Kangfuxiaoyan suppository (batch No. 20171225) was obtained from the Sunflower Pharmaceutical Group Co., Ltd. Sodium pentobarbital (batch No. 11011) was obtained from Merck. Mass spectrometry-grade acetonitrile and methanol, and HPLC-grade formic acid were obtained from Thermo Fisher Scientific. The water used for LC/MS analysis was purchased from Watsons.

Animal Sprague Dawley (SD) rats, female, weight 190–210 g, SPF grade, approval number: SCXK (Beijing) 2016-0011, raised in the Animal Experimental Center of Beijing University of Chinese Medicine, the temperature was controlled at 21–25°C, the relative humidity was 55-65%, and the cycle of day and night was 12 h alternately. The rats adapted to the environment for 7 days, during which they drank and ate freely.

Animal Model Establishment The rat model of CPID was established by implanting extraneou materials. The method of establishing CPID model is shown in Supplementary Material. The rats are grouped as follows: 20 rats were randomly divided into the normal group and the sham operation group, the normal group did not do any treatment, the sham operation rats received only a 2 cm incision wound in the middle of the lower abdomen, and then sutured without any treatment, and the other rats were used to establish the rat model of CPID by implanting extraneous materials. Finally, the rats were divided into normal group (n = 10), sham operation group (n = 8), model group (n = 10), KFXYS-dosed group (n = 11), levofloxacin-dosed group (n = 10).

Appropriate amounts of KFXYS and normal saline were added into the clean evaporation dish, heat in a water bath at 50°C until melting, and finally, obtain 0.4 g mL⁻¹ KFXYS solution. KFXYS was given by rectal administration according to the clinical equivalent dose of 408 mg kg⁻¹ for 21 days, and the rats in each group ate normal diet during the treatment.

Sample Collection For the identification of ingredients absorbed into blood of KFXYS and the study of pharmacokinetic, the blood was collected from orbital vein at different time points (15, 30 min, 1, 2, 3, 4, 6, 8, 12, and 24 h) after the last administration of KFXYS in CPID rats, and placed in a centrifuge tube covered with heparin sodium, centrifuged at 1,100 ×g for 10 min at 4°C, and the plasma samples were collected.

For the study of metabolomics, the blood samples were collected from the abdominal aorta in each group rats, and these samples were placed at room temperature for half an hour, centrifuged at 1,100 ×g for 10 min at 4°C, then the serum samples were collected, including eight sham operation group serum samples, 10 model group serum samples, and 11 KFXYS-dosed group serum samples.

Sample Preparation For the identification of ingredients absorbed into blood of KFXYS, mixed the plasma sample from each time point evenly, took 1 ml mixed plasma sample, three times the amount of methanol was added to precipitate the protein, and vortex for 30 s, centrifuged at 13,600 ×g for 15 min at 4°C, and the supernatant was dried with nitrogen. Then the residue was redissolved with 100 μL 70% methanol, centrifuged at 13,600 ×g for 10 min at 4°C, and the supernatant was taken as the plasma sample to be tested in each group. As to the study of pharmacokinetic, the plasma samples at each time point were prepared according to the above method.

For the study of metabolomics, in order to obtain comprehensive information of serum metabolites, the polar part and weak polar part of serum samples were treated respectively. The polar samples were processed as follows: taken 120 μL of the serum sample, added 360 μL methanol to precipitate protein, and vortex for 30 s, static 10 min, centrifuged at 13,600 ×g for 10 min at 4°C. After the 300 μL supernatant was dried with nitrogen, redissolved with 100 μL acetonitrile-water (1:1, v/v), vortex for 30 s, centrifuged at 13,600 ×g for 10 min at 4 °C, and the supernatant was collected. The weak polar samples were processed as follows: taken the serum sample 120 μL, added 300 μL chloroform-methanol (2:1, v/v), vortex for 2 min, static 10 min, centrifuged at 13,600 ×g for 10 min at 4°C, then taken 200 μL lower liquid (chloroform layer), dried with nitrogen, redissolved with 100 μL isopropanol-acetonitrile (1:1, v/v), vortex for 30 s, centrifuged at 13,600 ×g for 10 min at 4°C, and the supernatant was collected.

