Mechanism of salvianolic acid B neuroprotection against ischemia/reperfusion induced cerebral injury

The purpose of this study was to evaluate the cerebral protection of salvianolic acid B (Sal B) against cerebral I/R injury and investigate the underlying mechanism. As shown by 2,3,5-Triphenyltetrazolium chloride (TTC) staining and magnetic resonance imaging (MRI) analyses, Sal B significantly reduced cerebral infarct size, and accompanied with improved neurobehavioral functions as indicated by the modified Bederson score and Longa five-point scale. Sal B decreased the production of reactive oxygen species (p < .05, n = 10). The data of Western blotting and reverse transcription quantitative real time polymerase chain reaction (qRT-PCR) analyses showed that the expression of GFAP, Iba1, IL-1β, IL-6, TNF-α and Cleaved-caspase 3 was significantly reduced by Sal B in I/R injured brain tissues as compared to corresponding controls (p < .05, n = 10). Over activation of astrocytes and microglia were inhibited by Sal B as shown by immunostaining of GFAP and Iba 1. These data suggest that Sal B has neural protective effects against I/R-induced cerebral injury and could be an effective candidate for further development of clinical therapy.


Introduction
Stroke is a leading cause of death worldwide and remains a public health concern with severe economic and social burden (Shichita et al., 2009) . Acute cerebral infarction accounts for about 80% of all stroke. Currently, the most effective approach to treat acute brain infarction is to restore blood flow as early as possible to save ischemic-damaged tissues by thrombolysis, but reperfusion after thrombolysis can aggravate injury and seriously affect therapeutical efficacy (Hausenloy and Yellon, 2004;Matsumura et al., 1998).
Ischemia/reperfusion (I/R)-induced brain injury is a complex pathophysiological process. Oxidative stress and inflammation are considered two major factors involved in I/R-induced cell death. In endothelial cells, oxidative stress reduces the bioavailability of nitric oxide (NO), which is a potent vasodilator and an inhibitor of platelet aggregation and leukocyte adhesion (Yilmaz and Granger, 2010). Excessive accumulation of reactive oxygen species (ROS), which induce cell death by modulating a series of intracellular signaling pathways (Droge, 2002), particularly, the activation of mitogen-activated protein kinases (MAPKs) and phosphatidylinositol-3-kinase (PI3K)/Akt pathways (Jerome et al., 1994).
Inflammation represents another cascade of events triggered by I/R damage. Inflammation begins in the intravascular compartment after arterial occlusion. In blood vessels, ischemia /reperfusion leads to platelet aggregation and cytokine release, complement activating, and arachidonic acid metabolites (AA) releasing (Iadecola and Anrather, 2011). The inflammatory cytokines can mediate neutrophil chemotaxis to ischemic vascular endothelial. Microvascular permeability is changed and the blood brain barrier is damaged (Lindsberg et al., 2010). Peripheral inflammatory cells infiltrate to the ischemic brain tissue and cause cerebral edema and acute inflammation (Schroder and Tschopp, 2010 Astrocytes and microglia who play key role in maintaining normal function of neurons are two types of supportive cells in central nervous system (CNS). Astrocytes and microglia constitute the first line of responders to cerebral ischemic injury and can be activated a few minutes after cerebral ischemia (Streit et al., 2005), which contribute to post-ischemic inflammation by producing ROS, TNF, IL-1b and other pro-inflammatory factors. In the meantime, astrocytes release neurotrophic factors, which help to repair damaged neurons after cerebral ischemia, while microglia phagocytose debris and eliminate pathogens, secrete anti-inflammatory and nerve growth factor, promote the neuron survival around infarction area (Peachell, 2006).
Salvianolic acid B (Sal B) is the most abundant and bioactive compound of various Danshen phenolic acids extracted from Salvia miltiorrhiza mixture (Jiang et al., 2005;Zhou et al., 2005;Morrison and Li, 2011;Diaz et al., 2013), and is one of the most powerful antioxidants found in nature. Salvia miltiorrhiza Bunge (Danshen) is a popular traditional Chinese medicine and has been widely used in treating various diseases such as coronary artery diseases and myocardial infarction in China (Deng et al., 2015). Sal B was shown to prevent I/R-induced rat brain injury by reducing free radicals and improving energy metabolism (Chen et al., 2000;Zhao et al., 2008), improve regional cerebral blood flow in ischemic hemisphere, inhibit platelet aggregation, and promote recovery of motor function after cerebral ischemia/reperfusion injury (Liu et al., 2007). Sal B also showed anti-apoptotic after cerebral I/R-injury since it was able to up-regulate Bcl-2 expression after I/R-injury, and significantly reduced the expression of bax (Lay et al., 2003). However, the detail molecular mechanism of Sal B in protecting I/R-induced cerebral injury is still not well defined.
To evaluate the protective effect of Sal B against cerebral I/R-injury and explore underlying mechanism, we tested the effects of Sal B after I/R-induced cerebral injury in middle cerebral artery occlusion (MCAO) mice model by examining cerebral infarct size, neurological function, and cytokine production. We also evaluated the roles which astrocyte and microglia played in Sal B mediated brain tissue protection.

