SR and AM were purchased from Bozhou Chinese Medicine Exchange in China, SR originates from Inner Mongolia (China) and AM originates from Jiangxi (China). N′-Nitro-L-arginine-methylesterhydrochloride (L-NAME) (51,298-62-5) was obtained from Yuanye Biotechnology (Shanghai, China). Human chorionic extrachorionic trophoblast cell (HTR-8/SVneo) (lh-h129) was purchased from BioHarbor Biotechnology Company (Shanghai, China). Antibodies against PI3K (YM3503), AKT (T0185), and P-AKT (YP0006) were purchased from ImmunoWay Biotechnology Company (Texas, United States). eNOS (AB300071) was obtained from Abcam (Cambridge, UK). C57BL/6J mice were purchased from Nanjing Jicui Pharmachem Laboratory Animal Technology (Nanjing, China). LY294002(HY-10108) was purchased from MedChemExpress (New Jersey, United States).

100 g SR and 150 g AM (traditional medicine ratio 2:3), add 8 times water after washing, decoct for 60 min, while constantly stirring, filtering, retain the decoction, repeat three times. The three decoctions were mixed, concentrated to the extract, and freeze-dried for 48 h to obtain the extract of SA for use.

This extract was preliminarily analyzed using the HPLC-Q Exactive-Orbitrap-MS (Thermo Fisher Scientific, Waltham, MA, United States). The chromatographic separation was performed using the Ultimate 3000 HPLC system (2.1 mm × 150 mm x 1.8 um) column (Welch). The mobile phase comprised (A) 0.1% methanol in pure waterand (B) methanol with gradient elution: 0–1 min, 98% (A); 1–5 min, 98%–80% (A); 5–10 min, 80%–50% (A); 10–15 min, 50%–20% (A); 15–20 min, 20%–5% (A); 20–27 min, 5% (A); 27–28 min, 5%–98% (A); 28–30 min, 98% (A). The flow rate was 0.3 mL/min and the column temperature was maintained at 40°C. The samples were maintained at room temperature and the injection volume was 5 μL. The Q Exactive mass spectrometer equipped with a heated electrospray ionization interface was operated under both electrospray ionization (ESI) negative and ESI positive modes. The instrument was calibrated with the calibration solutions provided by the manufacturer. Data were acquired using Compound Discoverer 3.3 software (Thermo Fisher Scientific, Waltham, MA, United States), and the preliminary sorting data were searched and compared in the database (mz Cloud). The other parameters were set as follows: The source parameters, spray voltage of 3.2 kV (+)/3.2 kV (−); capillary temperature, 300°C; auxiliary gas heater temperature, 350°C, sheath gas pressure, 40 arb; auxiliary gas pressure, 15 arb.

L-NAME-induced hypertensive pregnant rats serve as a well-established animal model for preeclampsia. Pregnant mice were randomly divided into six groups (n = 6): control, model (L-NAME group), Aspirin (positive group, i. g, 15.2 mg/kg/d) + L-NAME, SA low-dose (i.g, 95 mg/kg/d) + L-NAME, SA middle-dose (i.g, 135 mg/kg/d) + L-NAME, SA high-dose (i.g, 175 mg/kg/d) + L-NAME. According to the Chinese Pharmacopoeia, the recommended daily dosages for adults are 3–10 g for Scutellaria baicalensis and 6–12 g for Atractylodes macrocephala (National Pharmacopoeia Commission, 2020). Based on the total amount of the compound formula and an extraction yield of 8.5%, the equivalent clinical extract dosage for adults was calculated to range from 95 to 270 mg/kg/day. Using the standard body surface area conversion factor of 9.01 to convert this dosage to mice, and considering toxicity data reported in the literature (Liu et al., 2014; Song et al., 2020; Yang et al., 2021), the low, medium, and high doses were finally set at 95, 135, and 175 mg/kg, respectively. Except for the control group, the other groups were given L-NAME (I.H, 50 mg/kg/d) subcutaneously daily from day 8 of pregnancy. The Aspirin and SA group were orally administered the respective compounds once daily, and the control and model groups received equal volume of saline. Each treatment was administered continuously on days 7 and 17 of pregnancy. Blood pressure was measured in each group on days 7 and 17 of pregnancy. Total protein was measured in urine on days 7 and 18 of pregnancy using the Total Urine Protein Test (UTP colorimetric method). Female rats were euthanized on day 18 of pregnancy, and placenta, fetus, and kidney tissue were collected. The study was approved by the Experimental Animal Center of Fuyang Normal University and conducted by the national and institutional guidelines regarding animal experiments. The Institutional Animal Care Committee of the university reviewed and approved the study protocol (Approval Number, K-2022-0529-1).

