Tetramethylpyrazine-2’-O-sodium ferulate attenuates blood–brain barrier disruption and brain oedema after cerebral ischemia/reperfusion

Abstract

Disruption of blood–brain barrier (BBB) and subsequent oedema are major causes of the pathogenesis in ischaemic stroke with which the current clinical therapy remains unsatisfied. In this study, we examined the therapeutic effect of tetramethylpyrazine-2′-O-sodium ferulate (TSF)-a novel analogue of tetramethylpyrazine in alleviating BBB breakdown and brain oedema after cerebral ischaemia/reperfusion (I/R). Then, we explored the potential mechanism of the protection on BBB disruption in cerebral I/R rat models. Male Sprague-Dawley rats (250–300 g) were subjected to 120 min middle cerebral artery occlusion (MCAO), followed by 48 h reperfusion. TSF (10.8, 18 and 30 mg kg⁻¹) and ozagrel (18 mg kg⁻¹) were administrated by intravenous injection immediately for the first time and then received the same dose every 24 h for 2 days. We found that TSF treatment significantly attenuated the cerebral water content, infarction volume and improved neurological outcomes in MCAO rats compared to I/R models. Moreover, we investigated the effect of TSF on the BBB for that cerebral oedema is closely related to the permeability of the BBB. We found that the permeability of BBB was improved significantly in TSF groups compared to I/R model group by Evans blue leakage testing. Furthermore, the expressions of tight junction (TJ) proteins junction adhesion molecule-1 and occludin significantly decreased, but the protein expression of matrix metalloproteinase-9 (MMP-9) and aquaporin 4 (AQP4) increased after cerebral I/R, all of which were alleviated by TSF treatment. In conclusion, TSF significantly reduced BBB permeability and brain oedema, which were correlated with regulating the expression of TJ proteins, MMP-9 and AQP4. These findings provide a novel approach to the treatment of ischaemic stroke.

Introduction

Ischaemic stroke and cerebral ischaemia/reperfusion (I/R) result in the blood–brain barrier (BBB) disruption, leading to the development of brain oedema. ^1,2 Despite the advances in the overall management of acute stroke, the propensity of ischaemic brain tissue to develop oedema remains one of the leading causes of death and severe disability in the world within the first few days of stroke. ^3,4 It is not surprising that strategies by which in protecting the BBB can be seen as an aspect of restoring normal brain homeostasis and this will be a promising management strategy for treatment of BBB dysfunction and cerebral oedema. ⁵

The BBB is a gatekeeper and modulator for the central nervous system, which is disrupted during ischaemic stroke and subsequent reperfusion. ⁶ Permeability contributes to vasogenic oedema, which further accelerates brain damage, activating matrix metalloproteinase-9 (MMP-9) and aquaporin 4 (AQP4), changing tight junction (TJ) proteins and eventually result in BBB breakdown. ⁷ However, cerebral oedema is an uncommon but potentially lethal complication post intravenous thrombolysis in addition to the risk of cerebral hemorrhage. ⁸ The presence of the massive cerebral oedema is usually associated with clinical worsening. It is postulated that the cerebral oedema after postrecanalization is composed of cytotoxic oedema caused by cerebral ischaemia and vasogenic oedema aggravated by reperfusion. ⁸

Tetramethylpyrazine is a kind of active ingredients, extracted from Rhizoma Chuanxiong, a Chinese herbal medicine (Figure 1(a)). It can protect the cerebral I/R injury. ⁹ Ferulic acid, a phenolic chemical composition, found in plant cell walls, also plays a vital role in cerebral I/R injury on the aspects of radical scavenging, anti-oxidation and anti-inflammatory. ¹⁰ Recently, sodium ozagrel, as a selective TXA2 synthase inhibitor, is widely used in clinical and it ameliorates platelet aggregation, vasoconstriction and brain oedema in acute cerebral ischaemia ¹¹ (Figure 1(b)). Tetramethylpyrazine-2′-O-sodium ferulate (TSF), a novel analogue of tetramethylpyrazine, exhibited anti-ischaemia and BBB protection properties in the previous research ¹² (Figure 1(c)). TSF, a new compound is modified Ligustrazine (2,3,5,6-tetramethyl pyrazine) as the parent compound on the molecular structure. ¹³ It has been reported that TSF showed significant effect against cerebral ischaemia injury in vitro. ¹² Phase I clinical trials have produced remarkable effect and phase II clinical trials are now under way.