Sal B improved neurobehavioral functions following I/R injury
The chemical structure of Sal B is shown in Supplemental Fig. 1A and the purity of Sal B was 99.4% as analyzed by HPLC (Suppl. Fig. 1B). Mouse cerebral I/R-injury was induced by 1 h of the middle cerebral artery occlusion followed by 24 h reperfusion, accompanied with the treatments of Sal B at designated concentrations. Sal B effects on the neurobehavioral functions following cerebral I/R injury were evaluated by modified Bederson scores and Longa five-point scales (Bederson et al., 1986;Longa et al., 1989), results are shown in Fig. 1. All mice in sham group had a neurologic grade of 0 (n = 10), 40% mice of I/R group were rated grade 3, another 40% in grade 4 and the left 20% in grade 5 (n = 10). The mice in 10 mg Sal B treated I/R group had 60% in grade 3 and 40% in grade 4 (n = 10). The mice in 20 mg Sal B treated I/R group, 60% of them rated grade 4 and 40% grade 3 (n = 10). The mice in 40 mg Sal B treated I/R group, 60% of them rated grade 2 and 40% grade 3 (n = 10). The mice in 60 mg Sal B treated I/R group, 80% of them rated grade 2 and 20% grade 3 (n = 10) (Fig. 1A).
While measured by Longa five-point neurobehavioral scores, all mice in sham group had a neurologic grade of 0 (n = 10). The mice of I/R group had 80% in grade 3 and 20% in grade 2 (n = 10), respectively. The mice in 10 mg Sal B treated I/R group had 60% in grade 3, 40% in grade 2 (n = 10). 60% of the mice in 20 mg Sal B treated I/R group were rated grade 2 and 40% grade 3 (n = 10). 60% of the mice in 40 mg Sal B treated I/R group were rated grade 2 and 40% grade1 (n = 10). 60% of the mice in 60 mg Sal B treated I/R group were rated grade 2 and 40% grade1 (n = 10). (Fig. 1B). These data indicated that Sal B administration was able to ameliorate neurobehavioral deficits caused by middle cerebral artery occluded cerebral (I/R) in mice.

Sal B reduced infracted area and preserved integrity of brain structure
Brains were removed rapidly at 24 h end point after I/R. The infarct volume was calculated as previously described using the formulae (Lin et al., 1993): Infarct volume (%) = [(volume of the normal hemisphere-non infarct volume of the infarct hemisphere)/volume of the normal hemisphere]Â100% Histological analyses of infarct sizes after cerebral I/R injury and Sal B treatments were performed by TTC staining. The infarct areas were analyzed by NIH Image J software and the data are shown in Fig. 2A. Ratio of infarct area relative to whole brain was calculated and shown in Fig. 2B. Compared with I/R control group (23.01 ± 1. 687% of total brain area) (p < .05, n = 10), we found that the infarct area of whole brain was significantly reduced in the groups treated with 20 mg/kg Sal B (15 ± 0.513%) (p < .05, n = 10), 40 mg/kg Sal B (9.633 ± 0.6013%) (p < .05, n = 10) and 60 mg/kg Sal B (8.623 ± 0.9 732%) (p < .05, n = 10). No significant difference was found in10mg/kg Sal B treated I/R group (22.38 ± 1.371%) (p < .05, n = 10).
The Sal B effects on reducing infarct size after cerebral I/R injury were further verified by MRI scanning (Fig. 2C). Ratio of infarct area relative to whole brain was calculated and shown in Fig. 2D. In line with our TTC staining results, the infarct area of whole brain was significantly reduced in I/R groups treated with 20 mg/kg Sal B (15 ± 0.513%) (p < .05, n = 8), 40 mg/kg (12.92 ± 1. 061%) (p < .05, n = 8) and 60 mg/kg Sal B (9.880 ± 1.100%) (p < .05, n = 8), respectively in comparison with I/R group (21.07 ± 0.5 461%) (p < .05, n = 8), no significant difference was foundin10mg/ kg Sal B treated I/R group (23.47 ± 7419%) (p < .05, n = 8). Both TTC staining and MRI scanning data showed that Sal B can effectively reduce brain infarct size of cerebral I/R injury and preserve integrity of brain structure.

