The effect and mechanism of combination of total paeony glycosides and total ligustici phenolic acids against focal cerebral ischemia

The root of Paeonia lactiflora Pall. (Chishao, CS) and Ligusticum chuanxiong Hort. (Chuanxiong, CX) were widely used as a drug pair in Chinese Medicine, and the combination of CS and CX showed a more significant inhibition on neuronal apoptosis in our previous study. In the present study, total paeony glycosides (TPGs) from CS and total ligustici phenolic acids (TLPAs) from CX were combined to evaluate the synergistic effects against focal cerebral ischemia both in vitro and in vivo. The combination of TPGs and TLPAs at 7:3 had the best anti-oxidative stress and anti-inflammatory effect on OGD-induced HUVEC. Additionally, the infarction area proportion and neuron apoptosis of rats by TPGs:TLPAs (7:3) was significantly lower than their alone in MCAO rats. Moreover, TPGs: TLPAs of 7:3 showed a more significant effect on decreasing the expression of MMP-2 and MMP-9, and increasing the protein expression or mRNA level of TIMP-1 than other combinations. The optimal ratio of TPGs and TLPAs at 7:3 could bring more remarkable protective effects against focal cerebral ischemia in MCAO rats by alleviating oxidative stress, inflammatory and neuronal apoptosis to protect the blood-brain barrier. Overall, the present study provided benefical evidence for clinical application of CS and CX as a “drug pair”.

The effect of TPGs and TLPAs at the best optimal ratio on MCAO rats. Improvement of TPGs:TLPAs at 7:3 on neurological deficit score of MCAO rats. As the results showed in Table 4, at 24 h after cerebral ischemia, the neurological deficit score of the sham-operated rats was 0, indicating that the sham-operated rats did not have neurological impairment symptoms. Compared with the sham-operated group, the neurological deficit scores in the model group were significantly increased. (P < 0.01). After administration, the score of neurological deficit in rats was significantly lower (P < 0.01, P < 0.05), and the changes were with the prolongation of administration time. TPGs:TLPAs at the ratio of 7:3 showed the best protective effect on the neurological deficit. Improvement of TPGs:TLPAs at 7:3 on infarction area of MCAO rats. As shown in Fig. 1A, infarction areas were shown in white. There was no infarction in the sham operation group, while infarction occurred in each group after the model was established. The cerebral infarct size ratio of the model group was reached 49.72%, which was significantly different from the sham operation group (P < 0.01). Compared with the model group, the infarct size of cerebral ischemia rats was significantly reduced after administration (P < 0.01, Fig. 1B). When TPGs:TLPAs at the ratio of 7:3, the treatment of infarct size in cerebral ischemia rats was most effective, showed a significantly lower in infarction areas than other combinations (P < 0.01, P < 0.05). There was no difference in cerebral infarct size ratio between TPGs:TLPAs treatment at 5:5 and TPG treatment alone, but the effect was better than that of the TLPAs alone group (P < 0.05).
Anti-oxidative stress activity of TPGs:TLPAs at 7:3 in MCAO rats. Anti-oxidative stress activity of TPGs:TLPAs combinations in MCAO rats was presented in Table 5. Comparing with sham control group, the activities of SOD, CAT and GSH-Px of model rats were significantly decreased (P < 0.01, P < 0.05), while the levels of MDA, and LPO were significantly increased (P < 0.01, P < 0.05). The activities of SOD, CAT, GSH-Px, MDA, and LPO  Table 3. Effect of XS on TNF-α, IL-6, sICAM-1expression in OGD-induced HUVEC (x ± SD, n = 3). Notice: ## P < 0.01 or # P < 0.05, compared with control group; **P < 0.01 or *P < 0.05, compared with model group. www.nature.com/scientificreports www.nature.com/scientificreports/ Effects of TPGs:TLPAs at 7:3 on functional recovery and anti-apoptosis of MCAO rats. Clear and integral tissue structure was observed in sham group with clear neuronal nuclei, relatively intact nuclear membrane and normal glial cells and capillary morphogenesis (Fig. 2C). However, in model group, edema of the neuropile were observed in model group, and most neurons in the damage areas appeared shrunken with eosinophilic cytoplasm and triangulated pyknotic nuclei. Histopathological abnormalities were significantly improved by TPGs and TLPAs combinations. Particularly, the treatment of TPGs:TLPAs (7:3) had the most obvious effect on improving the ischemic injury compared with other groups (P < 0.01, P < 0.05).
