MA, USA) as follows: Rabbit Ki67 (1:100; cat. no. 9129), rabbit phosphorylated-histone H3 (PH3; 1:50; cat. no. 9706), mouse BrdU (1:100; cat. no. 5292), rabbit doublecortin (DCX; 1:50; cat. no. 14802) and rabbit NeuN (1:200; cat. no. 12943). Following washing with PBS, the slides were incubated with DyLight 549-conjugated goat anti-rabbit (cat. no. A23320) or DyLight 488-conjugated goat anti-mouse secondary antibodies (cat. no. A23210) at a 1:200 dilution (Abbkine Scientific Co., Ltd., Wuhan, China) at 37°C for 2 h. The antibodies were diluted in 1% bovine serum albumin (Sigma-Aldrich; Merck KGaA) in PBS. Nuclei were stained with DAPI and slides were observed using a fluorescence microscope (BX51; Olympus Corporation). In the hippocampal DG, positive cells (Ki67⁺, PH3⁺ and BrdU⁺/DCX⁺) were counted in the granule cell layer of the dorsal DG of hippocampi (bregma from -2.3 to -4.5 mm).

Total RNA from the post-ischemic hemisphere was isolated using TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA) according to the manufacturer’s protocol. cDNA was obtained using Taqman reverse transcriptase (Applied Biosystems; Thermo Fisher Scientific, Inc.). TGF-β and β-actin cDNA were amplified using Power SYBR-Green (Applied Biosystems; Thermo Fisher Scientific, Inc.). Two-step qPCR was performed (95°C for 15 sec, 60°C for 60 sec for 40 cycles) with specific primers for TGF-β (forward, 5′-GTGTGGAGCAACATGTGGAACTCTA-3′ and reverse, 5′-TTGGTTCAGCCACTGCCGTA-3′) and β-actin (forward, 5′-AAGGCCAACCGTGAAAAGAT-3′ and reverse, 5′-GTGGTACGACCAGAGGCATAC-3′). The relative quantitation value is expressed as 2^−ΔΔCq, where ΔCq is the difference between the mean ΔCq value of duplicate measurements of the sample and β-actin control (30).

Data from the behavioral assessments (FST, SCT and OFT), RT-qPCR analysis and western blotting were analyzed by two-way analysis of variance (ANOVA) with surgery and housing condition as between subject factors. Other data were analyzed using one-way ANOVA to compare several groups, followed by Fisher’s Protected Least Significant Difference test for pairwise comparison of groups. The results are presented as the mean ± standard error of the mean. P<0.05 was considered to indicate a statistically significant difference. Statistical analysis was performed using GraphPad Prism 5 (GraphPad Software, Inc., La Jolla, CA, USA).

Depression-type behavior was evaluated using FST and SCT. The floating time in the FST was significantly increased (P<0.05; Fig. 2A) and consumption of sucrose was significantly decreased (P<0.05; Fig. 2B) in the post-stroke SI mice compared with the sham-operated SI mice at 14 dpi. Furthermore, compared with the sham-operated PH mice at 14 dpi, floating times in the post-stroke PH mice were significantly increased (P<0.05; Fig. 2A) and sucrose consumption was significantly decreased (P<0.05; Fig. 2B). The post-stroke PH mice exhibited a significant reduction in floating times (P<0.05; Fig. 2A) and a significant increase in sucrose consumption (P<0.05; Fig. 2B) compared with the post-stroke SI mice at 14 dpi.

Anxiety-type behavior was measured by the OFT. The post-stroke SI mice spent significantly less time in the center of the open field chamber than the sham-operated SI mice at 14 dpi (P<0.01; Fig. 2C). The percentage of time spent in the center of the open field chamber was further significantly reduced in the post-stroke PH mice compared with the sham-operated PH mice at 14 dpi (P<0.05; Fig. 2C). Post-stroke PH had a significant effect on OFT; the post-stroke PH mice spent a significantly increased duration in the center of the open field chamber compared with the post-stroke SI mice (P<0.05; Fig. 2C). No significant difference was observed in locomotor activity or exploration among the four groups in the OFT (P>0.05; Fig. 2C).

