Paeonia lactiflora Pall. (20140901), Rheum officinale Baill. (20140601), Rehmannia glutinosa(Gaertn.)Li-bosch. (20140102), Panax notoginseng (Burk.) F. H. Chen ex C. Chow. (20151021), Paeonia suffruticosa Andr. (20140901), Acorus tatarinowii Schott. (20140301), Bubalus bubalis Linnaeus (20140601), Pheretima aspergillum (E. Perrier). (20140601), were purchased from Tong Leng Hetian Chinese medicine company.

Total extract of LP (TLP): Heat reflux extraction with water technique was applied to obtain TLP. At first, all ingredients of LP were dry matter and smashed. Then 1 kg mixture of LP was soaked in distilled water for 60 min, extracted in a 20 L round-bottomed flask containing 10 L of distilled water for 1 h and repeated heat reflux extraction for 1 h with10 L of new distilled water again. Then double extraction solution were combined and evaporated under reduced pressure to remove most solvents by using a rotary evaporator. Finally the extraction was dried into lyophilized powder. A total of lyophilized powder of 641.15 g was got with 5 Kg mixtures of LP.

n-Butanol fraction extract (LPB): A total of 5 Kg mixture of LP was extracted with heat reflux extraction with water, Then double extraction solution were combined, evaporated and extracted by equivalent volume of n-butanol at five times. Finally, the n-butanol fraction liquid was combined, evaporated and dried into lyophilized powder of 109.60 g.

Volatile oil fraction extract (LPV): 1 Kg mixture of LP was put in a 20 L round-bottomed flask containing 10 L of distilled water, which was connected to a Clevenger-type apparatus with tap water for cooling. The mixture of LP was under method of heat reflux extraction with water for 1 h. Meanwhile, the obtained essential oil was collected in the side arm, then separated and dried with anhydrous sodium sulfate to eliminate moisture. Finally, the volatile oil of 9.10 ml was obtained with 5 Kg mixtures of LP (á = 0.91 g/ml).

Aspirin was purchased from by Bayer HealthCare with the batch number of J20130078). Enzyme-linked Immunosorbent Assay (ELISA) Kits including Rat interleukin 1 (IL-1), Rat S-100B protein, Rat nuclear factorκB (NF-κB), Rat neuron specific enolase (NSE), Rat high sensitive C reactive protein (HS-CRP), Rat estradiol (E2), Rat tumor necrosis factor-α (TNF-α), and Rat D-dimmer (D2D), were purchased from Nanjing Senbeijia Biological Technology Co., Ltd., Trizol was purchased from Invitrogen. Anti-AKT, Anti-PI3K, and Anti-β-actin antibody was all purchased from Abcam. All durgs (purity assay by HPLC ≥ 98%, power) including compounds of Emodin, Ginsenoside Rg1, Ginsenoside Rb1, Notoginsenoside R1, Baicalein, Geniposide, β-asarone, Taurine, Paeoniflorin, Paeonol, Apigenin, Rhein, and Catalpol, were purchased from Chengdu Must Bio-Technology Co., Ltd.

48-male-(8-week-old) SHR (180–210 g), 8 male Wistar Rats (180–210 g) were purchased from Vital River Laboratory Animal Technology Co. Ltd. (License: No. SCXK (Jing) 2012-0001). All mice were housed with standard controlled conditions (12/12 h light/dark with humidity of 40–60%, 21 ± 2°C), and were allowed free available to food and water. Then they were acclimatized in the laboratory environment for a week prior to the experiment. The health condition of rat was checked regularly.

Eight Wistar rats as blank control groups (Distilled Water group), 48 SHR were at random assigned into six groups (n = 8 in each group) including model group (SHR group only), positive control group (Aspirin), LPB high dosage group, LPB low dosage group, LPV high dosage group, and LPV low dosage group. Administration of gastric infusion was executed in the dose of 10 mL/kg once a day in each group, 15 days in total. On the 15th day, 1 h after administration, draw 5 ml blood from carotid artery, and transfer it into centrifuge tube containing 0.038 g/ml sodium citrate. Centrifuge at 3000 rpm for 15 min. Collect the top layer of plasma to determine indexes of NSE, IL-1, S-100B, NF-κB, TNF-α, E2, D2D, and HS-CRP.