In sham control group, polygons neurons, abundant cytoplasm, clear nucleoli and integral cell membrane were observed by TEM (Fig. 1D) by TEM. After MCAO modeling, neurons were swelled, organelles were broken www.nature.com/scientificreports www.nature.com/scientificreports/ or disappeared, and cell membranes were adhered with synapses. Meanwhile myelin in most synapses from surrounding stroma was disappeared (red arrow) or swelled (yellow arrow). After being treated with Ni, TPGs, TLPAs, TPGs:TLPAs at 7:3 and 5:5, apoptotic neurons were significantly ameliorated. Especially, the treatment with TPGs:TLPAs (7:3) had the most obvious improvement on neurons apoptosis compared with other combinations (P < 0.01, P < 0.05).

Protection of TPGs:TLPAs at 7:3 on blood-brain barrier in MCAO rats. Described as shown in
Protection of TPGs:TLPAs at 7:3 on apoptotic protein in MCAO rats. After modeling, protein expression of Bax and caspase-3 were significantly enhanced whereas Bcl-2 was apparently declined (P < 0.01). Compared with model group, protein expressions of Bax and caspase-3 in rats brain decreased significantly while Bcl-2 was significantly increased after treatment the with Ni, TPGs, TLPAs, TPGs:TLPAs at 5:5 and 7:3 for two weeks (P < 0.01, P < 0.05) (Fig. 3A). It was important that protection of TPGs:TLPAs at 7:3 on apoptosis in MCAO rats was significantly better than any other groups (P < 0.01, P < 0.05).
Compared with sham control group, mRNA level of Bax and caspase-3 in rats brain were apparently enhanced while mRNA level of Bcl-2 was apparently declined (P < 0.01) in model group. The treatment of Ni, TPGs, TLPAs, TPGs:TLPAs (5:5) and TPGs:TLPAs (7:3) significantly decreased the mRNA levels of Bax and caspase-3 whereas apparently increased Bcl-2 mRNA level when compared with model group (P < 0.01, P < 0.05) (Fig. 3B). This protection of TPGs and TLPAs at the ratio of 7:3 on apoptosis in MCAO rats was significantly better than other groups (P < 0.01, P < 0.05). The results of both Western blotting and qPCR are consistent.

Discussion
TCM, consisting of mutiple herbs, was benefical to synergistic effect of multi-components for the treatment and prevention of diseases. In traditional application, CS (Paeonia lactiflora Pall.) and CX (Ligusticum chuanxiong Hort.) were generally used as "drug pair" to treat the stasis and stroke. As known, CS was used for "eliminating blood stasis and relieving pain" (Chinese Pharmacopoeia 2015). Modern phytochemical and pharmacological studies have shown that monoterpene glycosides, galloyl glucoses and phenolic compounds are the major bioactive components of CS. Research showed its major compound paeoniflorin was benefical for auto-inflammatory and autoimmune disease treatment 9 . In this study, organic acids including ferulic acid and caffeic acid were isolated and identified from Ligusticum Chuanxiong Hort. These components have been proved to hold chemoprevention on apoptosis and reduction on the risk of ROS-mediated diseases 15 . According to our previously study, the combination of CS and CX have potential effects on curing cerebral disease or stroke. However, the best effeciacy of these components might be associated with its combination with the optimal ratio. Total paeony glycosides (TPGs) are the main active constituents in CS and total ligustici phenolic acids (TLPAs) are the major active components in CX. TPGs were reported to have a good effect on both cerebral ischemia and cardiac ischemia 16,17 .   Table 5. The effect of XS on SOD, CAT, GSH-Px activity and LOP, MDA content in rats serum of different treatment groups after cerebral ischemia (x ± SD, n = 24). Notice: ## P < 0.01 or # P < 0.05, compared with sham group; **P < 0.01 or *P < 0.05, compared with model group; $$ P < 0.01 or $ P < 0.05, compared with TLPAs group; && P < 0.01 or & P < 0.05, compared with TPGs group; %% P < 0.01 or % P < 0.05, compared with TPGs:TLPAs (5:5) group.
