A total of 60 adult male Sprague-Dawley (SD) rats (weight range, 220-270 g, 7 weeks old) were provided by Beijing Vital River Laboratory (Beijing, China). Animal treatments and experiments were carried out in accordance with the ‘Guide for the Care and Use of Laboratory Animals' of the National Institutes of Health (14). The present research protocol was approved the Institutional Animal Care and Use Committee of Jining No. 1 People's Hospital (Jining, China). Each rat was maintained in an individual cage with free access to water and food and allowed 1 week of adaptive feeding at 24˚C and relative humidity of 55-65% on a normal light cycle (12/12-h, lights on from 7:00 a.m. to 7:00 p.m.) prior to experimental surgery to minimize animal suffering.

A controlled cortical impact model of TBI was induced in vivo in rats as described previously (15). Animals were anesthetized by isoflurane inhalation (4% for induction and 1.8% for maintenance) mixed with pure oxygen at a flow rate of 500 ml/min (cat. no. R630; Shenzhen Ruiwode Life Technology Co., Ltd.); animals were maintained at 37±0.5˚C throughout the operation using thermal mats. Rats were placed on a stereotaxic frame and secured with an incisor bar and two ear bars. Subsequently, a craniotomy (diameter, 5 mm) was performed on the right parietal cortex (3 mm lateral and posterior from Bregma) with a dental drill. Disruption of the dura and the associated vasculature was avoided as much as possible during the process of craniotomy. A PCI 3000 PinPoint Precision Cortical Impactor (Hatteras Instruments, Inc.), equipped with an impactor tip (diameter, 4 mm), was used to deliver an impact at a velocity of 4.5 m/sec. the dwell time was 75 msec and the deformation was 1 mm. Following the induction of TBI, the piece of resected skull was immediately replaced and secured with bone wax. The incision was then closed using interrupted 4-0 silk stitches. An identical procedure with the exception of impact delivery was performed on sham rats.

Rats were randomized into three groups: i) Sham group (n=18); ii) vehicle-treated TBI group (n=21); and iii) HS-treated TBI group (n=21). For HS treatment, animals were given an intravenous injection of 10% HS (10 g NaCl in 100 ml deionized water) via the tail vein (0.08 ml/g; infusion rate, 0.3 ml/h) 3 h after the induction of TBI (16,17). For vehicle-treated rats, an equivalent amount of normal saline (0.9 g NaCl in 100 ml deionized water) was infused intravenously. The rats were sacrificed 2 days after treatment for further analyses.

Sequential MRI was conducted in 21 rats (including 5, 8 and 8 from the sham, HS-treated and vehicle-treated groups, respectively) at 0 and 48 h after TBI using a 7.0 T scanner (BioSpec Products, Inc.) equipped with a four-channel phased array rat head coil.

T2-weighted turbo spin-echo images (T2WI) were obtained using the fast spin-echo sequence (under conditions of field-of-view, 35x35 mm; matrix, 256x256; number of scans, 20; slice thickness, 1 mm; echo time, 33 msec; repetition time, 2838.2 msec; number of averages, 1). The obtained images were analyzed using the OsiriX software (version 4.2.2; http://www.osirix-viewer.com). To quantify the size of the lesion on T2WI, both the brain and lesion areas were delineated on a single slice by two researchers who reached a consensus, where the region of interest defined by the researchers on each slice containing the lesion was used. The brain and lesion volumes were summed for all animals.

Over the course of TBI, the contralateral brain area was determined according to the aforementioned method and was also delineated in the obtained images in accordance with rescaled drawings from the Paxinos and Watson Atlas (18). Hyperintense pixels in the ipsilateral cortex on the T2WI were regarded as significantly higher signals (P<0.05) compared with those in the contralateral hemisphere. Additionally, lesion volume on the T2WI (Vu) was calculated by multiplying the slice thickness by the hyperintense area on each slice. To compensate for the effect of brain swelling, the adjusted brain swelling (swelling %) and lesion volume (Ve %), which were presented in the form of volume increases in the affected hemisphere, were computed according to the equation below (19):

where Vu and Ve represent the uncorrected and corrected lesion volumes, respectively and Vc and Vi represent the contralateral and ischemic hemisphere volumes, respectively.

The modified neurological severity score (mNSS) of each group was evaluated at 48 h following TBI in a blinded manner according to a previously described method (20). Neurological function scores ranged from 0-18, with a higher score indicating a more severe injury (20).