PC12 cells, kindly donated from Professor Lu (Department of Chinese Medicine Pharmacology, Nanjing University of Chinese Medicine) were cultured at 37°C in DMEM containing 10% (v/v) heat-inactivated fetal bovine serum (FBS; GIBCO), 100 U/ml penicillin and 100 μg/ml streptomycin (Hyclone, J150019) under a humidified atmosphere of 95% air and 5% CO₂. For cell differentiation, cells were treated with 50 ng/mL of nerve growth factor (NGF; Sigma-Aldrich, USA) for 48 h. Afterwards, NGF-differentiated cultures were pretreated with different doses of each medicine at 1, 5, 10, and 20 μmol/L for 1 h, and then expored to10 mM L-glutamate for an additional 24 h. The untreated PC cells without the treatment of L-glutamate were used as control group.

PC12 cells were seeded into 96-well plates at a density of 50 cells/μL and were treated by drugs in the same way as described above, and then gently washed with PBS. Afterwards, 10 μl of CCK-8 solution and 90 μl of culture medium were added to each well. The 96-well plates were maintained at 37°C for 4 h and the absorbance was measured in a microplate reader (Rayto; RT-6000) at 450 nm. All groups were repeated in triplicate.

The PC12 cells were seeded in 6-well plates at a density of 1 × 10⁵/ml and incubated for 24 h. Cells were treated by drugs in the same way as described above. Cells were treated with10 mM glutamate for 24 h, then colleted and harvested with 0.25% trypsin and made into a single cell suspension, washed with cold Phosphate buffer (PBS) twice. PC12 Cells were resuspended as a single cell suspension with 50 μl PBS, fixed with l ml of cold 70% ethanol. Then 5 μl of Annexin V-FITC and 5 μl of PI stained cells in order. The cells were incubated at room temperature (20–25°C)for 15 min in the dark and determined with a FACS Calibur flow cytometer (Becton–Dickinson).

PC12 cells were seeded in 6 well plates and then resuspended at 1200 rpm for 10 min in DMEM. Total RNA was extracted with Trizol reagent (Invitrogen, America), and the concentration of total RNA was measured with Merinton SMA4000 and the purity was estimated with the ratio of A260/A280 between 2.0 and 2.3. The first strand of cDNA was obtained with Prime Script^TM RT Master Mix (Takara), and expression levels of mRNA was quantified with real-time PCR amplification instrument (Eppendorf, Germany). The program was included 1 cycle of 95°C for 5 min, 40 cycles of 95°C for 5 s, 60°C for 31 s. The mRNA levels of each gene took mRNA of GAPDH as the standard. The specific primers used in the reaction were as follows. The experiment was conducted in triplicate.

The PC12 cells were washed with pre-cooled PBS buffer for 3 times, and lysed with RIPA buffer (100 μl/50 ml) (Beyotime, Shanghai, China) and then placed for 30 min on ice. Cells were harvested by scraping, centrifuged at 12,000 rpm for 10 min. Bovine serum albumin (BSA) (2 μg/μl) was diluted in PBS, and protein concentration was detected with BCA Protein Assay Kit (Pierce, Thermo Scientific) at a 50:1 ratio. The lysates (2 μl) were diluted in double-distilled water (18 μl) (each sample in two wells). 200 μl of detection solution were added to each well in a 96-well plates. The OD values of samples were tested at a wave length of 490, and standard curve was drawn to calculate the concentrations of the proteins for every sample. The samples were first electrophoresed at 60 V, and then the voltage was rised to 120 V as soon as the samples reached the separation gel, followed by electrophoresis (conducted at 4°C) for 2 h. After electrophoresis, the proteins were transferred onto polyvinylidene fluoride (PVDF) membranes for 2 h (conducted at 4°C). The PVDF was removed from the transfer apparatus, and the samples were blocked with 5% evaporated milk in tris-buffered salinetween (TBST) and incubated at room temperature for 1 h. The membranes were individually incubated with anti-phospho-Akt(1:1,000, Abcam, United Kingdom) and anti-phospho-PI3K (1:1,000, Abcam, United Kingdom) with the samples at 4°C overnight. After three washes with TPBS, secondary antibodies (HRP; l:5000, Abcam, United Kingdom) corresponded to the respective primary antibodies at room temperature for 1 h. The images were captured with the Clinx3200 + System (Clinx Science Instruments Co., Ltd., Shanghai, China), and protein densitometry was analyzed with Quantity One software (Bio-Rad, Hercules, CA, United States). Each membrane was stripped and quantified as a ratio to β-actin.