In current study, we use the ischemic model both in vitro and in vivo to identify the best combined ratio of TPGs and TLPAs against focal cerebral ischemia. The optimal ratio of TPGs and TLPAs at 7:3 could bring more remarkable protective effects against focal cerebral ischemia in MCAO rats by alleviating oxidative stress, inflammatory and neuronal apoptosis to protect the blood-brain barrier. The present study provided benefical evidence for clinical application of CS and CX as a "drug pair".
smoking and tiredness, may cause irreversible structural changes in a specific vascular territory 18 . However, the effect of cerebral ischemia on BBB, a diffusion barrier essential for maintenance of homeostasis and normal function of the central nervous system, has been broadly investigated 10 . In current study, we investigated the protection of the combination of TPG and TLPA on BBB by anti-inflammation, anti-oxidative stress and anti-apoptosis effects.
Matrix metalloproteinases (MMPs), a family of zinc-binding proteolytic enzymes, contribute to the pathology of cerebral infarction by degrading a number of extracellular matrix molecules to breakdown the blood-brain barrier 21 MMPs-mediated BBB disruption has been found to be a critical pathological mechanism of cerebral disease or stroke 22 . However, this activation of MMPs is complexly modulated by transcriptional regulators and inhibitor tissue inhibitor of metalloproteinases (TIMPs) 23 . Under normal physiological conditions, MMPs are directly responsible for pathogenesis of cerebrovascular diseases and their activity is regulated by TIMPs 24 . The evidence showed that MMP-2 and MMP-9 might be up-regulated and activated during cerebral disease or stroke which are expressed particularly frequently in the nervous system 25 . MMP-2 also plays a significant role in the regulation of activity of platelets 26 , while MMP-9 damages matrix components of the basement membrane leading to neuro-inflammation after focal cerebral ischemia by maintaining the integrity of cerebral vasculature 27 . MMP-9, one of the most important MMPs correlated with BBB damage, accompanied with its endogenous inhibitor TIMP-1, has been confirmed to be important in vascular remodeling and neuro-protective effect 28 . This study indicated that there was a more significant effect of TPGs:TLPAs (7:3) on adjusting MMP-2, MMP-9 and TIMP-1 protein expression and mRNA level than other groups.
It has been proved that apoptosis is associated with cerebral disease or stroke leading to BBB dysfunction, inflammation and oxidative stress resulting in brain damage [29][30][31] . Histopathological changes and TEM images in current study indicated that TPGs and TLPAs especially the TPGs:TLPAs at 7:3 could alleviate neuronal apoptosis.

Conclusions
In this study, the optimal ratio of TPGs and TLPAs on improving cerebral ischemia damage was compared and screened according to ODG-induced HUVEC cells in vitro and MCAO rats in vivo. The results demonstrated the combination TPGs and TLPAs at 7:3 might be more benetifical to this protective effect via alleviating oxidative stress, inflammation and neuronal apoptosis. This study provided an evidence of the combination of CS and CX for clinical application as "drug pair", and more importantly, the study revealed that the proper ratio of TPGs and TLPAs could be critical for CS-CX "drug pair" in curing focal cerebral ischemia.  . Effect of TPGs and TLPAs on TUNEL staining in brain tissue of cerebral ischemia rats. (A) Images of TUNEL staining; (B) Apoptotic index. ## P < 0.01, compared with sham group; ** P < 0.01 or * P < 0.05, compared with model group; $$ P < 0.01 or $ P < 0.05, compared with TPGs group; && P < 0.01 or & P < 0.05, compared with TLPAs group; % P < 0.05, compared with TPGs:TLPAs (5:5) group.

