Tetramethylpyrazine nitrone activates the BDNF/Akt/CREB pathway to promote post‐ischaemic neuroregeneration and recovery of neurological functions in rats

Abstract

Background and Purpose

Neuronal regeneration from endogenous precursors is an attractive strategy for the treatment of ischaemic stroke. However, most stroke‐generated newborn neurons die over time. Therefore, a drug that is both neuroprotective and pro‐neurogenic may be beneficial after stroke. Here, we assessed the neurogenic and oligodendrogenic effects of tetramethylpyrazine nitrone (TBN), a neuroprotective drug candidate for stroke, in a rat model of ischaemic stroke.

Experimental Approach

We used Sprague Dawley rats with middle cerebral artery occlusion (MCAO). TBN was administered by tail vein injection beginning at 3 h post ischaemia. Therapeutic effect of TBN was evaluated by neurological behaviour and cerebral infarction. Promotion of neurogenesis and oligodendrogenesis was determined by double immunofluorescent staining and Western blotting analyses. Primary cultures of cortical neurons were used to assess the effect of TBN on neuronal differentiation in vitro.

Key Results

TBN reduced cerebral infarction, preserved and/or restored neurological function and promoted neurogenesis and oligodendrogenesis in rats after MCAO. In addition, TBN stimulated neuronal differentiation on primary culture of cortical neurons in vitro. Pro‐neurogenic effects of TBN were attributed to its activation of the AKT/cAMP responsive element‐binding protein through increasing brain‐derived neurotrophic factor (BDNF) expression, as shown by the abolition of the effects of TBN by a specific inhibitor of BDNF receptor ANA‐12 and by the PI3K inhibitor LY294002.

Conclusion and Implications

As TBN can simultaneously provide neuroprotection and pro‐neurogenic effects, it may be a promising treatment for both acute phase neuroprotection and long‐term functional recovery after ischaemic stroke.

Abbreviations

BDNF

brain‐derived neurotrophic factor

CNPase

2′,3′‐cyclic nucleotide 3′‐phosphodiesterase

CREB

cAMP responsive element‐binding protein

DG

dentate gyrus

DIV

days in vitro

LV

lateral ventricles

MBP

myelin basic protein

MCAO

middle cerebral artery occlusion

NPCs

neuronal precursor cells

NSS

neurological severity score

OPCs

oligodendrocyte progenitor cells

PSD95

post‐synaptic density protein 95

rCBF

relative cerebral blood flow

SVZ

subventricular zone

TBN

tetramethylpyrazine nitrone

TrkB

tyrosine receptor kinase B

Introduction

Stroke remains one of the major causes of death and disability worldwide (Mozaffarian et al., 2016). Despite notable advances in exploring the pathophysiology in the past decades, over 1000 therapeutic agents for treating ischaemic stroke have failed (Minnerup et al., 2014). In the ischaemic region, neurons rapidly die due to a cascade of biochemical changes. Therefore, pharmacological neuroprotection (Cook et al., 2012; Papadakis et al., 2013; Zhang et al., 2016b) and neuronal replacement from endogenous precursors (Arvidsson et al., 2002; Magnusson et al., 2014; Najm et al., 2015) or engrafted neural stem cells (Wang et al., 2016) might be attractive strategies for the treatment of ischaemic stroke.

There is evidence suggesting that ischaemic stroke markedly induces neurogenesis in the dentate gyrus (DG) and subventricular zone (SVZ) lining the lateral ventricles (LV) (Jin et al., 2001; Arvidsson et al., 2002), sites where the neuronal precursor cells (NPCs) in the adult brain is mostly located (Gross, 2000). New neurons migrate into the severely damaged area and replace the dead neurons. However, over 80% of the new neurons die during 2–6 weeks after‐ischaemia. Only a very small fraction, approximately 0.2% of dead neurons, are replaced by the new neurons, by 6 weeks after stroke (Arvidsson et al., 2002), for reasons not yet clear. If the formation and survival of new neurons could be chemically stimulated, a novel therapeutic strategy might be developed for stroke.

Tetramethylpyrazine, a main active ingredient of the herbal medicine Ligusticum wallichii Franchat (Chuanxiong), promotes the proliferation and differentiation of NPCs (Tian et al., 2010; Xiao et al., 2010) and enhances NPCs migration towards the ischaemic region in rats with middle cerebral artery occlusion (MCAO), a model of stroke (Kong et al., 2016). Tetramethylpyrazine nitrone (TBN), a novel nitrone derivative of tetramethylpyrazine, is a potent free radical scavenger with multifunctional neuroprotective effects in rat and monkey models of ischaemic stroke (Sun et al., 2008; Sun et al., 2012; Zhang et al., 2016b). TBN readily penetrates the blood–brain barrier (BBB) and displays adequate pharmacokinetic and safety profiles (Zhang et al., 2016b) and has advanced to Phase I clinical trial for the treatment of ischaemic stroke in China. While our previous studies mainly focused on neuroprotective effects of TBN at the acute stage of stroke in rats and monkeys, its long‐term effects on the restoration of neurological deficits and neuronal regeneration after stroke have not been explored. Here we report that TBN treatment promoted neuronal regeneration and long‐term functional recovery after ischaemic stroke in rats with MCAO.

We demonstrated that TBN reduced cerebral infarction, restored neurological function and promoted neurogenesis and oligodendrogenesis in the transient MCAO (t‐MCAO) rats. Activation of Akt/cAMP responsive element‐binding protein (CREB) through increasing the expression of brain‐derived neurotrophic factor (BDNF) was identified as contributing to the pro‐neurogenic effects of TBN. Most importantly, the combined properties of neurogenesis, oligodendrogenesis and neuroprotection in a single molecule, as in TBN, are more likely to be effective than any individual activity alone for treatment of neurodegenerative diseases, particularly for ischaemic stroke.

Methods

Animals

All animal care and experimental protocols conformed to internationally accepted ethical standards (Guide for the Care and Use of Laboratory Animals. NIH Publication 86–23, revised 1985) and were approved by the Institutional Animal Care and Use Committee of Jinan University (Guangzhou, China). All efforts were made to minimize the numbers of animals used and ensure minimal suffering. Animal studies are reported in compliance with the ARRIVE guidelines (Kilkenny et al., 2010; McGrath and Lilley, 2015). Sprague Dawley (SD) rats (obtained from Guangdong Medical Laboratory Animal Center, Guangzhou, China) weighing 290 ± 10 g were used. The rats were housed five to seven per cage on a 12 h light/dark cycle with free access to food and water in a temperature (20–25°C)‐ and humidity (30–50%)‐controlled animal facility.