Sal B down-regulated ROS production
Sal B is known to regulate the production of ROS, to further characterize its antioxidative effects, the production of ROS were evaluated by ELISA. Remarkably, we observed that Sal B significantly down-regulated ROS production after I/R injury in 60 mg/ kg Sal B treated group (62088 ± 1716 IU/ml) versus that in I/R group (67825 ± 1102 IU/ml) (p < .05, n = 10) (Suppl. Fig. 2).

Sal B regulated inflammatory response following I/R injury
Enhanced inflammatory response begins immediately after arterial occlusion and plays an important role in post-ischemic brain tissue damage (Iadecola and Anrather, 2011). Therefore, we examined the effects of Sal B on production of pro-inflammatory cytokines, IL-1b, IL-6 and TNF-a, following cerebral I/R injury. qRT-PCR data showed that mRNA expression of IL-1b and IL-6 in I/R injured brain tissues was significantly down regulated with increased doses of Sal B (20, 40 and 60 mg/kg) treatments compared to that without Sal B treatment (p < .05, n = 10), but not in10mg/kg Sal B treatment group ( Fig. 3A and B). The TNF-a mRNA transcription was significantly down regulated by 10, 20, 40 and 60 mg/kg Sal B treatments compared to that without Sal B treatment (p < .001, n = 10) (Fig. 3C). Western blotting data showed that, relative to none Sal B treatment I/R group, TNF-a was significantly suppressed by 20, 40, 60 mg/kg Sal B treatments following ischemia (p < .01, n = 8), but not in the groups treated with 10 mg/ kg Sal B ( Fig. 3D and G); IL-1b was significantly suppressed by 40 and 60 mg/kg Sal B treatments following ischemia (p < .01, n = 8), respectively ( Fig. 3D and E); IL-6 was significantly suppressed by The infarct size was evaluated as the percentage of infarct area relative to the whole brain (p < .01, n = 10). (C) Representative photomicrographs taken by MRI scanning. Infracted region was showed white. (D) The infarct size was scanned with MRI, evaluated as the percentage of infarct area relative to the whole brain (p < .001). The percentage change of infracted region is shown as mean ± SD. ( *** p < .001, n = 8).
2.5. Sal B affected the expression of GFAP, Iba1 and Cleaved-caspase3 following the cerebral I/R injury The inflammatory responses are often associated with astroglial and microglia activation following brain injury. To investigate the effects of Sal B on I/R injury induced astroglial and microglia activation, qRT-PCR and Western blotting were performed to examine changes of the mRNA and protein expression of glial fibrillary acidic protein (GFAP) and ionized calcium-binding adaptor molecule 1 (Iba1), respectively. GFAP is the major intermediate filament, a cell-specific marker used to distinguish astrocytes from other glial cells. Iba1 is an inflammation-associated Ca 2+ binding protein produced by microglia (Ito et al., 1998). qRT-PCR results showed that the mRNA expression level of GFAP in the groups treated with 20 mg/kg, 40 mg/kg and 60 mg/kg Sal B following ischemia decreased significantly relative to control I/R group (p < .05, p < .01, n = 10) (Fig. 4A). The mRNA expression level of Iba1 only significantly decreased in the group treated with 60 mg/kg Sal B following ischemia relative to I/R control group (p < .01, n = 10) (Fig. 4B). Western blotting results showed that the expression of GFAP protein was significantly attenuated in the groups treated with 40 mg/kg and 60 mg/kg Sal B, respectively, while the expression of Iba1 proteins was only attenuated in the group treated with 60 mg/kg Sal B (p < .01, n = 8) (Fig. 4C-E).
Caspase3 is involved in the activation cascade of caspases responsible for apoptosis execution closely related to neuronal death and the cleaved form of Caspase3 represents activation of Caspase3. Western blot data showed that the expression of Cleaved-caspase3 proteins was significantly attenuated in the groups treated with 40 mg/kg and 60 mg/kg Sal B following I/R (p < .01, n = 8), but not in the groups treated with 10 mg/kg and 20 mg/kg Sal B ( Fig. 4C and F). Data above suggested that Sal B possesses capability to prevent astroglial and microglia activation, and reduced cleaved-caspase3 following I/R injury.
We further conducted GFAP, Iba1 immunostaining and TUNEL analyses in penumbra area of hippocampus regions. The distribution and expression of GFAP was attenuated in 60 mg/kg Sal B treated group following I/R injury (p < .05, n = 6) ( Fig. 5A and D), indicating that Sal B treatment could decrease the activity of GFAP, suppressed astrocyte activation, Sal B could also restrain the activity of Iba1, suppressed microglia activation ( Fig. 5B and E). TUNEL labeling was used to investigate the programmed cell death due to I/R injury. The data clearly showed that less apoptotic cells presented in hippocampus region in the I/R group treated with 60 mg/kg Sal B compared to none Sal B treated I/R group (p < .05, n = 6) (Fig. 5C, 5F). These data suggested that Sal B could reduce I/R-induced brain cell apoptosis.