The screening of active compounds and identification of targets involved utilizing the Traditional Chinese Medicine Systems Pharmacology Database (TCMSP, http://lsp.nwu.edu.cn/) and the Chinese Academy of Sciences chemistry database (CASC, http://www.organchem.csdb.cn/scdb/main/) to identify active compounds from SR and AM, focusing on those with oral bioavailability (OB) ≥ 20% and drug-likeness (DL) ≥ 0.1, along with quality control compounds. Subsequently, details on human genes associated with preeclampsia were gathered from three databases: the Comparative Toxicogenomics Database (CTD, https://www.pharmgkb.org/), OMIM (https://www.omim.org/), and GeneCards (https://www.genecards.org/), with gene identifiers converted into official gene symbols using an R-script linked to PubMed (https://www.ncbi.nlm.nih.gov/gene/), ensuring only human genes were included for further analysis. Common target genes for the drugs and diseases were then obtained, leading to the construction of a network of compound-common targets, while protein-protein interaction (PPI) networks for the compound targets and preeclampsia targets were visualized using Cytoscape 3.7.1 software. Finally, bioinformatic analysis was performed through KEGG enrichment analysis using the Database for Annotation, Visualization, and Integrated Discovery (DAVID, https://david.ncifcrf.gov, v6.8), identifying pathways with significant changes (P < 0.05) for further analysis, selecting genes that significantly regulated these pathways for gene-pathway network analysis, and identifying key target genes from the resulting network.

The PDB file of PI3K was download from database (https://bitterdb.agri.huji.ac.il/dbbitter.php), Atractylenolide I and Baicalin mol2 files downloaded in PubChem database (https://pubchem.ncbi.nlm.nih.gov/), analysis software for AutoDockTools-1.5.7, visualization tools for PyMOL-2.6.0.

To assess histopathological changes, placenta and kidney tissue were fixed in 4% formaldehyde and embedded in paraffin. Sequential serial sections were deparaffinized and stained with hematoxylin and eosin (HE) for histological examination.

The samples were lysed in a RIPA buffer solution involving 1% protease inhibitor. After centrifugating, supernates were gathered. Determination of protein level was conducted via the BCA protein assay kit. The denatured protein was isolated via 10% SDS-PAGE and shifted to PVDF membrane through electrophoresis. Membranes were blocked with 5% skim milk and then co-incubated overnight under 4°C with primary antibodies (including PI3K, AKT, P-AKT and eNOS). After washing, membranes were incubated with secondary antibodies for 2 hours at ambient temperature. Quantification of the protein bands was performed via ImageJ.

HTR–8/SVneo cells were cultured in RPMI–1640 with 10% FBS, at 37°C with 5% CO₂. Cells were plated at 5 × 10⁴ cells/well in 96-well plates for 24 h and then treated with L-NAME for 2 h. Each well was added SR and atractylodis extract, baicalin, atractylenolide I, and then added MTT after 12 h and incubated at 37°C for another 5 h. The formazan crystals were dissolved in DMSO. The absorbance was measured at 490 nm.

Cells were plated at 5 × 10⁴ cells/well in 96-well plates for 24 h and then treated with L-NAME for 2 h. Each well was added SA, baicalin, atractylenolide I. The cell supernatant was aspirated and centrifuged to remove solid impurities from the supernatant. The absorbance of NADH was measured by spectrophotometer at 340 nm at room temperature.

Cells were plated at 5 × 10⁴ cells/well in 6-well plates for 48 h and then scratches were made on the bottom of the dishes with a lance tip. The well plates were divided into six groups, excluding the control group, the other five groups were added with the same concentration of L-NAME in each well, incubated in the incubator for 2 h. Each well was added SA extracts, baicalin, atractylenolide I. and compared of scratches photographed under microscope at 0 h, 12 h, and 24 h.

Cells were plated at 5 × 10⁴ cells/well in 6-well plates with spread for 48 h. Cells were fixed by soaking in ethanol and permeabilized using Triton X-100. The cells were restored to room temperature and incubated with appropriate fluorescence-conjugated secondary antibodies and DAPI. The images were captured with a Nikon confocal microscope and calculated with NIH ImageJ software.

The method of stable isotope dimethyl labeling was used to study the quantitative proteomic analysis of placenta. Analysis was performed based on published technical protocols using a Q-Exactive Orbitrap mass spectrometer coupled to an Ultimate 3000 LC system (Zhu et al., 2024). Proteins with an average ratio ≥1.5 or ≤0.67 in all samples were selected for KEGG pathway analysis and GO analysis by DAVID (https://david.ncifcrf.gov/). The selected proteins were imported into the STRING (http://stringdb.org/) database to construct a protein-protein interaction (PPI) network.