Figure 1.

Chemical structure of tetramethylpyrazine (a), ozagrel sodium (b) and TSF (c). TSF: tetramethylpyrazine-2′-O-sodium ferulate.

Our previous studies demonstrated that the administration of TSF could inhibit cerebral microcirculatory disturbance, neuron damage and cognitive impairment elicited by ischaemia. ^14,15 However, the effect of TSF on brain oedema and BBB breakdown after cerebral I/R injury have not been well explored. Therefore, we hypothesized that the decrease in the production of MMP-9 and AQP4 may be a possible mechanism, and TSF reverses the reduction of junction adhesion molecule-1 (JAM-1) and occludin expression, which eventually protects BBB from I/R injury. To provide evidence, the current study was designed to investigate the neurological deficits, infarction volume, brain oedema and BBB dysfunction after I/R, particularly illuminate the involvement of MMP-9, AQP4 and TJ proteins JAM-1.

Experimental procedure

Materials and animals

TSF was purchased from Hefei Medical Pharmaceutical Co. Ltd (Hefei, China). Adult male Sprague-Dawley (SD) rats weighing 250–300 g were purchased from the Animal Centre of Jiangsu University Health Science Centre (Certificate no. SCXK (SU) 2013-0011, Nanjing, China). The animals were acclimatized for 1 week in cages at (22 ± 2)°C and humidity of (40 ± 5)% under a 12-h light/dark cycle and received standard diet and water ad libitum. Before experiment, rats were fasted for 12 h but allowed free access to water. The experimental protocol and procedures were carried out in accordance with the Institutional Animal Care and conformed to international guidelines for the ethical use of experimental animals.

Animal experimental groups and drug administration

In the experiment, rats were randomly divided into six groups: Sham group, I/R model group, positive control group (ozagrel, 18 mg kg⁻¹), TSF treatment group (10.8 mg kg⁻¹), TSF treatment group (18 mg kg⁻¹) and TSF treatment group (30 mg kg⁻¹). In the positive control group and TSF treatment groups, animals were administrated with ozagrel and TSF by intravenous injection immediately for the first time and then received the same dose every 24 h for 2 days. Rats received a code that did not reveal the allocated treatment for the purpose of performing all further studies in a masked manner as experimenters analysing the samples were not aware of the treatment assigned to each rat.

Middle cerebral artery occlusion and reperfusion animal model

Male SD rats were subjected to middle cerebral artery occlusion (MCAO) and reperfusion as described previously. ¹⁶ Briefly, after anaesthetized with chloral hydrate (50 mg kg⁻¹) intraperitoneally, a rat was placed in the supine position with limbs taped to the operation table; a 4-0 silicon-coated nylon suture (diameter 0.26 mm) was inserted and passed through the left external carotid artery and then gently advanced into the internal carotid artery after the left external carotid artery was isolated, and its branches were ligated. Reperfusion was achieved by slowly retracting the suture back after 120 min occlusion. Rectal temperature was controlled at (37 ± 0.5)°C throughout the procedure from the start of the surgery until the recovery of the animals from anaesthesia with a thermostat-controlled heating pad.

Evaluation of vascular permeability

Evans blue (EB) extravasation was conducted to evaluate BBB opening as previously described. ¹⁷ EB dye (4% in 0.9% saline, 4 ml kg⁻¹) was administered intravenously through the vena caudalis. Three hours later, rats were anaesthetized with chloralic hydras, the rats’ chests were opened and the animals were perfused with physiological saline through the left ventricles at a pressure of 110 mmHg until colourless perfusion fluids were obtained from the right atriums. Brains were weighed and placed in formylamine solution at 56°C for 24 h. Following homogenization and centrifugation, the extracted dye was determined at 632 nm with a spectrometer. The tissue EB leakage rates were calculated as (left hemisphere integrated density (OD)/left hemisphere weight)/(right hemisphere OD/right hemisphere weight) × 100%.