Discussion
Present study showed that Sal B mediated cerebral protection against I/R induced injury in vivo as indicated by the reduction of infarct volume in cerebral cortex, which is consistent with previous observations (Chen et al., 2000). Sal B treatments also led to the improvement neurobehavioral deficits in the MCAO animal model. The reduction of infarct volume by Sal B was verified by TTC staining and MRI scanning. TTC staining requires sacrificing and processing mice, the whole process takes more than 2 h. In contrast, MRI scanning is able to continuously monitor brain injury without sacrificing the subjects, which is able to dynamically visualize the progress of brain injury resulted from I/R. MRI scanning greatly simplifies experimental process in comparison with TTC staining and avoids the error caused by artificial factors created in TTC staining, which makes the results more reliable (Sicard, 2006).
Oxidative stress, caused by ROS release, is considered to be a major destructive factor leading to nerve cell apoptosis induced by cerebral I/R injury (Carden and Granger, 2000). The cerebral I/R can cause the opening of the mitochondrial permeability transition pore, which results in the dissipation and uncoupling of the electrochemical gradient of the inner mitochondrial membrane, leading to release of ROS (Ginhoux et al., 2010). Sal B is an Fig. 3. Sal B regulated inflammatory response following I/R injury. Western blotting and qRT-PCR were performed to analyze the IL-1b, IL-6 and TNF-a expression following Sal B treat the I/R-injury mice. (A) IL-1b mRNA was down regulated with increased doses of Sal B (20, 40 and 60 mg/kg) treatments compared to that without Sal B treatment (p < .05, n = 10). (B) IL-6 mRNA was down regulated with increased doses (20, 40 and 60 mg/kg) of Sal B treatments compared to that without Sal B treatment (p < .05, n = 10). (C) TNF-a mRNA was down regulated by 10, 20, 40 and 60 mg/kg Sal B treatments compared to that without Sal B treatment (p < .001, n = 10). Western blotting patterns of IL-1b, IL-6 and TNF-a from Sal B treat the I/R-injury mice. The fold change of stripes were quantified and represented in E, F, G. (p < .001, n = 8). Quantization of mRNA and protein expression was determined after normalized to the b-actin control and presented as mean ± SD ( *** p < .001, n = 8-10).
antioxidant and able to eliminate free radical and superoxide anion. Its antioxidation capability is better than vitamin C, vitamin E, and Ginkgo biloba extracts (Gao et al., 2006). We had also observed the reduction of ROS in Sal B treated MCAO models, which suggested that Sal B antioxidative property might play a major role in protecting brain tissues from I/R-induced injury and preserve their integrity.
Accompanied with oxidative stress, persistent inflammation was another major factor which leads to central nervous system damage, and affects nerve cell survive (Yang et al., 2009;Coyne et al., 2006). We observed the up-regulation of several inflammatory cytokines after I/R injury, and found Sal B could suppress the expression of pro-inflammatory cytokines including TNF-a, IL-1b and IL-6 which had been associated with increased permeability induced by I/R injury (Guo et al., 2013;Wu et al., 2014;Homma et al., 2013) Since astrocytes is an important component of the blood brain barrier, the release of TNF-a can cause microvascular dysfunction (Kassiri et al., 2009). Sal B might suppress astrocyte activation and reserve microvascular integrity through downregulation of TNF-a.
What we found that Sal B reducing the secretion of inflammatory cytokines and protect cerebral nerve cells during I/Rinduced injury was in line with the previous reports Wang et al., 2010). Since the oxidative stress is highly correlated with inflammatory in response to tissue damage, thus the antioxidant effects of Sal B may play a major role to its antiinflammatory effects (Seif, 2013). Therefore, both antioxidant and anti-inflammatory effects induced by Sal B might improve the living environment of neural cells, which will lead to the improvement of nerve cell survival and function . However, it is still not clear if Sal B was able to target directly to microglia-mediated inflammatory pathways or indirectly target to its antioxidative properties, since we only found Sal B could inhibit the activation of microglia in relatively high concentration (60 mg/kg) (Figs. 4C and E, 5B and E).
In consistent with inhibiting ROS and inflammatory factors production mediated by Sal B treatments, our immunostaining data also showed that Sal B could suppress the activation of microglia and astrocytes effectively. Microglia and astrocytes are two major cell types in response to inflammation in brain tissues mentioned above. Microglias are phagocytic cells that reside in brain, and form stretch under the stimulation of inflammatory stress. Astrocytes functioned as an important component of the blood brain barrier. Although astrocytes are able to protect brain tissue damage by releasing nerve growth factors, excessive activation of astrocytes can also lead to over production of NO, TNF-a, matrix metalloproteinase, etc. which will promote ischemia-induced brain tissue necrosis (Chen and Swanson, 2003;Joe et al., 2012). Suppressing excessive reaction and inflammatory process of microglia and astrocytes has been tackled to alleviate the progression of neurological diseases such as stroke and neurodegenerative diseases (Kaushal and Schlichter, 2008). Therefore, Sal B may be able to ameliorate neurotoxic effects by targeting both of these two cell types.
In summary, Sal B increases the viability of microglia and astrocytes, reduces the infarct size and improves the neurobehavioral function against I/R-induced brain injury in mice, which suggested that Sal B can be developed for further clinical application of treating stroke and related diseases.