With the purpose of screening out the potential compounds in LP, OB and DL evaluation were applied. According to screening rules of OB≧30% and DL≧0.1, a total of 32 potential compounds of six herbal medicines in LP were selected out (Table 1). However, some compounds among them which couldn’t meet the requirements, were confirmed by experimental evidence and then were also allocated to potential compounds. For example, Catalpol has been displayed that it had the properties of anti-apoptosis, anti-inflammation and neuroprotection against Alzheimer’s disease, ischemic stroke and idiopathic Parkinson’s disease (Jiang et al., 2015). Geniposide enhanced growth factor signaling and inhibited apoptosis of neurone in the mouse model of Parkinson disease to exert neuroprotective effect (Jiang et al., 2015). β-Asarone was also added as potential compound for further analysis because β-Asarone treatment could suppress Beclin-1-dependent autophagy through the PI3K/Akt/mTOR signal pathway (Deng et al., 2016). Panax notoginseng saponins (PNS) were mainly composed of Ginsenoside Rb1, Ginsenoside Rg1, and Notoginsenoside R1,which had positive protection on ischemia brain damage with the activity of anti-inflammation and anti-apoptosis (Huang X.P. et al., 2015). Paeonol may possess anti-inflammatory and anti-oxidant functions with its active metabolites (Jin et al., 2016). In addition, Taurine was found that it promoted synapse development and proliferation of neuron (Liu et al., 2010). Lumbrokinase may relieve ischemia-reperfusion injury via TLR4 signaling (Wang et al., 2016). In all, It was found that 34 potential compounds in LP were preserved for subsequent analysis.

Traditional Chinese Medicine which involved many bioactive compounds may hit one or multiple targets from the network of biology. Then the biological system achieved new balance to reduce the harmful effects. Typically, Potential compound-targets- disease (P-T-D) network in LP on ICH was constructed to forecast and interpret the poly pharmacology action of multi-compound and multi-target. As shown in Figure 4, there are 34 potential compounds and 146 corresponding targets with the connections also reaching 506 items which indicates the close relationships between them. Lots of associated targets independently belong to inflammation (36/146), cell apoptosis (21/146), cell proliferation and differentiation (22/146), oxidative stress (7/146), destruction of blood-brain barrier (7/146), neurotransmitter (13/146), calcium overload(5/146), and coagulation cascade (11/146), which implied that the compounds of LP may intervene many pathologies of ICH such as neurologic function, thrombosis, oxidative stress, cell apoptosis, and inflammation. For example, prothrombin (F2) and coagulation factor VII (F7) are important coagulation factors in the thrombosis process. The injured vessels induce these circulating coagulation factors in a chain of interlacing reactions, causing development of a coagulum to generate thrombosis, and multi-compound in LP might regulate coagulation factors to put into effect.

A total of 142 corresponding targets rooted in P-T-D network were chosen as potential ICH targets to carry on Go analysis. As shown in Figure 5, 24 signaling pathways of ICH associated with potential compounds in LP were fully excavated by REACTOME pathways with 9841 available unique genes. The main pathways were involved in inflammation, coagulation, apoptosis and other pathologic links and relevant signaling pathways. Specifically, the reactome pathways were mainly related to Ligand-dependent caspase activation, signaling by ERBB2, SHC-related events triggered by IGF1R and intrinsic pathway for Apoptosis. Moreover, although some pathways didn’t occupy larger proportion, they also display tremendous importance, such as dissolution of Fibrin Clot, inflammasome pathway, AKT phosphorylates targets in the cytosol, and degradation of the extracellular matrix. These reactome pathways have been linked to the internal mechanism of hemorrhagic stroke and indicated the complexity of disease pathological process in more depth. Protein targets of ICH were shown to be involved in autophagy, ischemia, necrosis, apoptosis, calpain activation, inflammation, oxidative stress, caspase activation and cytokine secretion based on the current level of global pathway analysis. Go analysis resulted that these potential protein targets of potential compounds in LP were involved in specific pathological pathways basically.

