Background
Stroke, which can be hemorrhagic or ischemic, is caused by the bursting or blockade of blood vessels, respectively. It is one of the leading causes of death and disability worldwide, and 85 % of all stroke cases are ischemic [
1‐
3]. Recombinant tissue plasminogen activator (rtPA), an enzyme that breaks down blood clot aiming to restore blood reperfusion, is the only FDA-approved pharmacological treatment for ischemic stroke [
1‐
3]. However, its application is restricted because of a narrow administration time window of ~3 to 4.5 h after the initiation of ischemia [
4]. Hence, there is an urgent need for novel stroke treatments, and different neuroprotective strategies that counteract the molecular cascades responsible for ischemic injury have been proposed [
3,
5‐
8]. Mechanisms that contribute to ischemic injury include glutamate excitotoxicity, oxidative damage, inflammation and apoptosis; molecular targets for neuroprotection include calcium channel, glutamate receptor, GABA
A receptor, PSD95, nitric oxide synthase and extracellular adhesion molecule [
3,
5‐
8]. In particular, one neuroprotective strategy targets the N-methyl-
d-aspartate receptor (NMDAR) which plays a critical role in mediating glutamate excitotoxicity and neuronal death [
9‐
11]. Interestingly, NMDAR also mediates pro-survival signaling that is beneficial to neuronal survival upon cerebral ischemic attack, via Akt and ERK activation [
7]. Furthermore, NMDAR-mediated ERK activation in turn stimulates CREB, a transcriptional factor which regulates a number of genes critical for long-term neuroprotective effect [
12]. A prevalent view has emerged to explain the two opposing effects – excitotoxicity and pro-survival – on neurons upon NMDAR activation: they are mediated by different NMDAR subtypes containing GluN2B and GluN2A, respectively [
7].
Many neuroprotective compounds have been demonstrated to be effective in animal models of cerebral ischemia, and some candidates are being actively pursued in clinical trials (e.g. clinical trials with identifiers NCT02315443, NCT02222714, NCT01454154, NCT01221246, NCT01502761, NCT02144584 and NCT02535611), including four that mitigate the excitotoxicity mediated through NMDAR (NCT02315443, NCT01502761, NCT02144584 and NCT02535611), by blocking NMDAR channel current or channel interaction with PSD95 [
6,
8,
11,
13,
14]. It is known that some traditional Chinese medicine (TCM) preparations exert neuroprotection with few or no adverse effects, and some can inhibit NMDAR current [
15,
16].
According to traditional Chinese medicine, stroke is caused by
Qi deficiency and blood stasis [
17,
18]. For centuries, various multi-herbal formulas have been used by Chinese medicine practitioners to treat various diseases [
19‐
21]. Based on the theory of TCM, together with modern pharmacological knowledge of herbal medicines, a neuroprotective formula called Post-Stroke Rehabilitation (PSR), comprising seven herbs (Table
1) has been designed. Most chosen herbs have been used in herbal mixtures for stroke treatment, including Buyang Huanwu, Sheng Yu, Xiaoxuming, Gualou Guizhi and Wen Dan Decoctions as well as BNG-1, Hua Tan Tong Luo formula and Nao-Shuan-Tong [
22‐
30]. Moreover, extracts and/or isolates from all constituent herbs have been demonstrated to be neuroprotective in cell and/or animal models. For examples, extracts from
C. obtusifolia and
G. uralensis could protect cultured neurons against experimental induced-neurotoxicity and
P. lactiflora extract could induce proliferation of injured neurons or cultured Schwann cells [
31‐
35]; active molecules such as calycosin isolated from
A. membranaceus, tanshinones from
S. miltiorrhiza, chrysophanol from
C. obtusifolia, and ferulic acid and ligustilide from
L. chuanxiong and
A. sinensis were neuroprotective in rodent ischemic models [
36‐
41]. Four of the selected herbal components –
A. membranaceus,
P. lactiflora,
L. chuanxiong and
A. sinensis – are among the most commonly used herbs for stroke treatment [
42]. In addition to the prominent neuroprotective properties of
S. miltiorrhiza,
G. uralensis and
C. obtusifolia (Table
1), they were included in PSR because
S. miltiorrhiza, can promote blood circulation and remove blood stasis [
43];
G. uralensis can harmonize the actions of multiple herbs [
43];
C. obtusifolia has been shown to reduce hyperlipidemia [
44,
45], a known risk factor for stroke [
45]. The amount of each herb used in the PSR mixture is in accordance with the recommendation stated in the Chinese Pharmacopoeia [
43].