Cerebral ischemia model
All animal procedures were in adherence to the Guide for the Care and Use of Laboratory Animals approved by Fujian Provincial Office for Managing Laboratory Animals and was overseen by the Fujian Normal University Animal Care and Use Committee (ethics license numbers IACUC-2016Y0027). All in vivo studies complianced with the ARRIVE guidelines (Animal Research: Reporting in Vivo Experiments).
Male C57BL/6 mice from Silaike Experimental Animal LLC (Shanghai, China) were used for all experiments and housed in a temperature (22 ± 2°C) and humidity controlled environment on a 12 h light/dark cycle with free access to food and water. Eight to ten-week-old mice that weighed 20-25 g were anesthetized with sodium pentobarbital (60 mg/kg) intraperitoneally and used for generating cerebral ischemia model as follows. A laser Doppler flowmetry (moorVMS-LDF2) was used to measure local cortical blood flow supplied by the middle cerebral artery (MCA) during the operation. For the ligation of right common carotid artery (CCA), we inserted 7-0 nylon monofilament with a rounded tip into the right CCA. During the right middle cerebral artery occlusion (MCAO), 80% reduction of the blood flow was determined by the laser Doppler flowmetry. Sixty minutes after MCA occlusion, nylon filament was withdrawn to allow reperfusion of the right MCA. Sham surgery was performed by the same procedure without occlusion.
To detect the best ratio of Sal B, mice were randomly assigned into six groups: Sham-operated mice were given saline (Sham); ischemia-reperfusion mice were given saline (I/R); ischemiareperfusion mice were given 10 mg/kg Sal B (I/R-Sal B); ischemia-reperfusion mice were given 20 mg/kg Sal B (I/R-Sal B); ischemia-reperfusion mice were given 40 mg/kg Sal B (I/R-Sal B); ischemia-reperfusion mice were given 60 mg/kg Sal B (I/R-Sal B).
Sal B was dissolved in sterile endotoxin-free 0.9% NaCl as 10 mg/ml. Mice were administered with Sal B at the same time of reperfusion. The number of animals in each group was more than 60 to keep the animal number more than 10 for each group.