The perihematomal region resulted in a prominent form of cell death during the development of ICH. Therefore, inhibition of the neuron apoptosis has been expected to improve functional recovery and reduce tissue damage in ICH. Among the pathways of ICH associated with potential compounds in LP, the PI3K-AKT pathway wasn’t dominant position, but it had obvious upstream and downstream proteins which played a significant role in neuronal apoptosis, and an important contributor to early brain injury after ICH by taking part in cell survival and proliferation and regulating some important caspase proteins and apoptotic genes. As shown in Figure 6, PI3K stimulated the generation of the lipid second messenger, converting phosphatidylinositol (4,5) bisphosphate (PIP2) into phosphatidylinositol (3,4,5) trisphosphate (PIP3). PTEN antagonized action of PI3K, and Akt resorted its pleckstrin homology domain to bind the 3^′-phosphorylated inositol lipids. Compounds in LP have interferential effect on the phosphorylation of PI3K and AKT proteins based on the dissection of network pharmacology. PI3K may be targeted by 19 compounds from eight ingredients of LP such as Emodin, Baicalein, Ginsenoside Rg1, and so on. AKT may be targeted by 17 compounds from eight ingredients of LP such as Catalpol, Paeoniflorin, Paeonol, Ginsenoside Rg1, Ginsenoside Rb1, and so on. PI3K inhibitor could be targeted by only one compound Paeonol. The PI3K-Akt pathway acted an important part in inhibition of apoptosis and neuroprotection (Wang et al., 2013). The downsteam target AKT influences many other pathways such as inhibition of caspase and the regulation of Bcl-2 family proteins, NF-kβ pathway, JNK pathway, and ERK pathway (Jacquin et al., 2013; McNamara et al., 2013; Zhou et al., 2015; Velmurugan et al., 2017).

Based on the network pharmacological prediction, the compounds of LP may act on different pathologies of ICH such as neurologic function, thrombosis and inflammation. ICH stands for the acute presentation of cerebral small-vessel disease. Persistently raised intraluminal arterial pressure destories small-vessel walls and hypertension becomes the leading cause of ICH (Biffi et al., 2015). The level of IL-1, HS-CRP, NF-κB, S-100B protein, TNF-α, NSE, D2D, and E2 were detected in SHR treated with extracted fractions from LP. The LPV and LPB, two effective extracted fractions in LP for anti-inflammatory and activating blood on blood-heat and blood-stasis rat model, were adopted to estimate pharmacodynamic evaluation of LP on SHR model.

High sensitive C reactive protein, TNF-α, NF-kβ, and IL-1 levels in plasma in model group were obviously higher than that in control group (P < 0.01, Figure 7A). HS-CRP, TNF-α, NF-kβ, and IL-1 levels in aspirin group as positive medicine (Chong et al., 2012) was observably lower than that in model group (P < 0.01, P < 0.05). Compared to model group, after treatment with different doses of LPB and LPV, HS-CRP, TNF-α and IL-1 levels in serum were all significantly decreased (P < 0.05) and NF-kβ levels in LPB treatment groups were all significantly decreased (P < 0.01).

S-100B and NSE level in model groups increased remarkably in contrast to control group (P < 0.01, Figure 7B). In comparison with model group, level of S-100B and NSE in aspirin, LPV and LPB groups decreased significantly (P < 0.01). Compared with control group, D2D and E2 level in model group show a significant difference (P < 0.01, Figure 7C). Compared with model group, D2D and E2 level of aspirin, LPV and LPB groups reduced significantly (P < 0.05, P < 0.01). The D2D is very sensitive to intravascular thrombus and may be markedly elevated in disseminated intravascular coagulation. E2 exerted a protective effect by intracellular Ca²⁺ levels (Koh, 2014).

Cell proliferation was evaluated on the glutamate-induced PC12 cell treated with extracted fractions and compounds of LP. As is shown in Table 2, L-glutamate obviously inhibited cell proliferation in comparison to control group (p < 0.01), different dosages of LPB and TLP all can enhanced proliferation of PC12 cells (p < 0.01). The group dealt with LBV only enhanced cell viability at 6.435 μg/mL (p < 0.01).

In the meanwhile, Emodin, GRb1, Catalpol, Paeonol, GRg1, NR1, Baicalein, Geniposide, Taurine, β-asarone could effectively increase proliferation of PC12 cells at 20 μM (p < 0.05, p < 0.01, Figure 8). GRb1, Geniposide and Taurine showed fine effects with the lowest concentration of 5 μM (p < 0.01). Therefore, these compounds except Rhein, Paeoniflorin and Apigenin all increased cell proliferation on L-glutamate-induced PC12 cell.