Determination of brain water content

Brain water content was determined by the standard wet-dry method as described to evaluate brain oedema. ¹⁸ Brain water content was measured 48 h after MCAO. Briefly, at the end of the experiments, the brains were immediately removed and separated into ipsilateral ischaemic hemispheres and contralateral non-ischaemic hemispheres. The bilateral hemispheres were weighed to obtain the wet weight (without the brain stem and cerebellum). Brain specimens were then dried in an oven at 120°C for 24 h and weighed again to obtain the dry weight. The brain water content of each hemisphere was calculated with the equation as follows: Brain water content (%) = (wet weight − dry weight)/wet weight × 100%.

Determination of infarct size

Infarct volumes were measured at 48 h after MCAO using 2,3,5-triphenyltetrazolium chloride (TTC) staining, as previously described. ^19,20 TTC staining (Sigma, St Louis, Missouri, USA) was used to reveal cerebral infarct, whereby the infarct tissue remained unstained (white coloured), while normal brain tissues were stained to red. The rats were killed under deep anaesthesia and the brains were rapidly removed. In brief, brains were made coronally into serial 2-mm thick slices (from rostral to caudal). The set of slices from each brain was incubated for 30 min at 37°C in 1% TTC in 0.1 mol L⁻¹ phosphate buffer. Each TTC-stained brain portion was then photographed with a digital camera (iPhone A-1431, Campbell, California, USA). The unstained areas of the fixed brain slices were defined as infarcted. In order to avoid the influence of brain oedema, the infarct volumes were calculated by the equation as [(contralateral hemisphere volume-non-infarcted volume of the ipsilateral hemisphere)/contralateral hemisphere volume]. ²¹

Evaluation of neurological deficits

Neurological deficit scores were evaluated following a modified scoring system which developed according to a graded scoring system: 0 = no apparent neurological deficit; 1 = failure to extend contralateral forelimb fully; 2 = circling to the contralateral side; 3 = mild circling to contralateral side; 4 = no spontaneous motor activity. Each animal’s scores were estimated within approximately 1 min, and estimation was repeated another three times for consistency. Score of 0 corresponds to a normal neurological status, and the higher the neurological deficit score, the more severe behavioural deficit. ^16,22

Immunohistochemical staining

The rats (n = 6 for each group) were used for immunohistochemistry (IHC) assay for AQP4. Each animal was perfused transcardially with 150 ml 0.9% saline followed by 500 ml ice cold 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) at 24 h after reperfusion. Brain sections (5 μm thick) were blocked in 3% hydrogen peroxide, 3% normal goat serum, and then slides were incubated overnight with polyclonal goatanti-AQP4 (1:100 dilution; Abcam, UK) in 0.01 mol L⁻¹ phosphate-buffered saline. IHC was performed via the avidin–biotin technique, and then hematoxylin staining was selected as counterstaining, the Vect ABC kit containing the secondary antibodies, secondary biotinylated conjugates and diaminobezidine was purchased from Zhongshan Biology Technology Company, China. An examiner blinded to the experimental groups counted the cells labelled withAQP4 under a 400 light microscope. For each sample, three consecutive slices were observed. Firstly, the positive cells of each slice were counted in five random visual fields of ischaemic cortex around the infarct core. The average positive cell amount of each slice was obtained by divided the sum of numbers counted from each field by five, and the average positive cell amount of each sample was obtained by calculating the average number of three slices. At last, the average of each sample was used for statistical analysis.