Neurological function assessment
The injury severity of treated mice were assessed by modified Bederson score and Longa five-point scale as previously described (Bederson et al., 1986;Longa et al., 1989). The modified Bederson scores were following: 0, no deficit; mice with score 1, lost forelimb flexion; 2, as for 1, but plus decreased resistance to lateral push; 3, indicated unidirectional circling; mice with score 4, displayed longitudinal spinning or seizure activity; mice with score 5, showed no movement.
For Longa five-point scale (Longa et al., 1989): a score of 0 indicated no neurologic deficit; a score of 1 showed failure to extend left forepaw fully with a mild focal neurologic deficit; a score of 2 showed circling to the left and exhibited a moderate focal neurologic deficit; and a score of 3 showed falling to the left and had a severe focal deficit; mice with a score of 4 showed not walk spontaneously and had a depressed level of consciousness.

Calculation of infarct volume
The infarct volume was determined by TTC and MRI analyses. TTC measurements: animals were anesthetized with sodium pentobarbital (60 mg/kg), brain tissues were sectioned coronal and consecutively at 1 mm thick. Coronal brain slices were incubated in 1% of the 2,3,5-Triphenyltetrazolium chloride (TTC) buffer, 37°C for 20 min. After incubation, brain slices were fixed in 4% polyformaldehyde for 24 h, the slices were then photographed. The red regions represented non-infarct tissue part, and pale white regions displayed infarct tissue part. Morph metric measurement of the infarct areas were analyzed by Image-Pro Plus version6.0.
Magnetic resonance imaging (MRI) analyses were conducted as follows: after inhalation anesthesia by 2% isoflurane, mice were placed in a 7T MRI scanner (Bruker, Germany), using mouse specific coils. The image information of T2WI phase was collected to calculate infarcted areas. Thus, the changes of cerebral infarction in MRI could be determined.

Immunohistochemistry
Brains were fixed with 4% paraformaldehyde, cryoprotected with 20% sucrose-potassium-PBS (KPBS), embedded in optimum cutting temperature compound (OCT), cryostat sectioned into coronal (10 lm) sections and mounted on glass slides which were coated with OCT (Chen et al., 2009). Tissue-specific localization of astrocytes and microglia was examined immunofluorescent staining of the antibodies against specific marker protein GFAP and Iba1. The sections were blocked by incubating with 10% goat serum in PBS for 1 h at room temperature, then incubated with anti-GFAP antibody (diluted 1:1000, Abcam, catalog #ab7260) and Rabbit polyclonal antibody to Iba1 (Working Concentration: 2ug/ml, Wako, catalog #019-19741) at 4°C overnight, remaining processes was following standard procedures. After PBS-T washing, slides were incubated with Donkey anti-Rabbit IgG (H + L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor Ò 488 (Thermo Fisher, catalog # A-21206), and cell nuclear was stained with DAPI (ZSGB-BIO, catalog #ZLI-9058). Images were taken by a Zeiss fluorescence microscope and image analyses were performed using NIH Image software.
For TUNEL assay, brains were cryostat sectioned into coronal (10 lm) slides and mounted on glass slides which were coated with OCT. Apoptosis was detected by TUNEL labeling using Dead-End TM Fluorometric tunel system (Promega, catalog#G3250) following the manufacturer instructions. After TUNEL labeling, nuclei were stained with DAPI (ZSGB-BIO, catalog #ZLI-9058), and TUNEL positive cells were observed by Zeiss fluorescence microscope and image analyze were performed using NIH Image software.

Statistical analysis
All data were presented as mean ± SD. Statistical significance was determined using one way analysis of variance (AVOVA) with Dunnett post hoc t-test. Findings were considered to be significantly different when a p value < .05 was obtained. All statistical analyses were performed using the SPSS20.0 software program (SPSS Inc, Chicago, IL, USA).