First, the active ingredients of Chuanxiong were screened from the TCMSP database (Ru et al., 2014) (https://www.tcmsp-e.com/). To find the active ingredients, the DL value should be less than 0.18, and the OB value should be less than 30%. Second, each ingredient’s SMILE and SDF files were obtained from the PubChem database (Kim et al., 2019) (https://pubchem.ncbi.nlm.nih.gov/). Third, the targets associated with Chuanxiong were ascertained from SwissTargetPrediction. The targets related to Chuanxiong were demonstrated in SwissTargetPrediction (Daina et al., 2019) (http://www.swisstargetprediction.ch/). We chose the targets with probability values less than 0. Using “angiogenesis” as the search term, we used the DrugBank (https://www.drugbank.com/) to extract angiogenesis-related targets. The obtained Chuanxiong targets and disease targets were topologically calculated by Cytoscape 3.7.1, and the “drug–ingredient–disease–target” network was performed. Based on Chuanxiong targets for angiogenesis, protein–protein interactions (PPIs) were gathered from STRING (https://cn.string-db.org/cgi/input.pl).The network of PPI was constructed, and the topological features of the network were analyzed to screen out the core targets that play a crucial role in the PPI network. We used Metascape to perform GO enrichment and KEGG pathway enrichment analyses of target genes, screening results with a p-value of 0.01, a minimum count of 3, and an enrichment factor of 1.5%.

Quail hatching embryos were obtained from the Qingfeng Aichongyuan livestock farm, Shandong Rizhao, China. HUVEC cell lines were obtained from the Chinese Academy of Medical Sciences. Dulbecco’s modified Eagle’s medium (DMEM), heat-inactivated fetal bovine serum (FBS), streptomycin, and penicillin were obtained from Thermo Scientific (Waltham, United States). Power SYBR Green PCR Master Mix reagents and gel embedding medium were obtained from Sigma-Aldrich (St. Louis, Missouri, United States. Methanol (≥99.9%), pyridine (≥99%), and acetonitrile (HPLC grade) were obtained from Beijing Chemical Plant Co. Ltd. (Beijing, China). Compounds 2-chloro-L-phenylalanine (≥98%) and N, O-Bis (trimethylsilyl) trifluoroacetamide with 1% trimethylchlorosilane (BSTFA +1% TMCS) were obtained from Shyuanye Biochemical Co., Ltd. (Shanghai, China; CAS number 103616-89-3, 25,561-30-2). Double-distilled water was purified from the Millipore water purification system (Millipore, Bedford, MA, United States). Methoxyamine hydrochloride (≥98%) and thiazolyl blue tetrazolium bromide (MTT, ≥98%) were purchased from Beijing Biorigin Biotechnology Co., Ltd. (Beijing, China). Tetrahydropalmatine (CAS No.2934-97-6,99.6%) and tert-butylhydroperoxide (CAS No.75.91.2, 99.6%) were purchased from Innochem (Beijing, China).

The Chuanxiong decoction was purchased from Beijing Tong Ren Tang (Beijing, China). The Chuanxiong decoction was pulverized into powder using a disintegrator and sieved through a 24-mesh sieve. Then, 60 g of Chuanxiong powder was extracted with the 10-fold volume of deionized water for 60 min and refluxed four times (1 h/reflux). The filtrate is combined and used for lyophilization. The Chuanxiong extract was stored as a lyophilized powder at −20°C after 3 days of lyophilization.

Quail eggs (10 ± 2 g; 10/group) were incubated in an egg incubator under specific pathogen-free conditions, such as maintaining at 37°C and 60% humidity. The eggs were separated into five groups at random, namely, negative control, positive control, low dose (1 mg/mL), medium dose (2 mg/mL), and high dose (4 mg/mL), comprising 10 eggs per group. After 7 days of incubation, a window was opened on the egg shell, the medicine (1mL/egg) is injected into the egg through the window, and then, the eggs were put back in the incubator after the window was taped shut. After 48 h of treatment and subsequent incubation at 37°C, the in vivo system of vascular microcirculation was observed in the embryos by optical coherence tomography (OPTPPRORE MICRO-VCCULTIM, Beijing HealthOlight Technology Co., Ltd., China). Then, the CAMs were excised, unfolded, and photographed. The pictures were analyzed on the MATLAB platform.