SPSS 22.0 (IBM Corp., Armonk, NY, United States) and GraphPad Prism 8.2.0 (GraphPad Software Corp., San Diego, California, United States) were used for statistical analysis. Statistical analysis was performed with one-way ANOVA with Dunn’s test for multiple comparisons, while unpaired Student’s t-test for two individual comparisons. p < 0.05 is considered significant. Mean ± standard deviation (SD) is the expression of the data. The proteomic data were deposited to the ProteomeXchange consortium (https://proteomecentral.proteomexchange.org) via the iProX partner repository (Ma et al., 2019; Chen et al., 2022) with the dataset identifier PXD038535.

This extract was preliminarily analyzed using the HPLC-Q Exactive-Orbitrap-MS. The data acquired from HPLC-Q Exactive-Orbitrap-MS instrument were initially processed using Compound Discoverer 3.3 (CD 3.3, Thermo Fisher) and subsequently compared against the mzCloud database. In total, 636 compounds were identified in mz Cloud, with 132 compounds achieving a comprehensive score of mz Cloud best match exceeding 90 (Supplementary Table S1). The fingerprint of SA is attached in Figure 1. The major bioactive components commonly identified in S. baicalensis and Atractylodis Macrocephalae Rhizoma, such as Wogonin, Atractylenolide I, Baicalin, and Baicalein, were clearly annotated in the fingerprint chromatogram (Figure 1).

This part first evaluated the effects of SA on preeclampsia. Results showed that the systolic blood pressure and urinary protein levels of mice in model group were significantly higher than those in control group after 10 days of L-NAME administration (i.e., the 17th day of pregnancy (gd17)). However, SA treatment could decrease the systolic blood pressure and urinary protein of mice (Figures 2a–d). The organ index test showed that the weight of placenta and fetal rats in the SA treatment group was significantly higher than that in the model group (Figures 2e,f), suggesting that they can improve the growth of placenta and fetus. Further HE staining showed that the renal tubular epithelial cells in the control group were round and plump, arranged regularly, and the renal tubular cells were coated with nuclei. In the model group, the renal tubules were severely vacuolated (black arrow), a large number of nuclei were exposed and shed (red arrow), and some glomeruli were swollen. In the SA treatment group, vacuolization of renal tubules was reduced, and only a few nuclei were exposed and shed (red arrow) (Figure 2g). HE staining of placental tissue showed that in the control group, the boundaries of each layer were clear, the trophoblast cells were arranged regularly, the morphology was normal, and a large number of red blood cells could be seen. In the model group, the boundary between placental trophoblast layer and labyrinth layer was unclear, necrotic foci (red arrows) were seen in many places, trophoblast nuclei were pyknotic, deeply stained and fragmented, calcified foci (green arrows) were seen in necrotic foci, and the surrounding trophoblasts were arranged irregularly; Red blood cell content decreased. After treatment with SA, the above injury can be significantly improved (Figure 2h). As inflammation is a critical pathogenic mechanism and physiological manifestation of preeclampsia, we assessed the activation of the NF-κB pathway by analyzing the mRNA levels of key inflammatory factors, including IL-6, IL-1β, TNF-α, NF-κB-p50, and NF-κB-Rel A. The results demonstrate that SA can significantly modulate the expression levels of these inflammatory factors in the placental tissues of preeclampsia mice (Figure 2i).

After clarifying the protective effect of SA on preeclampsia, this part continued to observe its protective effect using HTR-8/SVneo cells. To demonstrate the pharmacological rationale for selecting specific concentrations, dose-response cell vability experiments firstly were conducted. SA pairs is safe for HTR-8/SVneo cells within the dose range of 0–4 mg/mL, as shown in Figure 3a. Such three doses (1,2,4 mg/mL)were selected for experiments of cell proliferation and migration. The results showed that compared with the control group, L-NAME treatment could lead to decreased cell proliferation, increased LDH secretion, and slowed wound healing. When treated with SA, compared with the L-NAME group, different doses of SA can improve the cell proliferation rate, reduce LDH secretion, and promote the migration level. After adding the PI3K inhibitor LY294002 to the HTR-8/SVneo cells, it was found that the promoting proliferation effect of SA on L-NAME-stimulated HTR-8/SVneo cells was significantly weakened (Figures 3c–f).