Preparation of TPGs and TLPAs. The preparation of TPGs was performed as previous studies with mod-
ifications. Pieces of CS with 1 kg were weighed and extracted twice (1.5 h/time) with 10 times amount of 70% (v/v) ethanol in reflux extraction device. The extraction was concentrated to 1 g decoction pieces/mL. TPGs was defined as the 30% ethanol eluant chromatographed with a D-101 macroreticular resin (Cangzhou Bon Adsorber Technology Co., Ltd, Cangzhou, China). On the basis of previous studies 32 , 1 kg CX were weighed and extracted twice (1.5 h/time) with 15 times amount of 80% (v/v) ethanol in reflux extraction device. The extraction was concentrated to 1 g decoction pieces/mL and was chromatographed with AB-8 macroreticular resin (Cangzhou Bon Adsorber Technology Co., Ltd, Cangzhou, China) column. Finally, the eluant of 30% ethanol was defined as TLPAs. The yield of CS and CX extraction was 17.81% and 16.63%, respectively. The content of TPGs was 3.62% after extraction with 70% ethanol. And after purification by macroporous resin, the content of the TPGs used in the study was 47.92% finally. While the content of TLPAs was 2.42% after extraction with 80% ethanol, and the content of the TLPAs used in the study was 41.43% finally after purification(The data shows in Supplementary Information). Cells culture. HUVEC was purchased from Shanghai Institute of Biochemistry and Cell Biology (Shanghai, China) and cultured in an incubator with 5% CO 2 at 37 °C. Cells were treated with RPMI 1640 medium containing 80 units/mL of penicillin, 0.08 mg/mL of streptomycin (KeyGEN, China) and 10% fetal bovine serum (FBS, Gibico, USA).
Cell OGD model and drug treatment. HUVECs modeled with oxygen glucose deprivation (OGD) was used as an ischemic model in vitro in according to previous reports. Cells of 90% confluence were cultured in 96-well plates (5000 cells/well, 100 μL). For simulated ischemia-reperfusion, cells were placed in a hypoxic incubator (5% CO 2 , 95% N 2 ) at 37 °C for 4 h.
Animal model and drug treatment. Male Sprague-Dawley rats (280 ± 20 g, licence number: SCXK (Su) 2015-0001) were supplied by Nantong University (Nantong, China). Rats were kept at a temperature of 25 °C and the relative humidity of 45%. All the experiment procedures were approved by the Institutional Animal Care and Use Committee of the Jiangsu Provincial Academy of Chinese Medicine in accordance with published National Institutes of Health guidelines, the ethics code number of our animal experimental was ZYY20151207. Before modeling, rats were housed for 7 d to adapt to the environment on a 12 h/12 h light/dark cycle with food and water ad libitum.
The middle cerebral artery occlusion (MCAO) model was performed as described previously. Rats were injected intraperitoneally 10% chloral hydrate (300 mg/kg) for anesthetizing. After anesthetized, rats were laid in dorsal recumbency and the right side of carotid artery was exposed and isolated. The origin of the right middle cerebral artery (MCA) was occluded by inserting a 0.26 mm heparin-dampened monofilament nylon suture (Ethicon, Inc., Osaka, Japan) with a heat-rounded tip from the external carotid artery (ECA) to the right internal carotid artery (ICA). Thread at the distal end of common carotid artery (CCA) was fastened when the suture was advanced closed to the origin of the MCA. Extravascular suture was cut before the closed neck incision. After preventing wound infection, rats in the sham group were performed with the same surgical procedures without insertion of the heparin-dampened monofilament nylon suture. The body temperature of these rats was maintained at 37 °C during surgery. After surgery, sodium penicillin (100,000 U/a rat/d) was given to prevent infection for three days successively.