Table 1
Herbal components of PSR and their implications for stroke treatment
1. Astragalus membranaceus (Fisch.) Bunge (Huangqi) | A component of Buyang Huanwu Decoction b (BYHWD), Sheng Yu Decoction and BNG-1 [ 22, 23, 27]; used with Honghua to treat stroke patients [ 17]; its isolates, calycosin and formononetin, were neuroprotective in cerebral ischemic rats [ 40, 81] |
2. Salvia miltiorrhiza Bunge (Danshen) | A component of Hua Tan Tong Luo formula and BNG-1 [ 27‐ 29]; polysaccharide extract and active isolates, lithospermic acid, salvianolic acid B, cryptotanshinone, tanshinone I and tanshinones IIA-B, could reduce damages caused by cerebral ischemia in rodents [ 41, 82‐ 85] |
3. Paeonia lactiflora Pall. (Chishao) | A component of BYHWD [ 22] and Nao-Shuan-Tong [ 30]; showed proliferative effects on injured rat peripheral neurons and Schwann cells [ 31, 32] ; an isolate, paeoniflorin, was neuroprotective in cerebral ischemic rats [ 86] |
4. Cassia obtusifolia L. (Juemingzi) | Used for hyperlipidemia treatment [ 44]; showed neuroprotective effects in hippocampal neurons and in cell and animal models of Parkinson’s disease in mice [ 33, 34] ; an isolate, chrysophanol, showed neuroprotective effect in ischemic mice [ 87] |
5. Ligusticum chuanxiong Hort. (Chuanxiong) | A component of BYHW, Sheng Yu, and Xiaoxuming Decoctions [ 22‐ 24]; butylidenephthalide, ferulic acid and ligustilide isolated from Chaunxiong were neuroprotective in rodent cells and ischemic rats [ 36‐ 38, 88] |
6. Angelica sinensis (Oliv.) Diels (Danggui) | A component of BYHWD and Sheng Yu Decoction [ 22, 23]; ferulic acid and ligustilide isolated from Danggui showed neuroprotection in rodent cells and ischemic rats [ 36‐ 38] |
7. Glycyrrhiza uralensis Fisch. (Gancao) | A component of BNG-1, Xiaoxuming, Gualou Guizhi, and Wen Dan Decoctions [ 24‐ 27]; used to treat diabetic acute ischemic stroke [ 89]; liquiritin, a major constituent, enhanced the neurotrophic effect of NGF [ 90] , showed protection against Aβ-induced toxicity in rat cortical neurons and ischemic mice [ 35, 91] |
To further understand if PSR is an effective treatment for stroke, preclinical and clinical studies are required. In this study, we aim to understand the action mechanism of PSR using in vitro primary embryonic neuron cultures, which express predominately GluN2B-containing NMDA receptors [
46], and in vivo using the rat middle cerebral artery occlusion (MCAO) model. Our study indicates that PSR can protect against NMDA-induced excitotoxicity, inhibit NMDA-induced current and activate the pro-survival ERK/CREB signaling in cultured primary embryonic neurons. Furthermore, PSR is neuroprotective and attenuates neurological impairments in MCAO rats when administered either before or after the onset of ischemia. The present findings offer insights into the underlying multi-targeted neuroprotective mechanism of PSR and provide a scientific basis for its further pharmacological development for managing stroke and other neurological diseases.