In order to further verify the effect of SA on the proliferation and migration of HTR-8/SVneo cell, we selected baicalin and atractylenolide, the main active and quality control components of SA, respectively. Results showed that its effect on the proliferation, LDH secretion and migration of HTR-8/SVneo cells was the same as that of SA (Figure 4).

Through the Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform (TCMSP), we identified 143 chemical constituents in SR and 55 in AM. By applying the selection criteria of oral bioavailability (OB) ≥ 20% and drug-likeness (DL) ≥ 0.1, along with the inclusion of pharmacologically active components commonly recognized in SR and AM, such as baicalin, baicalein, atractylenolide I, atractylodin, and atractylolide III, we ultimately obtained 38 active constituents from SR and 11 from AM. The effective components’ targets were mapped against disease targets using Venny 2.1.0, yielding 189 common genes (Figures 5a,b). A protein-protein interaction (PPI) network diagram was constructed for these 189 effective targets. Using Cytoscape software, 39 core targets were identified, with key target genes including TNF, IL1B, AKT1, and PTGS2, among others (Figures 5c,d). The aforementioned 189 targets were imported into the DAVID database for KEGG analysis. Given that this study primarily focuses on preeclampsia, signal pathways unrelated to the pathogenesis of this disease, such as those associated with cancer, were excluded. The top 10 signal pathways were visually processed, which mainly included the TNF signaling pathway, the PI3K/AKT signaling pathway, and the Relaxin signaling pathway (Figure 5e). Considering the significant role of the PI3K/AKT signaling pathway in the development of vascular diseases, further validation of this pathway will be pursued in subsequent studies.

To comprehensively elucidate the biological reaction mechanisms, including pathogenesis and pharmacodynamic effects. Proteomes extracted from placental tissues of pregnant rats in control, model, and medium-dose SA treatment groups were subjected to stable isotope dimethyl labeling followed by HPLC-MS/MS analysis.

Protein screening criteria were established as follows: (1) proteins with ≤0.67-fold change in normal vs. model group and ≥1.67-fold change in SA-treated vs. model group; (2) proteins with ≥1.67-fold change in normal vs. model group and ≤0.67-fold change in SA-treated vs. model group. The analysis revealed 285 significantly upregulated and 111 downregulated proteins in the model group. SA treatment resulted in the regulation of 270 downregulated and 170 upregulated proteins, which was illustrated by venn diagrams in Figure 6a (The tables of the top 20 differentially expressed proteins, including their fold changes and adjusted P-values, were listed in Supplementary Table S2, S3), with 39 targets suppressed after being pathologically overexpressed, and (right) 11 targets restored after pathological suppression. A four-quadrant plot was used to analyze the potential drug target proteins identified in venn diagrams (Figure 6b).

Furthermore, we conducted protein-protein interaction analysis using the STRING database for the 50 candidate proteins along with key pathway components PI3K, Akt, and eNOS. The network analysis identified eight proteins interacting with the PI3K/Akt/eNOS pathway, among which superoxide dismutase-2 (SOD2), apolipoprotein B (ApoB), cadherin-5 (CDH5), and echinoderm microtubule-associated protein-like 4 (EML4) demonstrated direct interactions (Figure 6c). The raw data of the eight key proteins identified in the PPI network were organized into bar charts to visually demonstrate the reversing effect of SA on their expression levels (Figure 6d). These findings suggest that SA pairs may exert therapeutic effects on preeclampsia through the PI3K/Akt/eNOS signaling pathway.

Molecular docking results further showed that atractylenolide I and baicalin had a good binding activity with PI3K, with the binding energy were −5.72 and −3.6 kcal·mol^-1, respectively ((Figure 7a). atractylenolide I and baicalin can form hydrogen bonds with Thr-827, Glu-800, Glu-209, Lys-213, Lys-883, etc. amino acid residues (Figure 7a). Besides, the expression levels of PI3K, p-Akt/Akt and eNOS in the model group were significantly downregulated, while the expression levels of them were significantly upregulated after treatment with SA in the in vivo experiment (Figure 7c). In the in vitro experiment, immunofluorescence staining results showed that the fluorescence intensity of eNOS in model group decreased, while the fluorescence intensity of baicalin, atractylenolide I, and SA treatment groups increased, suggesting that the SA treatment can reverse the reduction of eNOS expression caused by L-NAME (Figure 7b). Western Blot analysis showed that baicalin, atractylenolide I, and SA could reverse the decreased expression of PI3K, p-Akt/Akt and eNOS caused by L-NAME (Figures 8a–c). These results also showed that SA could upregulate PI3K, p-Akt/Akt pathway, increase eNOS expression, ultimately alleviate hypertension in the PE model.