Coronal sections of the cerebral cortical tissues around the contusion site were selected for H&E and IHC staining. Rats were sacrificed and perfused with PBS via the left ventricle followed by 4% paraformaldehyde for fixation. Then, the collected cerebral tissues were fixed at room temperature in buffered paraformaldehyde (4%) for 24 h embedded in paraffin and sectioned into 4-µm slices. H&E staining was performed after deparaffinization by xylene followed by a descending ethanol gradient at room temperature, where the stained tissues were examined under a light microscope (magnification, x400).

For IHC staining, rats were sacrificed and perfused with PBS via the left ventricle followed by 4% paraformaldehyde for fixation. Cortical tissues around the contusion site were collected and fixed at room temperature in 4% buffered paraformaldehyde for 24 h embedded in paraffin and sectioned into 4-µm slices, which were then deparaffinized using xylene and rehydrated by descending ethanol gradient. Hydrogen peroxide (3%) was diluted with deionized water and used to block endogenous peroxidase activity for 10 min at room temperature. Slices were incubated in citrate buffer (at pH=6.0) for 10 min at 98˚C for antigen retrieval and incubated with 10% goat serum (cat. no. ZLI-9022; Origene Technologies, Inc.) blocking solution for 20 min at 37˚C. Subsequently, the slices were incubated with the following primary antibodies: Rabbit anti-rat glial fibrillary acidic protein (GFAP; 1:2,000; cat. no. Z0334; Dako; Agilent Technologies, Inc.), chicken anti-rat albumin (1:1,000; cat. no. ab106582; Abcam) and rabbit anti-AQP4 (1:1,000; cat. no. ab128906; Abcam), overnight at 4˚C. The slices were then washed twice with PBS and then incubated at 37˚C with the relevant polymeric horseradish peroxidase (HRP)-labeled anti-rabbit immunoglobulin G secondary antibody (Super Vision IHC kit; cat. no. SV0002; Boster Biological Technology) for 30 min at 37˚C. After staining with 3,3'-diaminobenzidine for ~3 min at room temperature, all slices were counterstained with hematoxylin for 2 min at 37˚C, dehydrated and mounted on coverslips for light microscopic examination. IHC staining for albumin, GFAP and AQP4 was examined by ImageJ software 1.8.0 (National Institutes of Health). In total, six visual fields from each slice (magnification, x400) were examined and the average optical density (OD) was determined for all images.

Brain edema was evaluated by calculating the BWC of all animals according to a previously described method (21). The dry and wet weights of the brain were determined, where the water content was calculated for all samples based on the following equation: 100% x [(wet weight - dry weight)/wet weight].

Before the experiment ended, a 1-ml blood sample was collected from the rat tail vein under anesthesia as described previously (22). Blood sodium concentration was analyzed immediately using an ABL80 type automatic blood gas system (Radiometer Medical Ltd.) by another technician blinded to the experimental groups.

The plasma level of IL-1β was determined using ELISA. Blood samples (1 ml) were collected from the rat abdominal aorta under anesthesia as aforementioned and then subjected to 15 min centrifugation at 800 x g at 4˚C. The plasma IL-1β concentration was measured using a commercial ELISA kit (cat. no. F15810, Shanghai Xitang Biological Technology Co., Ltd.) in accordance with the manufacturer's protocol. Then, the OD at 450 nm was determined using a microplate reader.

Total RNA was isolated from cells and brain tissue from the area surrounding the edema using RNAiso plus (cat. no. 9109; Takara Biotechnology Co., Ltd.). RNA quantification was carried out using a spectrophotometer (NanoDrop 2000; PEQLAB Biotechnologie GmbH), and cDNA was synthesized from 1.0 µg RNA by reverse transcriptase M-MLV (cat. no. 2641; Takara Biotechnology Co., Ltd.) according to manufacturer's protocol. The temperature and duration for reverse transcription was as follows: 42˚C for 10 min and 95˚C for 2 min. The primer sequences used for RT-PCR were as follows: NF-κB forward, 5'-ACAGCCTGGTAGTGCGGTCGT-3' and reverse, 5'-TCAGCAAGTGGCTAGTCTGT-3'; IL-1β forward, 5'-AAAAGCTTGGTGATGTCTGG-3' and reverse, 5'-TTT CAACACGCAGGACAGG-3'; AQP4 forward, 5'-CTTTCT GGAAGGCAGTCTCAG-3' and reverse, 5'-CCACACCGA GCAAAACAAAGAT-3'; and GADPH (the reference gene), forward, 5'-CGGATTTGGTCGTATTGGG-3' and reverse, 5'-CTGGAAGATGGTGATGGGATT-3'. qPCR was performed with SYBR^® Premix Ex™ Taq (Takara Biotechnology Co., Ltd) using an ABI 7500 RT PCR System (Applied Biosystems; Thermo Fisher Scientific, Inc.). The thermocycling conditions were as follows: 95˚C for 10 sec, followed by 40 cycles of 95˚C for 5 sec and 60˚C for 30 sec. A dissociation curve was used to examine the detected signal specificity, which constituted a single peak. The 2^-ΔΔCq method was used for quantification of expression (23).