The TRIzol Reagent RA First Strand cDNA Synthesis Kit and an ABI Prism 7,500 sequence detection kit were used to isolate the total RNA from HUVEC samples. Life Technologies (Invitrogen; Thermo Fisher Scientific, Inc., MA United States) provided the specific VEGFR2, PI3K, and MAPK primers. The primer sequences are as follows: VEGFR2 primers were forward, 5′- AGC​ATA​GAC​AGC​CCT​TTG​GT -3′ and reverse 5′- CAC​AAT​CTC​TGC​TGG​TGC​AA -3′; PI3K primers were forward, 5′-ACT​GCC​GAG​AGA​TTT​TCC​CAC -3′ and reverse, 5′- TCA​CTC​ATC​TGT​CGC​AGG​CA -3′; and MAPK primers were forward, 5′- GAG​AGA​TGT​CTA​CAT​TGT​GCA​GGA​C -3′ and reverse, 5′- AAT​CTT​AAG​GTC​GCA​GGT​GGT​G -3’ (reverse primer). We used human GAPDH as an internal reference.

HUVECs were cultivated in Dulbecco's modified Eagle’s medium supplemented with 15% fetal bovine serum (FBS), 100 U/mL penicillin, and 100 g/mL streptomycin. In addition, 37°C, 5% CO₂, and 100% humidity were the conditions used to sustain the cultures. When the cells reached 70%–80% confluence, HUVECs were digested and centrifuged. Then, one 96-well plate of HUVECs was seeded for 24 h at 37°C before it was added for an additional 24 h of incubation. A Bio-Rad 550 spectrophotometer plate reader was used to detect the optical density at 550 nm (Bio-Rad, CA, United States).

HUVECs were cultured in 24-well plates until forming a monolayer. A clean scratch was made down the middle of the cell using a sterile pipette tip. Images of the same location were photographed using a light microscope (@10X, Nikon Eclipse TiE, Japan). Using Image-Pro Plus (version 5.0, National Institutes of Health, United States), we calculated the cell migration distance.

An assay for the closure of scratch wounds was used to investigate cell migration. In a 96-well plate, HUVECs were grown at a density of 1 × 10⁴ cells/cm² with a basement membrane matrix (Corning Matrigel, CAS number 356234) until the cell confluence reached 100%, and then, Chuanxiong was added for additional 6 h. Light microscopy (@10X, Nikon Eclipse TiE, Japan) and Image-Pro Plus (version 5.0, National Institutes of Health, United States) was used to measure the length of the tubes and capture pictures of the tube networks.

The cell samples from the cell lines were extracted using lysis buffer ultrasonically in ice. After 10 min of centrifugation at 4°C at a speed of 12,000 rpm, the supernatant was collected. Then, the concentration was measured using the BCA kit (Beyotime) according to the manufacturer’s instructions. SDS-PAGE gel electrophoresis was used to separate the proteins, and they were then transferred to a polyvinylidene fluoride film. After blocking in 5% BSA for 1 h, the membranes were treated with the corresponding primary antibodies (1:1,000) at 4°C overnight. After washing, the membranes were then incubated with the corresponding alkaline phosphatase-conjugated secondary antibodies (1:1,000) for 1 h. The Tanon 4200 SF Chemiluminescent Imaging System (Shanghai Tanon, Shanghai, China) was used to expose the membranes.

Data were analyzed by SPSS 16.0 software (Chicago, IL, United States). The data are displayed as mean ± standard deviation. In order to compare different groups, one-way analysis (ANOVA) of variance was used. Each two groups were contrasted using the independent t-test. These data discrepancies were deemed statistically significant when the differences reached a p-value of 0.05.

The TCMSP database had 189 active components of Chuanxiong in total, and seven active ingredients satisfied the criteria of OB value ≥30% and DL value ≥0.18. Then, 627 targets were obtained by SwissTargetPrediction. A total of 4,975 targets related to angiogenesis were obtained from the DrugBank. The intersection of component and angiogenesis targets yielded 51 common targets, and the drug–ingredient–disease–target network was constructed (Figure 1A).