The mRNA and protein expression levels of TGF-β in the ischemic hippocampi were significantly decreased in the post-stroke SI mice compared with those in the post-stroke PH mice at 14 dpi (P<0.05; Fig. 2D and E).

The numbers of general proliferation marker Ki67⁺ and M-phase-specific marker PH3⁺ cells in the ischemic hippocampi were significantly decreased in the post-stroke SI mice compared with those in the post-stroke PH mice at 14 dpi (P<0.01 and P<0.05, respectively; Fig. 3A-D). The administration of rTGF-β to the post-stroke SI mice significantly increased the numbers of Ki67⁺ and PH3⁺ cells in the ischemic hippocampus compared with those in the post-stroke SI mice treated with PBS at 14 dpi (P<0.05; Fig. 3A-D). The post-stroke SI mice treated with rTGF-β exhibited similar numbers of Ki67⁺ and PH3⁺ cells in the ischemic hippocampi as the post-stroke PH mice (P>0.05; Fig. 3A-D).

Cell proliferation was assessed at 13 dpi following treatment of the mice with BrdU to label S-phase cells; on 14 dpi, the dividing cells were analyzed. It was observed that the post-stroke SI mice exhibited a significantly decreased number of BrdU⁺ cells in the ischemic hippocampus compared with the post-stroke PH mice (P<0.01; Fig. 4A and B). The post-stroke SI mice administered with PBS had significantly fewer BrdU⁺ cells in the ischemic hippocampus compared with the post-stroke SI mice administered with rTGF-β (P<0.05), which themselves had a similar number of BrdU⁺ cells in the ischemic hippocampus as the post-stroke PH mice (P>0.05; Fig. 4A and B).

DCX, a neuron-specific microtubule-associated protein, expresses specifically in neuroblasts and functions as a marker for neuronal precursors and neurogenesis. Quantitative determination of BrdU/DCX double-labeled cells was performed to assess adult hippocampal neurogenesis. Significantly lower numbers of BrdU⁺/DCX⁺ cells in the ischemic hippocampus were observed in the post-stroke SI mice compared with the post-stroke PH mice at 14 dpi (P<0.05; Fig. 4A and B). Furthermore, significantly lower numbers of BrdU⁺/DCX⁺ cells were observed in the ischemic hippocampus of the PBS-treated post-stroke SI mice compared with the rTGF-β-treated post-stroke SI mice (P<0.05), which themselves exhibited a similar number of BrdU⁺/DCX⁺ cells in the ischemic hippo-campus as the post-stroke PH mice (P>0.05; Fig. 4A and B).

The percentage of correct alternations in the Y-maze was significantly lower in the post-stroke SI mice compared with the post-stroke PH mice at 14 dpi (P<0.05; Fig. 5A); no significant change in the number of total arm entries was observed (P>0.05; Fig. 5B). The post-stroke SI mice treated with rTGF-β exhibited a significant increase in the percentage of correct alternations compared with the PBS-treated post-stroke SI mice (P<0.05; Fig. 5A); no significant change in the number of total arm entries was observed (P>0.05; Fig. 5B).

In the NORT, the mice spent a comparable duration exploring each object, yielding a discrimination index of ~50% in all groups during the training session (Fig. 5C). During the retention trial performed 1 h following training, the post-stroke PH mice spent a significantly longer duration exploring the novel object compared with the familiar object (P<0.01; Fig. 5D), whereas the post-stroke SI mice explored the objects equally (P>0.05; Fig. 5D). The post-stroke SI mice treated with rTGF-β exhibited a significant preference for the novel object compared with the PBS-treated post-stroke SI mice (P<0.05; Fig. 5D).