The perilesional brain tissue was isolated from each treatment group 2 days following TBI, which was subjected to homogenization with cytoplasmic buffer supplemented with KCl (10 mM), HEPES (10 mM, at pH=7.9), EGTA (0.1 mM), EDTA (0.1 mM), Nonidet P-40 (0.15%), DTT (1 mM), NaF (10 mM), β-glycerophosphate (50 mM), and Na₃VO₄ (5 mM) and phosphatase inhibitors (Roche Diagnostics). The homogenates were subjected to 30 min centrifugation at 12,000 x g at 4˚C, following which the supernatant was discarded and the pellet was not disturbed. To extract nuclear protein, a nuclear buffer was supplemented with NaCl (400 mM), HEPES (20 mM, at pH=7.9), EGTA (1 mM), EDTA (1 mM), Nonidet P-40 (0.50%), DTT (1 mM), NaF (10 mM), Na₃VO₄ (5 mM), β-glycerophosphate (50 mM), in addition to the protease inhibitor cocktail and the buffer was added to the aforementioned pellet. Subsequently, the homogenates were subjected to a further 15 min incubation on ice after vortexing at room temperature for 20 sec, which was repeated four times. The homogenates were then subjected to 15 min of centrifugation at 12,000 x g at 4˚C to assess the nuclear NF-κB.

AQP4 expression was measured using total protein extracts. Protein was collected from brain tissues around the injury site using a total protein extraction kit (Bei Jing Pu Li Lai Gene Technology Co., Ltd.) in accordance with the manufacturer's protocol. Then, the resultant homogenates were subjected to 30 min of centrifugation at 12,000 x g at 4˚C and the supernatants were collected to analyze the amount of total protein extracted using a bicinchoninic acid protein assay kit (cat. no. P0012S; Beyotime Institute of Biotechnology). Protein lysates (20 µg) were subjected to 15 min of denaturation at 90˚C and 10% SDS-PAGE was performed before the proteins were transferred onto PVDF membranes (EMD Millipore). Subsequently, the membranes were subjected to 2 h blocking using 5% non-fat milk at 37˚C and incubated overnight at 4˚C with primary antibodies, including anti-NF-κB (1:1,000; cat. no. 8242; Cell Signaling Technology, Inc.), anti-AQP4 (1:800; cat. no. ab46182; Abcam), anti-zonula occludens-1 (ZO-1; 1:400; cat. no. 61-7300; Invitrogen; Thermo Fisher Scientific, Inc.), anti-claudin-5 (1:1,000; cat. no. 35-2500; Invitrogen; Thermo Fisher Scientific, Inc.), anti-occludin (1:400; cat. no. 33-1500; Invitrogen; Thermo Fisher Scientific, Inc.), anti-GAPDH (1:1,000; cat. no. AF1186; Beyotime Institute of Biotechnology) and anti-Histone H3 (1:1,000; cat. no. AH433; Beyotime Institute of Biotechnology). Then, the membranes were rinsed with PBST containing 0.1% Tween-20 for 10 min and repeated five times. Subsequently, the PVDF membranes were incubated with a HRP-labeled goat anti-rabbit IgG (1:5,000; cat. no. abs20002; Absin Biotechnology Co., Ltd.) or HRP-labeled goat anti-mouse IgG secondary antibody (1:5,000; cat. no. abs20001; Absin) at 25˚C for 1 h. The bands were visualized using the ECL (cat. no. P0018; Beyotime Institute of Biotechnology) method and analyzed with the ImageJ software 1.8.0 (National Institutes of Health). GAPDH was used to normalize the expression of the proteins, whilst histone H3 was used to normalize the expression of NF-κB in the nuclear extracts.

Each value was measured three times and data are presented as the mean ± SD. SPSS 18.0 (IBM Corp.) was used for statistical analyses. No repeated-measure (matched) values were available in the current study, and all of the values were examined by one-way ANOVA and Tukey's post hoc test. P<0.05 was considered to indicate a statistically significant difference.