Sample size calculations were based on the ability to detect a 10% difference in infarct sizes between drug and vehicle by the 2,3,5‐triphenyltetrazolium chloride (TTC) staining at a power of 0.8, α = 0.05 and an assumed standard deviation of 6% of group means according to our pilot study. Power and Sample Size Calculation software calculated that a sample size of eight rats per experimental group was needed. Based on this power calculation, for the proposed experiments, we used a sample size of 8–10 rats per experimental group to control for possible loss of animals. For therapeutic efficacy studies on permanent MCAO (p‐MCAO) and t‐MCAO rat, to avoid any influence of sex on the outcomes, we used half male and half female animals, total n = 14–18 per group.

MCAO surgery

SD rats were anaesthetized with 2.0–2.5% (v./v.) isoflurane in air by an animal anaesthesia ventilator system (RWD Life science, Shenzhen, China). The t‐MCAO rat was produced as previously described (Guan et al., 2012). Briefly, the left common carotid artery (CCA), external carotid artery (ECA) and internal carotid artery (ICA) were carefully isolated. A silicone‐coated nylon suture (0.34–0.36 mm diameter) was inserted into the ECA and was advanced retrogradely to the bifurcation of the CCA from where it was advanced a distance of approximately 18–22 mm into the ICA. After a 2 h occlusion of the MCA, reperfusion was achieved by removal of the suture. In the case of p‐MCAO, the surgery was similar, but the suture was left in place for 24 h without reperfusion.

To ensure the reproducibility of ischaemia and reperfusion, relative cerebral blood flow (rCBF) above the core of the middle cerebral artery territory (AP +2.0 mm, ML −4.0 mm) was monitored by a laser Doppler blood flow meter (PeriMed, Stockholm, Sweden). MCA occlusion resulted in a sharp decrease in rCBF to approximately 30% of the pre‐ischaemic baseline level. After reperfusion, the rCBF rapidly recovered to 80% of the pre‐ischaemic baseline level (Supporting Information Figure S1). No significant differences were detected between animals assigned to vehicle‐ or TBN‐treated groups (Supporting Information Figure S1). Venous blood gas was monitored by a blood gas analyser (ABL80basic, Radiometer) before and 5 min after the ischaemia. There were no significant differences among groups for blood gas parameters before and after ischaemia (Supporting Information Table S1). Body temperature was maintained at 37°C throughout the surgery with a heating blanket and feedback system (RWD Life Science). Rats were closely monitored for the following 24 h. Buprenex (0.5 mg·kg⁻¹, subcutaneous injection) was administered for analgesia. Penicillin (5 × 10⁴ U per rat) was also given to prevent infection, by intramuscular injection once daily for total 3 days.

Animals were excluded from analysis when the following occurred: neurological severity score (NSS) < 5 at 3 h after MCAO, intracerebral haemorrhage, underweight (animal's body weight loss exceeds 40% pre‐surgical weight) or died (Llovera et al., 2015). The numbers of animal included and excluded is shown in Supporting Information Table S2. The neurobehavioral assessments and quantitative analysis for cerebral infarction and immunohistochemistry/immunofluorescence were performed by investigators blinded to the experimental grouping and drug treatment.

Drug treatment

TBN was dissolved in 0.9% saline and administered i.v. through the tail vein. For the dose–response study with p‐MCAO rats, TBN or vehicle (0.9% saline) was given at 3 h and then 6 h after MCAO. For the therapeutic window study on p‐MCAO rats, at 3, 6 or 9 h after MCAO, the first dose (30 mg·kg⁻¹) was given; 3 h thereafter, the second dose (30 mg·kg⁻¹) was administered. For t‐MCAO rats, TBN or vehicle (0.9% saline) was given at 3 h and then 6 h after MCAO on the first day and then given twice daily with a 6 h interval for a total of 7 days.

Neurobehavioral assessment

Rats were trained in the sensorimotor tests for 3 days before ischaemia and measured at different time points after MCAO. Rotorod and adhesive tape removal tests were done as previously reported (Zhang et al., 2016a). Neurological deficits were scored by using the previously reported NSS (Chen et al., 2001), which is a composite of motor, sensory, reflex tests and beam balance. NSS was graded from 0 to 18 (normal score = 0, maximal deficit score = 18).

Novel object recognition task

Novel object recognition tasks were performed as previously described (Camarasa et al., 2010). Rats were habituated to an open‐field box (50 × 50 × 50 cm, length × width × height) at 28 days after t‐MCAO. In the first phase, for 10 min, two of the same objects (A1 and A2) were placed symmetrically from the wall. In a second phase, for 10 min, two dissimilar objects (object A2 was replaced with a novel object B1) were presented in the same box 1 h after the first trial. The amount of time an animal spent exploring each object during two phases was recorded. The objects and the box were cleaned with ethanol (75%) after each individual trial to eliminate olfactory cues. Digital cameras recorded the time of rats to explore the familiar object (TA1) and the novel object (TB1). The discrimination index was calculated according to the following expression: (TA1 − TB1/TA1 + TB1).

BrdU injection and analysis of brain damage

Twenty‐one days after MCA occlusion, animals were treated with twice daily i.p. bolus injections of 5′‐bromo‐2′‐deoxyuridine (BrdU) (50 mg·kg⁻¹; Sigma‐Aldrich, St. Louis, MO, USA) for the identification of proliferating cells. Twenty‐eight days after MCAO, animals were deeply anaesthetized with pentobarbital sodium (50 mg kg^‐1; i.v.) and were killed by decapitation under anaesthesia.

The wet weight of rat brain was measured with an electronic balance. For determination of the cortical width index, whole‐brain images were captured using a digital camera. The width at the midpoint of the forebrain was measured, and the ratio of left width to right width was defined as the cortical width index.

TTC staining and cresyl violet staining

TTC staining (24 h post‐stroke) and cresyl violet staining (8 days post‐stroke) were based on the methods as described previously (Sun et al., 2012; Lin et al., 2013). The areas of infarction were quantified by using Image J (National Institutes of Health, USA). The infarct area in TTC staining in each slice was obtained by subtracting the normal ipsilateral area from that of the contralateral hemisphere to reduce overestimation of infarct area resulting from oedema in the acute stage of ischaemia. Total infarct areas were calculated by summing the infarct areas of the seven slices and were presented as a percentage of the total areas of the contralateral hemisphere. In cresyl violet staining, a series of coronal sections (20 μm thick) from the anterior commissure to the hippocampus were imaged and digitized. Infarct volume was determined by subtracting the area of cresyl violet staining in the ipsilateral hemisphere from that of the contralateral hemisphere and was then multiplied by the section interval thickness.