Methods
Chemicals and reagents
Unless otherwise specified, all chemicals, antibodies, and culture reagents were purchased from Sigma (MO, USA), Cell Signaling Technology (MA, USA), and Life Technologies (CA, USA), respectively.
Animals and the MCAO model
Sprague–Dawley (SD) rats were obtained and reared from the Animal and Plant Care Facility of The Hong Kong University of Science and Technology. Rats were provided with food and water ad libitum under a 12/12 h light/dark cycle. Transient focal cerebral ischemia was induced by intra-luminal MCAO as described previously [
47,
48]. Male SD rats (280–320 g) were anesthetized and placed on heating pad to maintain rectal temperature at 37 °C. A nylon filament was advanced from the right external carotid artery into the lumen of the internal carotid artery to occlude the right MCA for 2 h, and the filament was subsequently withdrawn to allow reperfusion. The cerebral blood flow of the rats was monitored closely with Laser Doppler (PF 5010 – LDPM unit; PERIMED) during the operation. Sham-treated rats were subjected to the same operative procedure without the insertion of the nylon filament. The right femoral artery was cannulated to monitor arterial blood gases, pH, electrolytes, and hematocrit. All procedures were approved by the Animal Ethics Committee of The Hong Kong University of Science and Technology.
Plant collection and identification
PSR comprises
A. membranaceus,
S. miltiorrhiza,
P. lactiflora,
C. obtusifolia,
L. chuanxiong,
A. sinensis, and
G. uralensis (Table
1). All herbs were purchased from Lee Hoong Kee Limited, Hong Kong, and collected from different regions of China:
A. membranaceus (roots; voucher ID: TCM208-4) and
G. uralensis (roots and rhizomes; TCM080-4) from Inner Mongolia Autonomous Region;
S. miltiorrhiza (roots and rhizomes; TCM012-4),
P. lactiflora (roots; TCM037-3) and
L. chuanxiong (rhizomes; TCM111-3) from Sichuan;
C. obtusifolia (seeds; TCM238-2) from Anhui;
A. sinensis (roots; TCM800-2) from Gansu. All herbs were authenticated by one of the authors (Dr. Guang-Miao Fu, a pharmacognosist) and a voucher specimen of each herb with identification number indicated above was deposited in the herbarium of the Biotechnology Research Institute at the Hong Kong University of Science and Technology.
PSR preparation and HPLC fingerprinting
The PSR formula was prepared by the National Engineering Research Center for Modernization of Traditional Chinese Medicine, Zhuhai, China. The amount of each herb used in the mixture is in accordance with the recommendation stated in the Chinese Pharmacopoeia [
43]. Briefly, mixture containing
A. membranaceus (15 g),
S. miltiorrhiza (10 g),
P. lactiflora (7.5 g),
C. obtusifolia (7.5 g),
L. chuanxiong (6 g),
A. sinensis (6 g),
G. uralensis (6 g) were refluxed with water (w/v 1:8) for 1 h. The extracts collected were cool-dried by vacuum into a brown powder. Two batches of extract were prepared and the yields were ~18 %, respectively. The powder was used for in vitro and in vivo studies. An HPLC-diode array detection system (Waters) was used for fingerprint analysis. Each filtered sample (20 μL) was separated on a C18 column with a mixture of acetonitrile and 0.1 % trifluoroacetic acid in water at room temperature; a gradient elution of acetonitrile from 10–90 % was applied. Eluents were monitored by the absorbance at 254 nm.
PSR administration
PSR dissolved in water was orally administered to rats before and after MCAO. The human-equivalent dose (HED) in rats was calculated to be 1.1 g/kg. Rats were dosed with 1.1, 2.2 and 5.5 g/kg and the dosing volume was 5.5 mL/kg. For the pre-treatment regimen, PSR was administered daily for 5 days before and 6, 24, and 48 h after MCAO onset. For the post-treatment regimen, PSR was administered 6, 24 h, and then daily after MCAO onset. Deionized water was administered as control.