Western blot analysis

The total proteins of brains were extracted by ice-cold cell lysis buffer (Beyotime, Shanghai, China). Bicinchoninic acid assay protein assay kit (Beyotime) was used to determine the total protein content. Equal amount of protein preparations were run on 8%–12% sodium dodecyl acrylamide-polyacrylamide gels, electro-transferred topolyvinylidine difluoride membranes, blotted overnight with primary antibodies and reacted with appropriated peroxidase-labelled secondary antibodies (Zhongshan Biotechnology Co., Beijing, China). Immunoreactive proteins were detected by the chemiluminescence assay (ECL Millipore, Billerica, Massachusetts, USA) using LAS-4000 chemiluminescence imaging system (Fujifilm, Tokyo, Japan). Quantitative analysis was performed using AlphaEasaFC analysis software (Alpha Innotech, San Leandro, California, USA) with horseradish peroxidase-conjugated monoclonal antibody against β-actin serving as a control. ²³ The primary antibodies included occludin (1:1000; Epitomics, Burlingame, California, USA), JAM-1 (1:1000; Cell Signalling, Danvers, Massachusetts, USA), AQP4 (1:1000; Cell Signalling), MMP-9 (1:1000; Santa Cruz Biotechnology, Santa Cruz, California, USA) and β-actin (1:5000; Anbo Biotechnology Co., Ltd, San Francisco, California, USA).

Data analysis

All data are expressed as means ± SD and performed with SPSS 19.0 software (SPSS Inc., Chicago, Illinois, USA). One-way analysis of variance was conducted to compare and identify differences among all groups. The post hoc analysis among the data with equal variances was made by Bonferroni’s method, while Dunnett’s T3 method was used for the data with unequal variances. Student’s t test was used to analyse the differences between the two independent samples. Differences with p < 0.05 were considered statistically significant.

Results

BBB permeability

EB dye leakage after cerebral I/R is commonly used to evaluate BBB permeability. Following MCAO, the EB content in the ischaemic hemisphere was greater than that in the normal hemisphere. Moreover, as shown in Figure 2(b), EB extravasation in ischaemic hemispheres of the I/R group was markedly greater than that in the sham-operated group (2.14 ± 0.17 vs. 1.00 ± 0.01, p < 0.01), and the increase was significantly attenuated by ozagrel (18 mg kg⁻¹) or TSF (18 and 30 mg kg⁻¹) to 1.70 ± 0.13, 1.75 ± 0.11 and 1.82 ± 0.13 (ozagrel group p < 0.01 and both TSF group p < 0.05).

Brain water content

The protective effect of TSF on cerebral oedema was determined by detecting the cerebral water content (BWC) at 48 h after I/R. As shown in Figure 2(c), the BWC in the sham group was 74.3 ± 1.8%, which was markedly increased to 85.5 ± 1.6% in the I/R group. Compared with I/R group, in which the BWC treated with ozagrel (18 mg kg⁻¹) or TSF (18 and 30 mg kg⁻¹) evidently decreased to 80.1 ± 1.8%, 81.5 ± 1.4% and 78.9 ± 1.8%, respectively (p < 0.01). Although treatment with low dose of TSF (10.8 mg kg⁻¹) reduced BWC (84.0 ± 1.6%), no significant difference was obtained.

Figure 2.

Effect of TSF on cerebral I/R rat brains. (a) Typical images of EB extravasation at 48 h after MCAO. The blue region indicated the BBB broken (b) TSF-extenuated BBB hyperpermeability after ischemic stroke (n = 10). The pound sign denotes significant differences between I/R group and Sham group (^## p < 0.01). Asterisks denote significant differences between I/R group and other groups (**p < 0.01 and *p < 0.05). Results show evident EB leakage after I/R compared to the Sham groups (^## p < 0.01). Treatment with ozagrel (18 mg kg⁻¹) or TSF at 30 and 18 mg kg⁻¹ decreased BBB permeability (**p < 0.01 and *p < 0.05). (c) TSF-lessened brain oedema after I/R (n = 10). Results show significant brain oedema after I/R compared to the Sham groups (^## p < 0.01). Treatment with ozagrel (18 mg kg⁻¹) or TSF at 30 and 18 mg kg⁻¹ decreased BBB permeability (**p < 0.01). TSF: tetramethylpyrazine-2′-O-sodium ferulate; EB: Evans blue; MCAO: middle cerebral artery occlusion; I/R: ischaemia/reperfusion; BBB: blood–brain barrier.