Brain tissue preparation, immunohistochemistry and immunofluorescence

At 24 h or 28 days after MCA occlusion, animals were transcardially perfused with heparinized saline followed by 4% paraformaldehyde in 0.1 mM PBS. Brains were further fixed in 4% formaldehyde for overnight at 4°C and then transferred into 30% sucrose in PBS for 72 h. The brains were embedded and frozen in optimal cutting temperature compound. A series of brain coronal sections (20μm) were cut every sixth section between bregma levels +0.96 and −0.24 mm (six sections per rat).

Immunohistochemical staining of brain sections was carried out as previously described (Zhang et al., 2016a). For double immunofluorescent staining, frozen sections were incubated for 15 min in 1% Triton X‐100 to disrupt the cell membrane after being air‐dried. Sections were then placed in PBS and blocked with normal horse serum at 37°C for 30 min. Sections were then immersed in anti‐BrdU (1:500; Cell Signaling Technology, TraskLane Danvers, MA, USA), anti‐neuron‐specific nuclear protein (NeuN,1:500; Abcam, Cambridge, UK), anti‐Doublecortin (DCX, 1:500; Cell Signaling Technology), anti‐Nestin (1:500; Novus, Littleton, Colorado, USA) and anti‐adenomatous polyposis coli (APC,1:500; Abcam) antibodies. Brain sections were cover‐slipped and evaluated under a fluorescence microscope (Olympus, Japan). The images of immunohistochemical staining including NeuN⁺, glial fibrillary acidic protein positive (GFAP⁺), 8‐hydroxy‐2‐deoxyguanosine positive (8‐OHdG⁺), 4‐hydroxynonenal positive (4‐HNE⁺) and 3‐nitrotyrosine positive (3‐NT⁺) cells in the peri‐infarct area, and double immunofluorescent staining including BrdU⁺/NeuN⁺, BrdU⁺/DCX⁺ and BrdU⁺/APC⁺ cells in the peri‐infarct area and BrdU⁺/Nestin⁺ cells in ipsilateral subventricular zone were acquired with fluorescence microscope (Supporting Information Figure S2). Immunochemical and immunofluorescent positive cells were counted in six sections per animal. Results were expressed as the average numbers of positive cells in unit area per section of three animal brains.

Cell culture

Primary cortical neurons were prepared from E18 SD rat embryos and were cultured on poly‐D‐lysine (20 μg·mL⁻¹; Sigma‐Aldrich) coated dishes as previously described (Ip et al., 2011). Dissociated neurons were fed with Neurobasal medium (Life Technologies) supplemented with 2% B27 (Life Technologies) and 1% Glutamax (Life Technologies, Waltham, MA, USA). The seeding densities were 1 × 10⁵ mL⁻¹ for neurite outgrowth and differentiation of cortical neurons and Western blotting analysis.

Cell immunocytochemical staining

After 3 days in culture, cortical neurons were incubated with TBN for another 2 days. The primary cortical neurons were fixed in 4% paraformaldehyde at room temperature for 20 min and were then incubated in blocking buffer (10% horse serum, 0.1% Triton X‐100) for 1 h, followed by addition of mouse anti‐βIII tubulin (1:1000; Beyotime Biotechnology, Beijing, China) antibody at 4°C overnight. After washed twice with HBSS, FITC anti‐mouse secondary antibody (1:1000, Sigma‐Aldrich) and DAPI (2 ug·mL⁻¹, Beyotime Biotechnology) were added for 1 h at room temperature. The extended neurites were visualized using a fluorescence microscope (Olympus, Japan).

For analysis of differentiation of cortical neurons, different compounds were treated at cell plating (0 h or 0 days in vitro; DIV), and the neuronal morphology was analysed at 24 and 48 h later. Neurons were fixed with 4% paraformaldehyde and subjected to immunostaining with Tau1 (1:5000; Millipore) and MAP2 (1:5000; Millipore, Billerica, MA, USA) antibodies. After washing twice with HBSS, FITC anti‐rat secondary antibody (1:1000; Sigma‐Aldrich), TRITC anti‐rabbit secondary antibody (1:1000; Thermo Fisher, Waltham, MA, USA) and DAPI (2 ug·mL⁻¹; Beyotime Biotechnology) were added for 1 h at room temperature. Neuronal morphology was photographed using Zeiss LSM510 Meta Confocal Scanning Microscope (Carl Zeiss AG, Germany). At least 60 cells from six randomly selected fields were counted. Neurites with strong Tau1 signals at proximal ends were counted as axons. Neurite length was analysed by ImageJ software.

The quantitative analysis for in vitro cell culture studies was performed by investigators blinded to the drug treatment.

Western blotting

Brain tissue and primary cell samples were harvested and were lysed in RIPA buffer containing protease and phosphatase inhibitors. Generally, equal amounts of protein were subjected to SDS‐PAGE analysis, transferred onto a PVDF membrane and probed with primary antibodies against phospho‐Akt (p‐Akt), total‐Akt (Akt), p‐CREB, CREB, post‐synaptic density protein 95 (PSD95), Synaptophysin, β‐actin (1:1000; obtained from Cell Signaling Technology) or BDNF, vimentin, 2′,3′‐cyclic nucleotide 3′‐phosphodiesterase (CNPase) and myelin marker myelin basic protein (MBP; 1:1000; Abcam). After washing, the membranes were treated with the corresponding HRP‐conjugated secondary antibody. Chemiluminescence detection was carried out with ECL Western Blotting Detection Reagents (Thermo Fisher). Finally, the Carestream MI SE system was used to provide quantitative analysis.

Data and statistical analysis

The data and statistical analysis comply with the recommendations on experimental design and analysis in pharmacology (Curtis et al., 2015). All results were expressed as mean ± SEM, except that NSS results were displayed as median and interquartile range. All statistical analyses were performed with Graphpad Prism software version 6.0 (GraphPad Software, Inc., CA, USA). The animal NSS data were analysed by a non‐parametric Kruskal–Wallis test followed by Dunn's multiple comparisons. All other data were analysed using one‐way or two‐way ANOVA, and statistical differences between groups were analysed using the least significant difference or Bonferroni post hoc test. Significance was set at P < 0.05.