Evaluation of neurological deficit
Bederson’s neurological examination scoring system [
49,
50] was used to evaluate neurological deficit. Rats were scored 0 if they had no observable deficit (normal); 1 if they failed to fully extend the left forepaw (mild deficit); 2 if they circled towards the contralateral side (moderate deficit); and 3 if they lost the ability to walk and righting reflexes (severe deficit).
Measurement of cerebral infarction and edema extent
After the final evaluation of neurological deficit, rats were sacrificed and their brains were removed, fixed with 4 % paraformaldehyde, sectioned into seven 2-mm coronal slices, and stained with 2 % (w/v) 2,3,5-triphenyltetrazolium chloride (TTC) for 30 min at 37 °C [
49]. Images of the slices were taken and analyzed using ImageJ System software (National Institutes of Health, Bethesda, MD). The infarct area in each slice was determined by subtracting the TTC-stained area in the ipsilateral (i.e., ischemic) hemisphere from that of the contralateral (i.e., non-ischemic) hemisphere and expressed as the percentage of the TTC-stained area of the contralateral hemisphere. Infarct volume was calculated on the basis of the infarct area and slice thickness. The ischemia-induced gain in hemispheric volume (i.e., edema) was calculated by subtracting the volume of the contralateral hemisphere from that of the ipsilateral hemisphere and expressed as the percentage of the volume of the contralateral hemisphere.
Primary neuronal cultures
Embryonic cortical and hippocampal neurons express high levels of GluN2B-containing NMDA receptors and therefore were used for this study [
46]. They were prepared from embryonic day 18 (E18) SD rats as described previously [
51]. Cortical neurons were cultured on poly-
d-lysine–coated plates and maintained in neurobasal medium supplemented with 2 % B-27 supplement, penicillin (100 units/mL), and streptomycin (100 μg/mL) at 37 °C in a humidified atmosphere with 5 % CO
2. Hippocampal neurons were grown similarly, except the medium was supplemented with 0.5 mM L-glutamine.
NMDA excitotoxicity assay and immunohistochemistry
Since a larger quantity of cortical neurons can be obtained from one embryo compared to hippocampal neurons, cortical neurons were used for NMDA excitotoxicity assay. Cortical neurons at 11–12 days in vitro (DIV) were pre-treated with PSR or vehicle (DMSO) for 2 h and subsequently rinsed with Locke’s solution (5 mM KCl, 128 mM NaCl, 2.7 mM CaCl
2, 1 mM Na
2HPO
4, 5 mM HEPES, and 10 mM glucose) followed by a 15-min incubation with glycine-containing (10 μM) Locke’s solution. The neurons were then co-treated with PSR (or vehicle) and 20 μM NMDA dissolved in Locke’s plus glycine solution for 20 min. The neurons were then incubated with fresh growth medium. After 24 h, NMDA excitotoxicity was assessed by either the lactate dehydrogenase (LDH) release assay (Roche) or immunohistochemistry, as described previously [
52]. The cell death in the presence of PSR and NMDA was quantified by normalizing the LDH release to the maximum (vehicle with NMDA, the control) and minimum (no NMDA) LDH releases. Neurons were immunostained with β-tubulin type III antibody (1:1000, Sigma) and labeled with FITC-conjugated secondary antibody. DAPI (4′, 6-diamidino-2-phenylindole) was used to stain the nuclei. Images were captured and visualized by fluorescent microscopy (Leica).