TSF lessened infarct volume

Infarct size is a main index to estimate ischaemic injury. Compared with sham-operated group, the rats in the I/R group showed a prominent brain infarct size of 42.9 ± 4.8% and administration with ozagrel (18 mg kg⁻¹) or TSF (18 and 30 mg kg⁻¹) alleviated the cerebral infarct size significantly (p < 0.01; Figure 3).

Figure 3.

Effect of TSF on cerebral I/R rat brains. (a) Representative photographs of rat brain TTC staining in different groups at 48 h after I/R. The pale region was the infarct brain tissue and the red region was normal. (b) Post-treatment with TSF alleviated ischemic infarction after I/R (n = 10). Quantitative analysis of infarct size in different groups. The pound sign denotes significant differences between I/R group and Sham group (^## p < 0.01). Asterisks denote significant differences between I/R group and other groups (**p < 0.01). Results showed significant increase of ischemic infarct size in I/R rats compared to the Sham groups (^## p < 0.01). Treatment with ozagrel (18 mg kg⁻¹) or TSF 18 at and 30 mg kg⁻¹ could extenuate ischemic infarction (**p < 0.01). (c) TSF-attenuated neurological deficit. Compared with I/R group, the behavioural scores were significantly reduced in ozagrel (18 mg kg⁻¹) or TSF at 30 and 18 mg kg⁻¹ group (**p < 0.01 and *p < 0.05). But there was no significant difference between I/R group and TSF 10.8 mg kg⁻¹ group. TSF: tetramethylpyrazine-2′-O-sodium ferulate; EB: Evans blue; MCAO: middle cerebral artery occlusion; I/R: ischaemia/reperfusion; BBB: blood–brain barrier; TTC: 2,3,5-triphenyltetrazolium chloride.

TSF improved neurological deficit

As shown in Figure 3, the animals in the I/R injury group exhibited severe neurological deficit (3.1 ± 0.7) and showed circling towards the contralateral side and had a reduced mobility when compared to the sham group (p < 0.01). The rats in the ozagrel (18 mg kg⁻¹) or TSF (18 and 30 mg kg⁻¹) treated groups showed significant improvement in the behaviour when compared to the I/R (2.9 ± 0.37, 2.7 ± 0.25 and 2.9 ± 0.27, respectively; p < 0.01 and p < 0.05). Taken together, TSF post-treatment is beneficial for neurological deficit after I/R.

IHC analysis of AQP4

As shown in Figure 4 under microscope, the positive cells were brown and the expression of AQP4 cells were found orange–brown stained into view. The results of IHC analysis showed that AQP4 expression increased after I/R. Compared to the sham-operation group, the AQP4 expression in I/R group increased significantly (p < 0.01). Compared to I/R group, the AQP4 expression in ozagrel (18 mg kg⁻¹) group decreased significantly (p < 0.01), and TSF (18 and 30 mg kg⁻¹) group also decreased significantly (p < 0.05 and p < 0.01, respectively). There was no statistical difference of AQP4 expression compared I/R group with TSF (10.8 mg kg⁻¹).

Figure 4.

(a) IHC analysis of aquaporin-4 expression in the striatum at 48 h after I/R (n = 4) (IHC, ×400). (b) Semi-quantitative analysis of AQP4 expression according to the OD and grey value of positive cells by a professional image analysis software Image-Pro Plus 6.0 ( $\overline{x}$ ± s). ^## p < 0.01, compared with the sham-operation group; **p < 0.01; *p < 0.05, compared with I/R group. IHC: immunohistochemistry; OD: integrated density; AQP4: aquaporin-4; I/R: ischaemia/reperfusion.

Western blot analysis of JAM-1, occludin, AQP4 and MMP-9 expressions

Western blot analysis shown in Figure 5 revealed that the expression of JAM-1 and occludin in I/R group markedly reduced compared to the sham group. Interestingly, the reduced expression levels of JAM-1 and occludin caused by I/R injury were significantly reversed by ozagrel (18 mg kg⁻¹) or TSF (18 and 30 mg kg⁻¹). As shown in Figure 6, the expression of MMP-9 and AQP4 in I/R-treated group was upregulated compared with the sham group. Meanwhile, compared to I/R group, the administration of ozagrel (18 mg kg⁻¹) or TSF (18 and 30 mg kg⁻¹) significantly reduced the expression of MMP-9 and AQP4 (p < 0.01).