For the identification of ingredients absorbed into blood of KFXYS, the UPLC-Q-TOF/MS analysis was performed on Waters ACQUITY UPLC I-CLASS liquid phase system equipped with SYNAPT G2-SI high-resolution mass spectrometer. Chromatographic separations were performed at 40°C on ACQUITY UPLC BEH C18 column (2.1 × 100 mm, 1.7 μm, Waters, UK). The mobile phase composed of 0.1% formic acid-water (A) and acetonitrile (B). The elution program was as follows: 0–2 min, 2%–5% B; 2–3.5 min, 5%–8% B; 3.5–6 min, 8%–11% B; 6–8.5 min, 11%–15% B; 8.5–9.5 min, 15%–17% B; 9.5–11 min, 17% B; 11–13.5 min, 17%–20% B; 13.5–15 min, 20%–23% B; 15–17 min, 23%–30% B; 17–20.5 min, 30%–40% B; 20.5–25 min, 40%–60% B; 25–29 min, 60% B; 29–32 min, 60%–100% B; 32–32.1 min, 100%–2% B; 32.1–35 min, 2% B. The flow rate was set to 0.25 ml min⁻¹ and the injection volume was 2 μL.

An electrospray ionization source (ESI) was used both in positive and negative mode. The MS conditions were as follows: capillary voltage, +3 KV/-2.5 KV; cone voltage, 40 kV; ion source temperature, 100°C; cone gas flow, 50 L h⁻¹; desolvation gas temperature, 400°C; desolvation gas flow, 600 L h⁻¹. The Lockmass solution was Leucine encephalin. The collision energy was set as 6 eV (trap) for low-energy scan, and 10–65 eV ramp (trap) for high-energy scan. The mass range was 100-1,200 m/z and the data acquisition mode was 3D data acquisition in Continuum mode.

As to the study of pharmacokinetics of main ingredients, the UPLC-Q-TOF/MS analysis was performed according to the above method, but the data acquisition mode was MRM mode.

For the study of metabolomics, all serum samples were performed on Waters Acquity^TM UPLC analysis system equipped with Xevo^TM G2 Q/TOF tandem quadrupole time-of-flight mass spectrometer, and chromatographic separations were performed at 45°C on ACQUITY UPLC BEH C18 column (2.1 × 100 mm, 1.7 μm, Waters, UK). The UPLC-Q-TOF/MS analysis conditions are shown in Supplementary Material. Quality control (QC) samples were prepared by mixing an equal amount of each sample, and the stability and repeatability of sample analysis were monitored by analyzing QC samples after every 10 samples.

Based on the previous research, the identification steps of the ingredients absorbed into blood of KFXYS were as follows: 1) Summarizing the possible metabolic pathways of various chemical components in vivo by comparing with some representative ingredients and tracking the literature. 2) The prototype components and metabolites in the KFXYS-dosed plasma were discovered by comparing with the KFXYS-treated and KFXYS sample. Then the ingredients absorbed into blood of KFXYS were comprehensively identified based on the mass fragmentation pattern, characteristic fragments and chromatographic retention behavior.

The optimized ion pairs of the main components and parameters of cone voltage and capillary voltage are shown in Table 1.

The raw files were imported into Progenesis QI software for denoising, peak identification, peak alignment, normalization and other operations. The raw data were converted into a data matrix containing t _R -m/z ion pairs, sample names and peak intensities, and then the matrix was imported into multivariate statistical software EZinfo2.0 (Waters Corporation, Milford, MA, United States) for principal components analysis (PCA) and orthogonal partial least-squares discriminant analysis (OPLS-DA). The metabolites were identified by searching and comparing the mass fragment information and accurate mass number with data from METLIN (http://metlin.scripps.edu/), HMDB (http://www.hmdb.ca/) and KEGG (http://www.genome.jp/kegg/) databases. The MetaboAnalyst (https://www.metaboanalyst.ca/) was used for metabolic pathway analysis.

The potential targets of components absorbed into blood of KFXYS were obtained from PharmMapper (http://www.lilab-ecust.cn/pharmmapper/) and TCMSP (https://old.tcmsp-e.com/tcmsp.php) databases. The disease targets related to CPID were searched from OMIM (https://omim.org/) and GeneCards (https://www.genecards.org/) databases. The target protein names were corrected to official gene names by Uniprot (https://www.uniprot.org/). Through the Venn diagram, the components targets and the disease targets intersected, and the intersectional targets were considered potential therapeutic targets of KFXYS in CPID. R-packet ClusterProfiler was used to analyze the potential targets for GO and KEGG enrichment analysis.