Electrophysiology
Embryonic hippocampal neurons express larger NMDAR current compared to cortical neurons and were used for patch clamp recording of the NMDAR current. Whole-cell recording was performed at room temperature (~22 °C) as previously described [
51,
53]. The pipette solution contained (in mM): 135 CsCl, 0.5 EGTA, 2 MgCl
2, 4 Na
2ATP, 0.4 NaGTP, and 10 HEPES (pH 7.25). The bath solution contained (in mM): 5.3 KCl, 0.4 KH
2PO
4, 4.2 NaHCO
3, 138 NaCl, 0.3 Na
2HPO
4, and 27.8 glucose, 2.5 CaCl
2, 0.01 glycine, 0.0005 tetrodotoxin, 0.02 bicuculline methiodide (Tocris), and 5 HEPES (pH 7.4). NMDA alone (20 μM) or with PSR were added to the bath solution. Pipettes were fire-polished, and resistance was 3–5 MΩ. Currents were recorded at −70 mV using a Multiclamp 700B amplifier and data were analyzed using pClamp 10 (Molecular Devices, USA).
Western blotting
After treatment, cortical neurons were lysed with RIPA buffer at different time points. Western blot analysis was performed as described previously [
54]. Briefly, protein samples (30 μg) were separated by SDS-PAGE and then transferred to nitrocellulose membrane, which was subsequently incubated with antibodies specific to phospho-p44/42 mitogen-activated protein kinase (p-ERK) (1:1000), p44/42 mitogen-activated protein kinase (ERK) (1:1000), phospho-cAMP response element-binding protein (p-CREB) (1:1000), or CREB (1:1000). The membranes were then washed and incubated with horseradish peroxide-conjugated secondary antibody (1:2000). Signals were detected using the SuperSignal West Pico chemiluminescent substrate kit (Pierce, Rockford, IL, USA). Band densitometry was measured and analyzed by ImageJ.
Luciferase reporter assay
Freshly isolated cortical neurons were transiently transfected with the luciferase reporter construct (CRE-Luc) [
55] using Lipofectamine 2000. Transfected neurons (9–11 DIV) were treated with PSR or vehicle (DMSO) for 6 h and then lysed with luciferase lysis buffer. Lysates were transferred onto white-bottom plates and assessed using a luciferase assay kit (Promega). Luciferase (i.e., CRE) activity was quantified using a Promega GloMax™ 96 microplate luminometer.
Statistical analysis
Results are reported as mean ± SEM (or SD if number of experiments is 2). Data with two dependent variables (e.g. drug dose and time/coronal slice) were analyzed by two-way ANOVA followed by Bonferroni post-test; data with one dependent variable were analyzed by one-way ANOVA followed by Newman-Keuls post-test, except when stated otherwise. The level of significance was indicated by *, **, and *** which corresponds to P < 0.05, <0.01, and <0.0,01, respectively.
Discussion
PSR is a novel neuroprotective formula designed for stroke treatment. Its constituent herbs have all been demonstrated with prominent neuroprotective properties. The further development of PSR for stroke treatment relies on both preclinical and clinical studies. The work reported here is a pioneer preclinical study of PSR demonstrating its neuroprotective efficacy. We first demonstrated that PSR extract could prominently attenuate cell death using cultured NMDA challenged-neurons. This in vitro finding validate the neuroprotective character of the selected herbal components of PSR. Furthermore, it indicate that the net activity of all the neuroprotective compounds within the PSR extract persists despite possible neutralizing interactions that can occur with the many other molecules presence in the extract. Support by the in vitro data, PSR was tested in vivo using MCAO rats, and demonstrated to be effective in reducing cerebral injury and improving neural behavior. Further in vitro studies on the effects of PSR on NMDAR currents and ERK and CREB activation shed insights into its mechanism of action.