Figure 5.

TSF upregulated occludin and JAM-1 expressions. Representative photographs of western blot of (a) occludin; (b) JAM-1 and actin control in the brain at 48 h after MCAO; (c) and (d) The bar graph of western blot. The pound sign denotes significant differences between I/R group and Sham group (^# p < 0.05). Asterisks denote significant differences between I/R group and other groups (**p < 0.01 and *p < 0.05). Results showed significant diminished the expression of occludin and JAM-1 in I/R rats compared to the Sham group (^# p < 0.05). Treatment with zzagrel (18 mg kg⁻¹) or TSF at 18 and 30 mg kg⁻¹ could upregulate occludin and JAM-1 (*p < 0.05). TSF: tetramethylpyrazine-2′-O-sodium ferulate; JAM-1: junction adhesion molecule-1; MCAO: middle cerebral artery occlusion; I/R: ischaemia/reperfusion.

Figure 6.

TSF downregulated AQP4 and MMP9 expressions. Representative photographs of western blot of (a) AQP4, (b) MMP-9 and actin control in the brain at 48 h after MCAO. (c) and (d) The bar graph of western blot. The pound sign denotes significant differences between I/R group and Sham group (^# p < 0.01). Asterisks denote significant differences between I/R group and other groups (**p < 0.01). Results showed significant over expressions of AQP4 and MMP9 in I/R rats compared to the Sham groups (^# p < 0.01). Treatment with ozagrel (18 mg kg⁻¹) or TSF at 18 and 30 mg kg⁻¹ could downregulate AQP4 and MMP9 (*p < 0.05). TSF: tetramethylpyrazine-2′-O-sodium ferulate; JAM-1: junction adhesion molecule-1; MCAO: middle cerebral artery occlusion; I/R: ischaemia/reperfusion; AQP4: aquaporin-4; MMP-9: matrix metalloproteinase-9.

Discussion and conclusions

It remains a challenge for clinicians to cope with the brain oedema after I/R. Ozagrel is a selective TXA2 synthase inhibitor. ²⁴ It ameliorates platelet aggregation, vasoconstriction and brain oedema in acute cerebral ischaemia. Ozagrel has been widely used for the treatment of thrombotic or lacunar stroke over the last few years. It modulates the arachidonic acid cascade thereby reducing TXA2 and increases PGI2. Ozagrel administration improves the balance of PGI2/TXA2 during the acute phase of cerebral ischaemia. ²⁵

This study demonstrates that the new compound, TSF, which belongs to the same class with ozagrel, shows apparent protective effect on ischaemic stroke. TSF could be select as an alternative to ozagrel.

The results of the present study revealed TSF to be a promising management in this circumstance, demonstrating that post-treatment with TSF significantly reduced the BBB damage and cerebral oedema after I/R and, more importantly, relieved neurological deficits and brain infarction. Furthermore, TSF was found to attenuate I/R-induced degradation of the TJ proteins and suppressed in the expression of MMP-9 and AQP4.

In this study, we set the dosage of administration (10.8, 18 and 30 mg kg⁻¹) according to the phase I clinical trials dosage and the preliminary pharmacokinetics experiment of TSF. Phase I clinical trials have got very good effect. Our group have demonstrated that TSF ameliorates cerebral ischaemia disturbance, neuron damage and cognitive impairment elicited by ischaemia, consistent with previous research. ²⁶ However, to date, the effect of TSF on BBB disruption associated with cerebral I/R has not been reported.

Ischaemic stroke and subsequent reperfusion cause numerous severe clinical complications including BBB disruption and brain oedema. ² Likewise, in the present work, in the MCAO model of the study, an obvious BBB breakdown and brain oedema were observed after 48 h of reperfusion following 2 h of ischaemia, as evidenced by EB leakage and brain water determination. Interestingly, we demonstrated that the I/R elicited the permeability of BBB and brain oedema was alleviated significantly at 48 h after reperfusion compared with I/R model group, indicating the efficiency of TSF as a therapeutic strategy for this lesion. Here, we showed that TSF prevents BBB disruption and brain oedema in a cerebral MCAO and reperfusion rat model for the first time.