Materials

TBN (Lot number: 20101002; purity: 99.3%) was synthesized by Shanghai Medicilon Inc. (Shanghai, China). ANA‐12 was supplied by Maybridge (Cornwall, UK) and LY294002 was supplied by Sigma‐Aldrich.

Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Southan et al., 2016), and are permanently archived in the Concise Guide to PHARMACOLOGY 2017/18 (Alexander et al., 2017a,b,c).

Results

TBN reduces cerebral infarction and preserves neurological functions at early stage of stroke in rats 24 h and 8 days after ischaemia

Firstly, in the dose–response study on p‐MCAO rats, TBN at 30 and 90 mg·kg⁻¹ given i.v. at 3 h after ischaemia significantly reduced brain infarction in a dose‐dependent manner (Figure 1A). In agreement, TBN treatment notably decreased the NSS 24 h after ischaemia (Figure 1B). In the therapeutic time window study on p‐MCAO rats, treatment with TBN (30 mg·kg⁻¹) at 3 or 6 h, but not at 9 h post‐ischaemia, was effective in reducing brain infarction, as well as in decreasing NSS (Figure 1C, D).

Figure 1.

TBN attenuates brain infarction and neurological deficits in rats subjected to MCAO. (A–D) Effect of TBN on p‐MCAO rat model. (A, B) For dose–response study, TBN or vehicle (0.9% saline) was injected i.v. through the tail vein, at 3 h and then 6 h after MCAO. (C, D) For the therapeutic window study, at 3, 6 or 9 h after MCAO, the first dose (30 mg·kg⁻¹) was given; 3 h thereafter, the second dose was administered. (A, C) Neurological severity score was assessed at 3 and 24 h after MCAO. (B, D) Brain infarct size was examined at 24 h after MCAO by TTC staining. (E–H) Effect of TBN on t‐MCAO rat model. Rats were subjected to 2 h ischaemia and then reperfusion for 8 days. TBN or vehicle was injected i.v. through the tail vein, at 3 h and then 6 h after MCAO on the first day, followed by twice daily treatment at 6 h intervals for another 7 days. Representative serial histological sections stained with cresyl violet at 8 days after MCAO (E) indicated that TBN dose‐dependently reduced brain infarction (F). Neurological severity score (G) and rotarod performance (H) were assessed at 3 h and 8 days after MCAO. Statistical analysis was carried out by. *P < 0.05 significantly different from vehicle treatment group; one‐way ANOVA with Bonferroni post hoc test (A, C, F) and two‐way ANOVA with Bonferroni post hoc test (H) or a non‐parametric Kruskal–Wallis test followed by the Dunn's multiple comparisons test (B, D, G).

Next, in the therapeutic efficacy study on t‐MCAO rats, TBN given at 3 h after ischaemia and then twice daily for another 7 days significantly reduced the cerebral infarction by 22, 31 and 36%, respectively, at the dose of 10, 30 and 90 mg·kg⁻¹ (Figure 1E, F). TBN also decreased the NSS (Figure 1G) and increased the time on rotarod (Figure 1H). There was no sex difference observed with any of the parameters assessed, and treatment with TBN had similar effects on the brain infarction and neurological severity score in male and female rats in both p‐MCAO and t‐MCAO models (Supporting Information Figure S3). Therefore, only male rats were used in therapeutic time window study and long‐term (28 days) post‐stroke recovery study.

In addition, TBN (30 mg·kg⁻¹) treatment significantly sustained the viable neurons (NeuN⁺) 24 h after ischaemia in the cortical peri‐infarct region of p‐MCAO rats, suppressed astrocyte activation (GFAP⁺) and reduced oxidative DNA damage (8‐OHdG⁺), lipid peroxidation (4‐HNE⁺), and protein nitrosylation (3‐NT⁺) (Supporting Information Figure S4).

TBN restores long‐term neurological functions of stroke rats 28 days after ischaemia

To further assess the effect of TBN for a 7‐day treatment regime on long‐term (28 days after ischaemia) neurological function recovery and neural regeneration, we designed the t‐MCAO experiment as illustrated in Figure 2A. A battery of neurological assessments including NSS, adhesive tape removal, rotarod, open field and novel object recognition tests were conducted throughout the 28 day observation period. All rats except those in the sham group exhibited similar substantial neurological deficit 3 h after MCAO, and the neurological function gradually recovered during the 4 weeks after the initial insult. The NSS of rats treated with TBN was significantly lower than that of vehicle‐treated rats at day 7 (Figure 2B). Treatment with TBN also significantly improved the performance in adhesive tape removal, rotarod and open field tests from 14 days post‐ischaemia compared with vehicle treatment (Figure 2C–E). All improvements persisted up to 28 days after stroke. Moreover, compared with vehicle, TBN significantly increased the discrimination index of novel from familiar objects in the novel object recognition test (Figure 2F), reflecting the improvement of recognition memory (Faivre et al., 2012). Accordingly, TBN‐treated rats displayed a better recovery of body weight, compared with vehicle‐treated rats (Figure 2G).

Figure 2.

TBN improves long‐term functional outcomes after t‐MCAO. (A) Experimental workflow: Rats were subjected to t‐MCAO (2 h ischaemia and then reperfusion). TBN (30 mg·kg⁻¹) or vehicle was injected i.v. at 3 and 6 h after MCAO and twice daily for another 7 days and then rats received BrdU (50 mg·kg⁻¹, i.p.) twice daily during days 21–28. The behavioural tests were assessed before MCAO and at 3 h and on day 7, 14, 21 and 28 after MCAO. Stroke rats treated with TBN presented an improvement of NSS (B), sensorimotor function (adhesive tape removal test, C) and motor coordination and balance (rotarod test, D), as well as locomotor activity (open field test, E) compared with vehicle‐treated rats. (F) TBN‐treated rats displayed better performance of recognition memory than vehicle‐treated ones in a novel object recognition test at 4 weeks after MCAO. (G) The change of body weight in each treatment group during 28 days. ^## P < 0.05, significantly different from sham, *P < 0.05, significantly different from vehicle; non‐parametric Kruskal–Wallis test with Dunn's multiple comparisons test (B), two‐way ANOVA (C–E) and one‐way ANOVA (F) with Bonferroni post hoc tests.