In the passive-avoidance task, no significant differences were observed in step-through latency times without shocks among groups (P>0.05; Fig. 5E). Post hoc analysis of the animals assessed following training revealed that the post-stroke PH mice exhibited significant higher step-through latency times 24 h following shocks compared with the post-stroke SI mice at 14 dpi (P<0.05; Fig. 5F). Treatment with rTGF-β significantly restored the reduced step-through latency times in the post-stroke SI mice during the retention test (P<0.05; Fig. 5F).

Intranasal administration of hyperforin to the post-stroke SI mice significantly increased the mRNA and protein levels of TGF-β in the ischemic hippocampi compared with those in the NS-treated post-stroke SI mice at 14 dpi (P<0.05; Fig. 6A and B).

The floating times in the FST were significantly decreased (P<0.01; Fig. 6C) and sucrose consumption was significantly increased (P<0.01; Fig. 6D) in the hyperforin-treated post-stroke SI mice compared with the NS-treated post-stroke SI mice at 14 dpi. Application of a TGF-β-neutralizing antibody significantly reversed the hyperforin-mediated decrease in floating times (P<0.05; Fig. 6C) and increase in sucrose consumption (P<0.05; Fig. 6D) in the post-stroke SI mice.

The percentage of time spent in the center of the open field chamber was significantly higher in the hyperforin-treated post-stroke SI mice compared with that in the NS-treated post-stroke SI mice at 14 dpi (P<0.05; Fig. 6E). TGF-β-neutralizing antibodies reversed the hyperforin-mediated increase in the time spent in the center of the open field chamber in the post-stroke SI mice (P<0.05; Fig. 6E).

Intranasal administration of hyperforin to the post-stroke SI mice led to a significant increase in the number of BrdU⁺ and BrdU⁺/NeuN⁺ cells in ischemic hippocampi compared with that in the NS-treated post-stroke SI mice at 14 dpi (P<0.01 and P<0.05, respectively; Fig. 6F and G). TGF-β-neutralizing antibody treatment reversed the hyperforin-mediated increase in BrdU⁺ and BrdU⁺/NeuN⁺ cells in the ischemic hippocampi of post-stroke SI mice at 14 dpi (P<0.01 and P<0.05, respectively; Fig. 6F and G).

The percentage of correct alternations in the Y-maze was significantly higher in the hyperforin-treated post-stroke SI mice compared with the NS-treated post-stroke SI mice at 14 dpi (P<0.05; Fig. 7A) and no significant change in the number of total arm entries was observed (P>0.05; Fig. 7B). TGF-β-neutralizing antibody treatment significantly reversed the hyperforin-mediated increase in the percentage of correct alternations in the post-stroke SI mice at 14 dpi (P<0.05; Fig. 7A), without significant changes to the number of total arm entries (P>0.05; Fig. 7B).

During the training session in the NORT, mice spent a comparable time exploring the same objects in all groups (P>0.05; Fig. 7C). During the retention trial performed 1 h following training, the hyperforin-treated post-stroke SI mice spent significantly more time exploring the novel object compared with the familiar object (P<0.05; Fig. 7D) and the NS-treated post-stroke SI mice explored the objects equally (P>0.05; Fig. 7D). The hyperforin-treated post-stroke SI mice treated with TGF-β-neutralizing antibody exhibited significantly impaired discrimination between familiar and novel objects at 14 dpi (P<0.05; Fig. 7D).

In the passive-avoidance task, no significant differences in step-through latency times without shock were observed among groups (P>0.05; Fig. 7E). Post hoc analysis following training revealed that hyperforin treatment significantly increased step-through latency times compared with those in the NS-treated post-stroke SI group at 14 dpi (P<0.01; Fig. 7F). TGF-β-neutralizing antibody treatment significantly reversed this hyperforin-mediated increase in step-through latency times (P<0.05; Fig. 7F).