To ascertain the anti-apoptotic effect of LP, apoptosis rate of pretreatment with TLP, LPB, and LPV were evaluated on L-glutamate-induced PC12 cell. 10 mM glutamate was able to induce PC12 cells apoptosis from3.79 to 19.03% in 24 h (P < 0.01, Figure 9A). However, pretreatment with TLP, LPB, or LPV for 1 h prior to glutamate exposure decreased the apoptosis rate to 5.39% (LPB), 4.06% (TLP), and 5.65% (LPV), respectively (P < 0.01). Therefore, LP had obvious neuroprotective effect when PC12 cells were exposed to amino acid damage.

For determining whether the six herbal ingredients in LP were different or similar on compound composition, total compounds in six herbs were compared under important properties including MW, nHAcc, nHDon, OB, MlogP, and DL. On one hand, the distribution interval of various properties for all compounds in six medicinal herbs of LP show obvious difference. On the other hand, properties of total compounds of blood-activating herbs such as PS and PN undifferentiated with PR in MW, nHDon, nHAcc, MlogP, and OB, properties of total compounds of heat-clearing herb such as RO only had subtle differences in MW and MlogP compared to RG (p < 0.05, Table 3). However, six properties of total compounds in AT for resuscitation and palinesthesia of consciousness almost showed significant differences with other herbs (P < 0.01). The results showed that although there were various compounds in the various herbs of LP, the properties of heat-clearing herbs as PR and PS were similar, the same as RO and RG as blood-activating herbs.

According to the prediction result of PI3K/AKT pathway on ICH in LP (Figure 2), effect of compound combination in LP on L-glutamate-induced PC12 cell proliferation was further investigated with the three-factor two-level factorial design. The result showed that first-level interactions existed between Taurine and Paeonol, Taurine and Geniposide, Paeonol and Geniposide, Ginsenoside Rg1 and Ginsenoside Rb1 (p < 0.01, Table 4), but secondary interaction didn’t exist among compound combination. Taurine was a characteristic compound from BB and Geniposide was an important compound from RG. BB and RG were classic herb pair in TCM, therefore compound combination of taurine and geniposide was reserved for further study. Potential compounds Ginsenoside Rg1 and Ginsenoside Rb1 were from the same herb of PN, which also were reserved to explore the effect of compound combination by PI3K/AKT pathway.

Based on the effect of the compound combination of LP in the experiment above (Table 4), combination of Ginsenoside Rg1 and Ginsenoside Rb1, Taurine and Geniposide on neuronal apoptosis were evaluated with the model of L-glutamate-induced cell apoptosis. The apoptosis radio of cells was increased from 3.79 to 19.03% after the cells were treated with 10 mM glutamate for 24 h (p < 0.01, Figure 9B). The combination of Taurine and Geniposide, Ginsenoside Rg1and Rb1 at 5 μM effectively decreased percentage of apoptosis on L-glutamate-induced PC12 cell and normal PC12 cell obviously (p < 0.01), and were the same as at 10 μM respectively (P < 0.01). The results showed that the combination of Ginsenoside Rg1 and Rb1, taurine and geniposide in LP at low concentration had a significant increase in efficiency on inhibition of neuron apoptosis.

In order to verify regulation effect of compounds and their combination in LP on PI3K/AKT signaling pathway, the mRNA and protein levels of PI3K and AKT were measured with RT-QPCR and western blot (Figure 10). Compared with normal groups, L-Glu-damaged groups showed down-regulated of PI3K and AKT mRNA and protein expression (P < 0.01). Compared with L-Glu-damaged groups, GRb1 upregulated PI3K and AKT mRNA expression (P < 0.01). The combination of taurine and geniposide at 5 μM upregulated AKT mRNA the expression (P < 0.05). However, Taurine or Geniposide at 10 μM, and combination of Taurine and Geniposide at 5 μM up-regulated p-AKT and p-PI3K protein expression on L-Glu-damaged PC12 cell (P < 0.01). GRb1 at 10 μM, and combination of GRb1and GRg1 at 5 μM also up-regulated p-AKT and p-PI3K protein expression on L-Glu-damaged PC12 cell (P < 0.01). In addition, combination of GRb1and GRg1 at 5 μM up-regulated p-AKT and p-PI3K protein expression on normal PC12, while combination of Taurine and Geniposide at 5 μM down-regulated p-AKT and p-PI3K protein expression on normal PC12. The results showed that Taurine, Geniposide and GRb1 in LP individually up-regulated the level of PI3K and AKT on the level of protein obviously. Furthermore, the combination of Taurine and Geniposide, GRb1 and GRb1 at half-dosage respectively had a significant increase in efficiency on up-regulation of protein expression of PI3K and AKT.