TBN promotes post‐ischaemic regenerative neurogenesis in rats

To confirm whether the long‐term therapeutic effects of TBN on ischaemic stroke benefit from the proliferation of newly formed neuronal cells, we assessed the NeuN/BrdU, DCX/BrdU and Nestin/BrdU double labelled, presumably newborn mature neurons, migrating and immature neurons (neuroblasts) (Gleeson et al., 1999) and NPCs (Arvidsson et al., 2002) respectively. In the sham rats of our study, almost no NeuN⁺/BrdU⁺ and DCX⁺/BrdU⁺ cells were found in cortex and striatum, which mainly contain the damaged zone with ischaemic stroke. However, both the numbers of NeuN⁺/BrdU⁺ and DCX⁺/BrdU⁺ cells were markedly increased in the lesion zone of vehicle‐treated stroke rats (Figure 3). Importantly, the double‐positive cells of TBN‐treated rats were significantly higher than that of the vehicle‐treated ones (Figure 3A–E). In the LV zone of the stroke rats, the Nestin⁺/BrdU⁺ NPCs markedly expanded, compared with the sham group. Furthermore, the number of Nestin⁺/BrdU⁺ NPCs cells of TBN‐treated group was significantly higher than that of the vehicle‐treated group (Figure 3C, F).

Figure 3.

TBN enhances neural reconstruction by stimulating neurogenesis after ischaemic stroke. Representative images of ipsilateral hemisphere sections co‐stained with antibodies against (A) NeuN (green, a marker of mature neurons) and BrdU (red, a marker of proliferating cells); (B) DCX (green, a marker of migrating neuroblasts) and BrdU (red); (C) Nestin (green, a marker of NPCs) and BrdU (red) cells. DAPI (blue) indicates nucleus. White dotted lines indicate borders between ischaemia‐affected and unaffected region. Insets show a higher magnification view of double‐positive cells. Scale bar = 200 μm for whole slice, 50 μm for inset magnification. (D–F) Quantitative analysis of newly formed mature neurons (BrdU⁺/NeuN⁺, D), migrating neuroblasts (BrdU⁺/DCX⁺, E) in the damaged region and proliferating NPCs (BrdU⁺/Nestin⁺, F) in the LV on day 28 post t‐MCAO. n = 3 brains per group. (G–I) TBN induced cortical expansion after ischaemic stroke. Representative images of rat brain (G), the cortical width index (H, the width of midpoint of the forebrain hemisphere was measured, and the ratio of left width to right width was defined as the cortical width index) and brain weight (I) of rats at day 28 after ischaemia. n = 8 in sham group; n = 9 in vehicle and TBN treatment groups. ^# P < 0.05, significantly different from sham; * P <0.05, significantly different from vehicle; one‐way ANOVA with Bonferroni post hoc tests.

We also measured cortical width index (Wang et al., 2016) to determine the cerebral cortical expansion. The cortical cavitation, brain weight loss and decrease in cortical‐width index caused by ischaemic injury were significantly attenuated by TBN treatment (Figure 3G, I). These data suggest that an enhanced and substantial neuroregenesis possibly contributes to the accelerated structural and functional recovery after TBN treatment post‐ischaemic stroke.

TBN enhances post‐ischaemic oligodendrogenesis and preserved myelin in rats

Axonal myelin disruption occurs in white matter stroke, accounting for up to 25% of stroke and constitutes the second leading cause of dementia (Iadecola, 2013). Generation of mature oligodendrocytes from oligodendrocyte progenitor cells (OPCs) is crucial for remyelination and functional recovery after brain ischaemia (Sozmen et al., 2016). By double staining with BrdU and APC (also known as CC1), a marker for mature oligodendrocyte cell bodies (Han et al., 2015), we found that TBN treatment dramatically amplified the number of APC⁺/BrdU⁺ cells in the peri‐infarct areas compared with vehicle (Figure 4A, B). Accordingly, TBN treatment time‐dependently augmented the protein expression of immature oligodendrocytes marker CNPase (Saneto and de Vellis, 1985) and MBP (Mi et al., 2009) (Figure 4C–E). Similar to our recent findings in the monkey stroke model, TBN treatment significantly up‐regulated the decreased CNPase expression in the stroke peri‐infarct tissue at 28 days post‐ischaemia (Zhang et al., 2016b).

Figure 4.

TBN stimulates oligodendrocyte proliferation and preserves myelin after ischaemic stroke. (A) Representative images of ipsilateral hemisphere co‐staining of APC⁺ (green; a marker of oligodendrocyte)/BrdU⁺ (red) cells. DAPI (blue) indicates nucleus. Insets show a higher magnification view. Scale bar = 200 μm for whole slice, 50 μm for inset magnification. (B) Quantitative analysis of proliferated oligodendrocytes (BrdU⁺/APC⁺ cells) on day 28 after t‐MCAO. n = 3 brains per group. Representative images of immunoblotting (C) and quantification of CNPase (cyclic nucleotide 3′‐phosphohydrolase, D) and MBP (myelin basic protein, E). n = 4 in each group. ^# P < 0.05, significantly different from sham; *P < 0.05, significantly different from vehicle; one‐way ANOVA (B) and two‐way ANOVA (D, E) with Bonferroni post hoc tests.

TBN promotes post‐ischaemic neuroregenerative repair by activation of Akt/CREB through increasing BDNF expression

BDNF is considered to be a key neurotrophic factor that enhances proliferation of NPCs (Benraiss et al., 2001; Pencea et al., 2001; Schabitz et al., 2007). BDNF promotes NPC migration, differentiation, survival, maturation and behavioural recovery by binding with its specific receptor‐tyrosine receptor kinase B (TrkB) (Song et al., 2015). We wondered if BDNF and its related molecular events participated in the mechanisms underlying the enhanced neuroregeneration by TBN. Western blotting analysis of the peri‐infarct tissue revealed that TBN treatment significantly increased BDNF expression at both 7 and 28 days post‐ischaemia, compared with vehicle (Figure 5A, B). It has been reported that Akt and CREB are downstream targets of BDNF (Massa et al., 2010; Clarkson et al., 2015). We found that at 7 days after ischaemia, CREB was markedly activated by ischaemia compared with sham rats, and TBN treatment more strongly activated CREB than vehicle. At 28 days after ischaemia, there were no notable differences on CREB signal among the three different groups (Figure 5A, C). In contrast, Akt was significantly activated by TBN at both 7 and 28 days after ischaemia (Figure 5A, D).

Figure 5.