Neurological examination was conducted to screen the successful models after modeling for 24 h. The successfully modeled rats were randomly divided into 7 groups (n = 24/group): model group (MCAO, Saline 10 mL/ kg); sham control group (Saline 10 mL/kg); positive control group (Nimodipine 7.53 mg/kg); TPGs group; TLPAs group; TPGs:TLPAs at 7:3 group and TPGs:TLPAs at 5:5 group. The treatment drug dose was 500 mg/kg according to the previous research 17 . Equivalent saline was given to the sham rats and model rats. All rats were treated by gavage administration for two weeks consecutively.
Immunohistochemistry. Brain tissues were removed and immediately immersed in formaldehyde solution. Paraffin embedded tissues were sectioned (5 μm thick) and labeled with 3% H 2 O 2 , 3% normal goat serum. Briefly, sections were incubated with anti-mouse antibodies overnight. Then secondary antibodies, secondary biotinylated conjugates and diaminobenzidine were added as the instructions of DAB kit. Finally, slides were dehydrated, cleared, and mounted for visualization using IX51 microscope (Olympus Corporation, Japan).
Transmission electron microscope (TEM). Preparation of samples was performed according to previous method 36 . Sections of 1 mm3 were taken out from hippocampal CA1 region of rats' brain and fixed with 5% glutaraldehyde followed by PBS (pH = 7.4) washing. Then sections were postfixed in 1% osmic acid and dyed with 2% uranyl acetate. Sequentially, sections were dehydrated by 50%, 70%, 90% and 100% acetone and embedded. Samples were analyzed using the JEM-1200EX transmission electron microscope (JEOL, Japan) at a voltage of 80 kV.
Western blotting analysis. Protein levels in brain tissues were determined by Western blotting analysis.
Briefly, proteins were extracted from tissue homogenates and an equal amount of total protein was separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred from the SDS-PAGE gel to PVDF membrane (Millipore, USA). After being blocked with 5% BSA in tris-buffered saline Tween-20 (TBST), membranes were incubated with the primary antibodies at a dilution of 1:1000 overnight. Subsequently, membrane was washed (three times, 5 min), and incubated with secondary antibody (1:5000) at 37 °C for 30 min. The blots were visualized with ECL-Plus reagent (Santa Cruz, USA) and analyzed with Image pro plus (IPP 6.0, Media Cybernetics, MD, USA) software. β-actin was used as loading control.
Quantitative PCR (Q-PCR). Total RNA of all brain samples were extracted by TRIzol reagent. Then it was reverse transcribed with a SuperScript III First-Strand Synthesis System for quantitative real-time polymerase chain reaction (q-PCR) following manufacturer's indications. The Primer sequences used in this study were shown in Table 7. SYBR green of 10 µL, 6 µL molecular grade water, 1 µL of each forward and reverse primers and 2 µL cDNA were contained in each reaction of the PCR plate. The amplification was performed under the following conditions: 10 minutes at 95 °C, 50 cycles at 95 °C for 15 seconds and 60 °C for 60 seconds. Q-PCR was performed under standard conditions and all experiments were run in triplicate. GAPDH was used as the internal reference.
TUNEL staining. Cell apoptosis in the brain was detected by TUNEL staining according to the manufacturer's protocol (KeyGEN, Nanjing, China). Briefly, 5 μm slide sections were deparaffinized in Histoclear, rehydrated and treated with NaCl (0.85% w/v) for 5 min, then washed with 4% w/v paraformaldehyde and permeabilized with Proteinase K. After that sections were equilibrated and incubated with Biotinylated Nucleotide mix and recombinant Terminal Deoxynucleotidyl Transferase (rTdT) at 37 °C for 1 h till the