Brain oedema occurs commonly after a longer period of ischaemia, which is a mixed form consisted of cytotoxic oedema and vasogenic oedema. ²⁷ It is generally agreed that cytotoxic oedema occurs within 30 min after the ischaemia, ²⁸ while the vasogenic oedema appears at least several hours after an ischaemic injury, ²⁹ like the case in this study, and is mostly vasogenic. The BBB is a highly special structure between brain and blood circulation that maintains the homeostasis of the neural microenvironment and protects brain from undesirable penetration by external compounds, cells or excess fluid. ³⁰ We choose 48 h for the rest of our study with the purpose of exploring clinical relevance of early treatment of brain oedema formation after I/R injury.

It is generally believed that TJs affect the BBB permeability, which mediates the paraendothelial transport. ^{31 –33} The degradation of the tight-junction protein (TJP) is a decisive step in ischaemic BBB breakdown in stroke. The previous studies showed that oestrogen-α and oestrogen-β activation reduce TJ disruption after ischaemic injury and targeting oestrogen is a useful strategy for protecting the BBB from ischaemic stroke. ³⁴

Drugs that upgrade occludin expression have been demonstrated to increase transendothelial resistance and decrease BBB permeability. ³⁵ JAM-1 is a family of immunoglobulin protein localized within the intercellular cleft of TJ, which has a single transmembrane domain. JAM-1 plays a similar role in regulating the integrity of the BBB. ^36,37 Although a great deal has been learned about TJP, the exact mechanism of TSF on TJP remains unclear. Our experiments of western blotting and IHC staining demonstrated that both occludin and JAM-1 are upregulated, which are implicated in the BBB permeability modulation by TSF.

To further investigate the underlying mechanisms of protective effect of TSF against brain oedema, we detected the expression of AQP4 and MMP-9 in rat brains. AQP4, a water channel protein, is mainly expressed in astrocyte end feet around the surface of capillaries associated with the BBB, glia limitans and ependyma. ³⁸ Because of its critical role in water homeostasis and brain oedema formation, AQP4 is regarded as an attractive therapeutic target for various brain disorders, including hydrocephalus and cerebral ischaemic injury. ³⁹

MMPs are a family of zinc-dependent proteases that usually cleave components of the extracellular matrix. MMP-9 damages the BBB by degrading a number of extracellular matrix molecules, and thus contributes to the pathology of BBB, and thus contribute to the pathology of cerebral ischaemia. When the integrity of the BBB is compromised, inflammatory cells and fluid will infiltrate the brain, causing vasogenic oedema. ^{40 –42}

Notably, excessive AQP4 and MMP-9 contributes to weakening of the BBB, the major finding in our present study is that TSF treatment significantly downregulates AQP4 and MMP-9 protein expression after I/R, which results in the attenuation of cerebral oedema and suggests that the protecting effect of TSF may due to the amelioration of BBB.

Our experiments of western blotting and IHC staining suggested that the protective effect of TSF on brain oedema may rely on the regulation of TJ proteins and inhibition ofAQP4 and MMP-9 of BBB. Nonetheless, the details of the underlying role of TSF in attenuation of BBB permeability after I/R are not clear at present and need further elucidation. The results of the present study showed that the administration of TSF resulted in an apparent improvement in BBB damage and cerebral infarction, as demonstrated by EB leakage, TTC staining and neurological score, in addition to the attenuation of brain oedema.

In conclusion, this study has shown that TSF displays protective effect in brain against I/R-induced BBB disruption and brain oedema by interference in the degradation of TJ protein and suppressing AQP4 and MMP-9 expression, accompanying by an improvement in cerebral infarction and neurological score. Combined with previous results, these findings suggest TSF as a promising alternative approach for brain oedema induced by cerebral I/R damage. Nonetheless, much more works are needed to clarify this issue.