TBN activates BDNF/Akt/CREB signalling pathway to provide neuro‐repair in mice. Representative immunoblotting (A, E) and quantification of the relative protein level of BDNF (B), p‐CREB/CREB (C), p‐Akt/Akt (D), synaptophysin (SYP, F), PSD95 (G) and vimentin (H) in the brain infarct region of sham, vehicle or TBN‐treatment rats on days 7 and 28 after t‐MCAo. n = 4 in each group. ^# P < 0.05, significantly different from sham; *P < 0.05, significantly different from vehicle; two‐way ANOVA with least significant difference tests.

Cerebral ischaemic reperfusion can disrupt synaptic transmission and results in learning and memory deficits (Li et al., 2013). BDNF plays an important role in hippocampal synaptic function and plasticity (Patterson et al., 1996). A fundamental question is whether TBN can restore the disrupted neural synaptic function in the post‐ischaemic rat brain. Here, we examined the expression of the major pre‐synaptic scaffold proteins, synaptophysin (SYN) (Greengard et al., 1993) and PSD95 (Feyder et al., 2010). We found that SYN and PSD95 profoundly decreased at both 7 and 28 days post‐ischaemia. Rats treated with TBN exhibited a significantly higher expression of SYN and PSD95 than those treated with vehicle (Figure 5E–G), implying that TBN possibly induced formation of synaptic connectivity between host neurons and NPC‐derived new neurons (Wang et al., 2016). Moreover, the markedly increased vimentin (a prominent marker of microglia and reactive astrocytes) (Jiang et al., 2012) after ischaemia was significantly diminished by TBN (Figure 5E, H).

TBN promotes neuronal differentiation through activation of BDNF and Akt in vitro

Only if the neuroblasts extending from the LV to peri‐infarct regions can differentiate into subtypes of damaged neurons, are they likely to be a reliable source for neuronal replacement after stroke injury. However, it seems that the micro‐environmental signals in damaged area are inadequate for neuronal differentiation (Arvidsson et al., 2002; Parent et al., 2002). To better characterize the effect of TBN on neuronal differentiation and the possible mechanism involved, we conducted in vitro neuronal differentiation on cultured rat embryonic cortical neurons. During neuronal differentiation, the formation of axons and dendrites, neuronal polarization is a prerequisite for neurons to integrate and disseminate signals within the CNS (Witte and Bradke, 2008). Our preclinical pharmacokinetic study demonstrated that after 30 mg·kg⁻¹ TBN injected i.v. in rats, the maximum plasma concentration of TBN can reach 196 μM, and the brain concentration is up to 120 μM at 30 min (data not shown). To examine the development of axonal polarity in vitro, we incubated TBN (100 μM) with dissociated embryonic cortical neurons immediately after plating (0 DIV) and examined neurite morphology by immunostaining of MAP2 and Tau‐1 (somato‐dendritic and axonal markers, respectively) at 24 h (Supporting Information Figure S5) and 48 h (Figure 6). TBN notably promoted neuronal polarization, as increasing the percentage of neurons with polarization from about 40 ± 14% to 55 ± 9% by 1 DIV (Supporting Information Figure S5), and from about 65 ± 7% to 80 ± 10% by 2 DIV (Figure 6A, B).

Figure 6.

TBN promotes neuronal differentiation though BDNF and PI3K signalling pathways in vitro. (A, B) TBN enhances neuronal polarization. Embryonic rat cortical neurons at 0 day in vitro (DIV) were treated with or without TBN in the presence or absence of ANA‐12 (a TrkB antagonist) or LY294002 (LY, a PI3K inhibitor) for 48 h. (A) Representative images of cortical neuron morphology immuno‐stained with the dendritic marker MAP2 (green) and axonal marker Tau‐1 (red). Nuclei were visualized by DAPI staining (blue). (B) Quantitative analysis of neurons at un‐polarization and polarization stages (control, n = 92 neurons; TBN 100 μM, n = 127 neurons; TBN 100 μM + ANA‐12 10 μM, n = 84 neurons; TBN 100 μM + LY 10 μM, n = 104 neurons). (C–E) TBN stimulates neurite outgrowth. Embryonic cortical neurons at 3 DIV culture were treated with or without TBN in the presence or absence of ANA‐12 or LY for 2 days and then immuno‐stained with anti‐βIII‐tubulin (green) and DAPI (blue). (C) Representative images of neurite outgrowth of cortical neurons. Scale bar = 50 μm. (D, E) Quantification of the total length of neurites (D) and number of branches (E). Untreated control, n = 28 neurons; TBN 30 μM, n = 30 neurons; TBN 100 μM, n = 30 neurons; TBN 300 μM, n = 33 neurons; TBN 100 μM + ANA‐12 10 μM, n = 38 neurons; TBN 100 μM + LY294002 10 μM, n = 28 neurons; neural growth factor (NGF) 200 ng·mL⁻¹, n = 25 neurons. Representative immunoblotting (F) and quantification of relative protein expression of BDNF (G), p‐Akt/Akt (H), PSD95 (I) and SYN (J). n = 4 per group. *P < 0.05, significantly different from untreated control (Ctrl). ^# P < 0.05, significantly different from TBN alone group; one‐way ANOVA with Bonferroni post hoc test (B, D, E) or with least significant difference tests (G–J).

Regulation of neurite outgrowth is an important aspect of neuronal plasticity and neuronal regeneration from injuries (Tang, 2001). TBN treatment for 2 days from 3 DIV resulted in a significant growth in the length and branch numbers of β‐III tubulin (an indicator for neurite outgrowth)‐positive neurites (Figure 6C–E). To determine whether BDNF and Akt signals were involved in the effect of TBN on neuronal differentiation, ANA‐12, an antagonist of the BDNF receptor TrkB (Cazorla et al., 2011), and LY294002, an inhibitor of PI3K, were used to treat cortical neurons with the addition of TBN. Interestingly, both ANA‐12 and LY294002 almost completely inhibited neuronal polarization and neurite outgrowth enhanced by TBN (Figure 6A–E). Furthermore, our Western blotting result confirmed that TBN increased BNDF expression and activated Akt, resulting in up‐regulated expression of synaptic scaffold protein PSD95 and SYN (Figure 6F–J). All the up‐regulated protein expression was abolished by ANA‐12 and LY294002, except that ANA‐12 did not significantly affect the expression of BDNF (Figure 6F–J).