Our in vivo study demonstrates that PSR pre- and even post-treatments can effectively decrease ischemic infarct size and reduce functional neurological impairment in MCAO rats. The effectiveness of pre-treatment suggests that PSR may be developed for the prophylactic management of patients with a high risk of stroke recurrence [
61]. Two aspects of the present post-treatment data are particularly outstanding. First, PSR acts quickly to mediate neuroprotection and elicits sustained improvement in neurological function in MCAO rats. Improvements in neurological scores were observed as early as 18 h after the first dose of PSR and extended up to day 14. Second, the therapeutic time window of PSR in rats is at least 6 h after the onset of MCAO, which is longer than many other tested compounds/drugs in similar animal models. As patients with stroke onset longer than 3 h before therapeutic intervention can be practically applied is common, the 6 h therapeutic time window of PSR in MCAO rats increases the chance of PSR to be successfully translated to human stroke patients [
14,
62].
Our in vitro finding that PSR can reversibly inhibit NMDA-induced currents with a fast off-rate provides one possible mechanism underlying the neuroprotective effect of PSR. A fast off-rate is a desirable pharmacological property of an NMDAR current inhibitor critical for safety [
11,
53], and the fast off-rate of PSR is consistent with the long safety history of the herbal components of PSR to treat humans. In contrast to memantine, another NMDAR blocker with fast off-rate failed to protect against ischemic impairment in rat when administered 1 h after the onset of MCAO [
47], the superior therapeutic effect of PSR when administered 6 h after ischemia suggests that mechanisms other than NMDAR current inhibition is involved in mediating its potent neuroprotective effect.
Increasing evidence shows that single-targeted drug candidates are ineffective for stroke treatment [
63]. As TCM formulas contain multiple herbs designed and assumed to act synergistically [
19‐
21], their pharmacological actions are usually multi-targeted. PSR indeed has multiple targets – in addition to inhibiting NMDAR-mediated current, PSR can activate ERK and CREB, which may potentiate the expression of CREB-mediated genes such as growth factors and anti-apoptotic proteins to exert neuroprotective effects [
64‐
67]. Two of the constituent herbs,
A. sinensis and
L. chuanxiong, have been demonstrated or implicated in activating CREB – pretreatment of cortical neuron culture with extract of
A. sinensis prevented the reduction of phospho-CREB induced by amyloid beta peptide, and extract of
L. chuanxiong induced CREB phosphorylation in PC12 cells [
68,
69]. However, to the best of our knowledge, no decoction has been reported to enhance ERK and CREB signaling in cultured primary neurons. Only until recently, a decoction of thirteen herbs to treat memory impairment of schizophrenia patient was shown to activate ERK and CREB in hippocampi of treated animals [
70]. Moreover, PSR likely acts on other targets in addition to NMDAR and ERK/CREB, as all its seven constituent herbs except
C. obtusifolia have been implicated to have both anti-oxidative and anti-inflammatory effects, while
C. obtusifolia has been implicated to have only anti-inflammatory effect [
71‐
77]. In future experiments, it will be important to demonstrate the anti-oxidative and anti-inflammatory effects of PSR extract, and explore other plausible mechanisms that can contribute to its neuroprotective effect.
There are two possible directions regarding the further development of PSR. First, the PSR extract, although a complex molecular mixture, can be further investigated and developed for stroke treatment. Such effort requires the preparation of a quality-controlled extract from the seven constituent herbs, which in turn relies on standardized good practice to ensure the use of high quality herbs, a consistent procedure for extract preparation and chemical profiling to validate the quality the extract [
78,
79]. Alternatively, the extract can be simplified by systematically purifying molecular fractions from the seven herbs, identifying the active molecules, and mixing the active molecules in optimized molar ratio to generate a cocktail which can achieve the best neuroprotective effects. The optimized molar ratio of the active molecules can be derived by using computational algorithm coupled to bioassays [
80].
Acknowledgements
We thank Prof. M.E. Greenberg (Harvard Medical School) for the luciferase reporter gene CRE-luciferase construct; Venus W.S. Lee and Yuki Y.K. Tang for their excellent technical assistance; Dr. Y.P. Ng and Dr. Ho-Kee Koon for critical comments on the manuscript.