MRI was first performed to determine the efficacy of HS treatment on lesions induced by TBI, as observed on the T2WI (Fig. 1A). On day 3 following TBI, the corrected T2WI Ve were found to be 18.25±4.13 and 25.59±6.13% in the HS and vehicle groups, respectively (P<0.05; Fig. 1B and C). In addition, HS treatment significantly alleviated brain swelling compared with that in the vehicle group, where the HS and vehicle groups exhibited values of 13.73±2.98 and 18.63±3.87%, respectively (P<0.05; Fig. 1D).

Compared with the sham group, the mNSS scores of the TBI group were found to be significantly increased at 48 h post-TBI. HS treatment significantly reduced the mNSS scores compared with those of TBI alone, suggesting that HS significantly reduced neurological deficits (P<0.05; Fig. 2A). Subsequently, TBI-induced brain edema was evaluated by measuring the BWC, where it was demonstrated that the BWC was significantly increased in the ipsilateral cortex in the TBI group compare with that in the sham group, which was significantly reversed by HS treatment (Fig. 2B).

H&E staining results identified a regular arrangement of neurons in the sham group, along with capillary morphogenesis and normal glial cells. However, 2 days following TBI, the majority of cells were disorderly arranged where the nuclei became pyknotic or were severely shrunken. In addition, cellular swelling and hypervacuolization were observed in certain areas. HS treatment significantly improved TBI induced nuclei shrinking and cell swelling (Fig. 2C).

BBB leakage was next assessed by IHC staining for albumin and western blotting for claudin-5, ZO-1 and occludin. It was found that albumin staining was largely negative in the sham group, whilst a significant increase in albumin expression was observed in tissues from the TBI group (Fig. 3A and B), indicating BBB injury. Albumin expression was found to be significantly downregulated in tissues in the HS group compared with that in the TBI group alone.

In addition, tight junction (TJ)-associated proteins, including occludin, ZO-1 and claudin-5, were also revealed to be significantly downregulated by TBI, which was significantly reversed by HS treatment (Fig. 3D). Therefore, these results suggested that HS treatment alleviated BBB destruction following TBI.

Astrocyte activation was evaluated by IHC staining for GFAP (Fig. 4). GFAP expression was detected in the sham group, which identified few astrocytes with finely branched protrusions. In the TBI group, GFAP expression was demonstrated to be elevated with astrocyte hyperplasia, cell body enlargement, protrusions, thickening and extravascular collagen fibers all observed, suggesting the activation of astrocytes. However, after HS treatment, GFAP expression remained to be elevated compared with that in the sham group, but was significantly lower compared with that in the TBI group, suggested that the astrocyte protrusions did not completely recover their fine morphology. Therefore, it was speculated that HS mitigated astrocyte activation around the site of TBI.

AQP4 expression tissues around the contusion site was evaluated using IHC (Fig. 5). A small amount of AQP4 expression was observed in the sham group, which was significantly upregulated in the TBI group, where AQP4 staining was clearly observed around the blood vessels. After HS treatment, AQP4 expression was significantly downregulated compared with that in the TBI group, where the positively stained particles were substantially smaller.

AQP 4 expression in peri-TBI tissues was also measured using western blotting and RT-qPCR. Compared with the sham group, the protein and mRNA expression levels of AQP4 in the TBI group were found to be markedly increased, which was reversed by HS treatment.

The blood sodium concentration was analyzed using an automatic blood gas system, where the results showed that HS treatment remarkably increased the blood sodium ion concentration from 144.50±0.94 mmol/l (sham group) to 158.11±2.2 mmol/l (P<0.05; Table I).

The expression levels of inflammatory factors in each group were next measured (Fig. 6). Western blotting and RT-qPCR results suggested that, compared with the sham group, the protein and mRNA expression levels of NF-κB in the TBI group were significantly higher whilst HS treatment significantly reduced NF-κB expression compared with those in the TBI group.

Serum IL-1β levels were next measured using ELISA. Compared with the sham group, the serum IL-1β level in the TBI group was found to be significantly higher, which was significantly reversed by HS treatment. RT-qPCR was also performed in the tissue samples to detect the changes in IL-1β mRNA expression. Compared with the sham group, IL-1β mRNA expression in the TBI group was significantly higher, but HS treatment significantly downregulated IL-1β expression compared with that in the TBI group.