Discussion

Previously, we have demonstrated that a single dose of TBN, injected i.p., reduced the brain infarction and ameliorated the neurological deficits 24 h after ischaemia in rats (Sun et al., 2008, 2012). In this report, we systematically investigated the dose–response and therapeutic time window of TBN given i.v. in rat stroke models with severe ischaemic insults, reasoning that efficacy in these models, although it produces more severe strokes than are usually seen in humans, might maximize the chance of ultimate clinical use.

We have shown that TBN is a potent neuroprotective agent, which provides neuroprotection via scavenging free radicals and inhibiting intracellular calcium influx at the early stage of cerebral ischaemia (Sun et al., 2008, 2012). In this report, we focus on the long‐term recovery of neurological function and neuroregeneration post‐ischaemic stroke in rats. TBN accelerated the recovery of sensorimotor behaviour and recognition during the 28 day post‐ischaemia observation period. These results are consistent with our recent findings in the monkey ischaemic stroke model (Zhang et al., 2016b). Newborn neurons play a key role in the long‐term recovery of motor and learning ability after ischaemic stroke (Zhao et al., 2015; Wang et al., 2016). Our recent quantitative proteomics study in the monkey model of ischaemic stroke found that TBN promoted neurogenesis and preserved myelin, which may offer long‐term benefit (Zhang et al., 2016b). Our results suggest that TBN has both neuroprotective and pro‐neurogenic effects, leading to recovery of neurological function at both the early‐stage and long‐term post‐ischaemia. TBN's novel mode of action was summarized in Figure 7. TBN increased NPC proliferation, migration, differentiation and newborn neuron survival and preserved myelin, eventually repairing the neuronal damage after ischaemic stroke. Up‐regulation of Akt/CREB by increasing BDNF expression possibly contributed to the post‐ischaemic neurogenic effect of TBN.

Figure 7.

Proposed mechanisms underlying the action of TBN on promoting post‐ischaemic neuroregenerative repair in cerebral ischaemic environment. TBN enhances endogenous neurogenesis and oligodendrogenesis by increasing NPC proliferation, migration, differentiation and newborn neuron survival and eventually repairs the neuronal damage after ischaemic stroke. Up‐regulation of the BDNF/Akt/CREB signalling loop participates in the post‐ischaemic neurogenic effects of TBN.

The presence of neurogenesis in the DG and SVZ of adult mammalian brain raised the possibility of neural replacement from endogenous NPCs after injury (Jin et al., 2001; Arvidsson et al., 2002; Parent et al., 2002; Zhao et al., 2015). In the post‐ischaemic brain, neurogenesis can be controlled at three major steps – the proliferation, migration and differentiation of NPCs into the neural phenotype destroyed by the ischaemic lesion (Iwai et al., 2002). In the present study, there was a significant increase in the number of Nestin⁺/BrdU⁺ cells in the LV and DCX⁺/BrdU⁺ cells in the peri‐infarct region after TBN treatment, demonstrating that TBN stimulated the proliferation of NPCs and enhanced the formation and migration of neuroblasts. A recent study reports that stroke elicits a latent neurogenic programme in the striatal astrocytes under Notch signalling regulation (Magnusson et al., 2014). Whether TBN also modulates the neurogenic programme in the striatal astrocytes is unknown and warrants further investigation. Many studies have revealed that the survival of newly formed neurons in the ischaemic brain is very low, probably due to an unfavourable environment with lack of trophic support or a toxic environment exposed to severely damaged tissue, and only a small portion can differentiate into mature neurons (Arvidsson et al., 2002; Doeppner et al., 2011). Importantly, in our study, TBN treatment also significantly increased the number of NeuN⁺/BrdU⁺ mature cells in the peri‐infarct region compared with the vehicle treatment. All these results imply that treatment with TBN possibly, on one hand, directly protects newborn neurons and on the other hand modifies the local micro‐environment to be more supportive of the long‐term survival of newborn neurons.

Even with the increased NPC proliferation and subsequent neuroblast migration, only if newly generated neurons are able to differentiate into neuronal subtypes appropriate to the injured region, are they likely to be a useful source for neuronal replacement after injury. Differentiated mature neurons are highly polarized cells that have two main structurally and functionally distinct units: axons and dendrites. In our work, TBN enhanced the axonal polarization and dendrites sprouting in vitro, which was beneficial to TBN's neurogenic potential. Moreover, PSD95 and SYN, two major synaptic and neurotransmission‐regulating proteins (VanGuilder et al., 2010), were up‐regulated by TBN in vitro and in vivo, which further confirmed the ability of TBN to stimulate neuronal differentiation. Despite all these encouraging results, whether these differentiated cells are capable of functionally integrating into the adherent structure and reconnecting to the correct pathways in vivo needs further investigation.

Myelination allows the rapid transfer of information that is required for physiological cognitive and behavioural functions (Deoni et al., 2011). Oligodendrocytes are the myelin‐forming glial cells in the CNS. After brain ischaemic injury, NPCs rapidly proliferate and transform into OPCs and then migrate to fill the demyelinated area, differentiate into mature oligodendrocytes and reinstate myelin sheaths (Gallo and Armstrong, 2008). In our study, treatment with TBN after stroke stimulated the proliferation of APC⁺ cells and increased the expression of CNPase and MBP, a marker of immature oligodendrocytes and myelin respectively. These results suggest that TBN stimulated oligodendrogenesis and preserved myelin after stroke. These effects may provide therapeutic benefit for treating multiple sclerosis and other demyelinating disorders (Najm et al., 2015).

Mechanistic studies have revealed that TBN treatment increased the expression of BDNF after cerebral ischaemia. BDNF is the most abundant neurotrophin in the brain and is known to play a critical role in protecting brain tissue from injury and in promoting neuronal plasticity and neurogenesis (Greenberg et al., 2009). Our current findings are consistent with studies by others, showing that increasing cerebral BDNF level exerts direct effects on promoting neurogenesis, oligodendrogenesis and myelination (Benraiss et al., 2001; Pencea et al., 2001; Kobayashi et al., 2006). However, most of those studies raised BDNF levels using a recombinant viral approach or intracerebral administration (Benraiss et al., 2001; Pencea et al., 2001; Kobayashi et al., 2006), which clearly has its limits in practical application. Our current study showed that TBN is a potent agent capable of stimulating neurogenesis after ischaemic stroke by increasing BDNF expression. As our previous data demonstrate that TBN readily penetrated the BBB in the rat and monkey (Sun et al., 2012; Zhang et al., 2016b); systemic TBN would readily be able to modulate BDNF levels in the brain.