The target interaction network graph is obtained with the confidence score set at 0.4 (Figure 1B). The nodes represent differentially expressed proteins, and the edges represent their interactions. There were 51 nodes and 203 edges in the network. Seven nodes with degrees higher than or equal to the median (degree = 8) of all nodes were selected as the core components and core targets of Chuanxiong in angiogenesis, including PPARG, HDAC1, CDK4, JAK2, HSP90AA1, CASP3, and SYAT3. These targets might be crucial to Chuanxiong’s ability to affect angiogenesis. A total of 44 cellular component terms, 81 molecular function terms, and 439 biological process terms were significantly enriched by GO enrichment analysis. The top 10 enriched pathways with significant differences were selected to construct a bubble chart (Figure 1C). Similarly, according to the p-value, the top 20 pathways in the KEGG enrichment study are shown (Figure 1D).

Using qCAM tests, Chuanxiong’s effectiveness in inhibiting angiogenesis was assessed (Figure 2A), and Pos was selected as the positive control according to previous studies. A measure of 1, 2, and 4 mg/mL were determined as different doses based on the previous dose screening results. Compared to the control group, the vessel numbers and vessel area significantly (p < 0.05) increased in the medium- and high-dose group (Figure 2B). It indicated that Chuanxiong has a notable dose-dependent proangiogenic effect. Additionally, the macroscopical results and the optical coherence tomography findings, which examined the vascular microcirculation of qCAMs in vivo, were in agreement (Figure 2C).

On the basis of the proliferation assays of HUVECs, the effects of Chuanxiong on angiogenesis were evaluated. The administration concentrations chosen were 25 µg, 50 μg, and 100 µg . Based on the previous report, Pos (40 µM) was chosen as the positive control. The survival rate of HUVEC was represented by MTT assays after 4 h of exposure to 5.5 μM tert-butylhydroperoxide (t-BHP), according to the preliminary assay that we performed (Cui et al., 2021). The results of MTT assays showed that middle-dose and high-dose groups could considerably (p<0.05) restore cell viability in t-BHP-damaged cells, protecting them against further harm (Figure 3A). Furthermore, VEGFR2 messenger RNA (mRNA) expression levels of the gene were measured by quantitative real-time PCR. The result showed that the messenger RNA (mRNA) expression levels in middle-dose and high-dose groups were significantly (p<0.05) higher than those of models, and these changes were dose-dependent (Figure 3B).

The tube formation and wound scratch assays of HUVECs were used to further evaluate the effects on angiogenesis of Chuanxiong in vitro. The model condition and positive control stayed together with the proliferation assays. As shown by the results, Chuanxiong could promote scratch healing percentages and increase the branch points of tubes (Figures 3C–F).

The results of Chuanxiong’s target prediction indicated that the main mechanism of Chuanxiong (2/5 with low p-value of top term) was connected to metabolic pathways. Therefore, we used DESI-MSI based on non-targeted metabolomics to explore the possible metabolic pathway of the pro-angiogenic effect of Chuanxiong. A scoring plot with R²Y (cum) of 0.779 and Q² (cum) of 0.463 was obtained, showing that the principal component analysis (PCA) and orthogonal structure to latent structure discriminant analysis (OPLS-DA) projection had good adaptability and predictive power for the potential model (Figures 4A, B). The variables highly correlated with group separation were selected using the projection (VIP) values of the OPLS-DA model and the variable importance of the s-plot. In other words, variables with |p (corr)| ≥ 0.58 and VIP value > 1 were selected as key metabolites and identified with METLIN (Figure 4C). Thus, the results showed seven potential biomarkers were identified by analysis (Table 1). Pathway analysis was performed using MetaboAnalyst, and it was concluded that the metabolites affecting glycine, serine, and threonine metabolism in Chuanxiong were more variable (p < 0.05).

To estimate the underlying mechanism of Chuanxiong in HUVECs under oxidative stress, the mRNA expressions of MAPK and PI3K were detected in each group. The result shows that the activity of MAPK and PI3K, two indicators of angiogenesis, was reduced with t-BHP stimulation and improved after 100 μg/mL Chuanxiong treatment (Figure 5A). In addition, as determined by Western blotting, Chuanxiong increased the expression of the AKT, FOXO1, and RASA1 after t-BHP stimulation (Figure 5B). These data demonstrate that the ability of Chuanxiong to protect against impaired angiogenesis depends on PI3K/AKT/RAS/MAPK signaling.