Previous studies have shown that BDNF can increase both p‐Akt and p‐CREB (Massa et al., 2010; Clarkson et al., 2015). The PI3K/Akt signalling pathway plays a central role in regulating cell growth, proliferation and survival and is also involved in axonal sprouting, which is an important mechanism underlying post‐stroke functional recovery (Li et al., 2010). Moreover, disruption of BDNF signalling leads to a decrease in CREB transcription, which is critical for processes of synaptic plasticity and learning (Mowla et al., 1999). We observed an increase in BDNF, p‐Akt and p‐CREB expression after TBN treatment and a TrkB inhibitor ANA‐12 and a PI3K inhibitor LY‐294002 significantly blocked the neural differentiation effect of TBN in vitro, which would highlight the BDNF/Akt/CREB signalling pathway as being crucial for mediating functional recovery after TBN treatment in stoke. Although many studies identified that Akt and CREB are downstream targets of BDNF (Massa et al., 2010; Clarkson et al., 2015), a recent study reported that CREB was phosphorylated by Akt, leading to CREB‐mediated expression of genes crucial for neuronal survival, including BDNF (Lu et al., 2013). Our reversal study found that a TrkB inhibitor ANA‐12 could down‐regulate p‐Akt expression, whereas a PI3K inhibitor could also down‐regulate BDNF expression. Combining our findings with those of others, it would appear that BDNF, Akt and CREB are involved in a regulatory loop. A number of studies have shown that BDNF is released by neurons and is mainly secreted via dendritic release (Brigadski et al., 2005; Matsuda et al., 2009). Our in vitro data also demonstrated that TBN could increase BDNF expression in primary cortical neuron cultures. However, how TBN increases BDNF expression and the direct target of TBN therapy is still unclear and needs to be further explored.

In summary, we have demonstrated that TBN promoted neuronal regeneration after ischaemic stroke by up‐regulation of Akt/CREB via increasing BDNF expression. The ability of TBN to simultaneously provide neuroprotection and promote neurogenesis/oligodendrogenesis is likely to make it more effective than a single functional compound. TBN is thus a promising new treatment for both acute phase neuroprotection and long‐term functional recovery after ischaemic stroke.

Author contributions

Y.Q.W. and Z.J.Z. designed and supervised the project, G.X.Z., T.Z., Y.W.S., C.M.L., Y.H.T., L.C.S. and G.Y. performed the stroke experiments, collected and analysed the data. N.L., L.M.W., W.L., C. Z., P.Y. and X.F.Y. performed the cell culture experiments, collected and analysed the data. Z.J.Z., T.Z. and Y.Q.W wrote the manuscript. All authors have read and approved the final manuscript.

Conflict of interest

The authors declare no conflicts of interest.

Declaration of transparency and scientific rigour

This Declaration acknowledges that this paper adheres to the principles for transparent reporting and scientific rigour of preclinical research recommended by funding agencies, publishers and other organisations engaged with supporting research.

Supporting information

Figure S1 Changes of cerebral blood flow in MCAO. The changes of cerebral blood flow monitored by a laser Doppler blood flow meter during ischemia and reperfusion in rats. The results were expressed as means±SEM. n = 8 in sham group; n = in vehicle group; n = 9 in TBN group. ^### P<0.001 versus sham group, ***P<0.001 versus vehicle group. Statistical analysis was carried out by two‐way ANOVA with Bonferroni post hoc tests.

Figure S2 Representative images of brain coronal section stained with hematoxylin‐eosin. The observed area of double‐immunofluorescent staining was outlined by the red dotted line (up panel), and the peri‐infarct region was magnified and indicated by a black dotted line (low panel). Scale bars: 2 mm in the up panel; 1 mm in the low panel magnified images. LV: lateral ventricle.

Figure S3 TBN attenuates brain infarction and neurological severity score in rats subjected to MCAo. (A, B) Effects of TBN on male rats of p‐MCAO. (C, D) Effects of TBN on female rats of p‐MCAO. (E, F) Effects of TBN on male rats of t‐MCAO. (G, H) Effects of TBN on female rats of t‐MCAO. All data were expressed as mean±s.e.m.

Figure S4 TBN preserves neuronal viability, suppresses astrocyte activation and reduces oxidative damage in the peri‐infarct area at 24 h after p‐MCAO. (A) Representative images of the peri‐infarct area by immunohistochemistry with antibodies against NeuN (a marker of mature neuron), GFAP (a marker of astrocyte), 8‐OHdG (a marker of DNA oxidation), 4‐HNE (a marker of lipid peroxidation) and 3‐NT (a maker of protein tyrosine nitrosylation). Scales bar=50 mm. (B‐F) Statistical analyses of NeuN, GFAP, 8‐OHdG, 4‐HNE and 3‐NT positive cells. Results were calculated as ratio to sham or vehicle group and expressed as mean±s.e.m. n = 4 in sham group. n = 4 in vehicle group. n = 7 in TBN group. Statistical analysis was carried out by One‐way ANOVA with Bonferroni post hoc test. ^## P < 0.01 vs. sham group. *P < 0.05 and *P < 0.01 vs. model group.

Figure S5 Neuronal polarization promoted by 24 h TBN treatment. The morphology (A) and quantitative analysis (B) of neuronal polarity in primary cortical cultures after drug treatment for 24 h. (Ctrl, n=32 neurons; TBN 100 μM, n= 127 neurons; TBN 100 μM + ANA‐12 10 μM, n=64 neurons; TBN 100 μM + LY294002 10 μM, n=104 neurons). Scales bar=50 μm. The results were expressed as means ± s.e.m. #P <0.05 and versus untreated control (Ctrl); ***P<0.001 versus TBN treatment. Statistical analysis was carried out by One‐way ANOVA with Bonferroni post hoc tests.

Table S1 Changes of blood gas parameters.

Table S2 The total number of rats in each group, as well as the number of rats included and excluded in each group.

Zhang, G. , Zhang, T. , Li, N. , Wu, L. , Gu, J. , Li, C. , Zhao, C. , Liu, W. , Shan, L. , Yu, P. , Yang, X. , Tang, Y. , Yang, G.‐Y. , Wang, Y. , Sun, Y. , and Zhang, Z. (2018) Tetramethylpyrazine nitrone activates the BDNF/Akt/CREB pathway to promote post‐ischaemic neuroregeneration and recovery of neurological functions in rats. British Journal of Pharmacology, 175: 517–531. doi: 10.1111/bph.14102.