By comparing the potential pathways enriched by the network pharmacological targets with the metabolic pathways enriched by differential metabolites, the common metabolic pathways which were regarded as the key pathways were screened out. The KEGG database was used to obtain the proteins contained in the key metabolic pathways to screen the key targets from the potential targets of network pharmacology. Through the literature research to reveal the possible mechanism of KFXYS in the treatment of CPID.

In this study, in order to determine whether KFXYS reverse inflammatory indexes to exert anti-inflammatory effect, we measured the level of IL-1, IL-6, IL-10, TNF-α in the uterine tissue homogenate. As shown in Figure 2A, compared with the sham operation group, the uterus index of the model group was significantly increased (p < 0.05), and the results indicated that the CPID model was successfully established. Compared with the model group, the uterus index decreased significantly after administration of KFXYS (p < 0.05), indicating that KFXYS can effectively improve the degree of uterine swelling in CPID model rats. In the uterine tissue homogenate, the level of IL-6 in the model group decreased significantly (p < 0.05), and the level of IL-1 increased significantly (p < 0.01). After administration of KFXYS, the changes of these indicators were significantly improved. The result shows that KFXYS has the effect of reversing inflammatory indexes (Figures 2B–E).

Eighty chemical components with high response and 43 chemical components with weak response from KFXYS were identified by high-resolution mass spectrometry (Supplementary Figure S1), and the components information is shown in Supplementary Table S1.

We developed a new method of real-time binary comparison to discover the ingredients of KFXYS absorbed into the blood (Figure 3), a total of 60 KFXYS-derived compounds were identified, including 10 prototype compounds (Table 2) and 50 metabolites (Supplementary Table S2). According to the metabolic rules established by representative reference substances, we consider that these 60 ingredients were mainly derived from SFR, AH, TH and VH, mainly including alkaloids, flavonoids and organic acids etc.

In order to explore the drug-like properties of compounds, we investigated the pharmacokinetics of the KFXYS-derived components with good peak shape and considerable content in the blood. It is reported that the alkaloid in SFR has definite anti-inflammatory activity, such as matrine (Pu., et al., 2016), oxymatrine (Wang et al., 2015), sophocarpine (Jiang et al., 2018b); sophoridine (Wang et al., 2021), and oxysophocarpine (Yang et al., 2015). Therefore, we studied the pharmacokinetics of the alkaloids from SFR, including matrine, sophocarpine, sophoridine, oxymatrine and oxysophorine, formononetin from SFR, aloin A from aloe, and esculetin from VH and TH. The results showed that matrine, sophocarpine, aloin, esculetin-O-glucuronide, 7,4′-dihydroxyisoflavone-O-glucuronide (7-hydroxy-4′-methoxyflavonol-CH₂+C₆H₈O₆), and 4′-methoxyisoflavone-7-O-glucuronide had good ADME behavior in vivo (Figure 4). Due to the low content of other components in blood, we only discovered these six components showed good ADME behavior in vivo. Thus, we speculated that matrine, sophocarpine, aloin, esculetin, and 7-hydroxy-4′-methoxyisoflavone and their metabolites may be the potential pharmacological components of KFXYS.

By dividing the serum sample into polar part and weak polar part, more comprehensive metabolic information was obtained, and the BPIs in positive and negative ion mode are shown in Supplementary Figure S2. QC samples in positive and negative ion mode showed that both the instrument precision and method precision can carry out good metabolic profile analysis.

The serum samples of sham operation group, model group and KFXYS-dosed group were subject to multivariate statistical analysis. In order to distinguish and visualize the differences between different groups, PCA and OPLS-DA analyses were performed. PCA diagrams showed that the model group and sham operation group are clearly separated, and the KFXYS-dosed group tends to be adjusted to the sham operation group ( Figures 5A,B). The serum samples of sham operation group and model group were used for Student’s t-test (t-test) and OPLS-DA analysis (Figures 5C,D) to screen the metabolites with significant differences (p < 0.05, VIP >1) (Figures 5E,F), respectively. Among these differential metabolites (p < 0.05, VIP > 1), those with the opposite trend in the KFXYS group and model group were defined as final differential metabolites. As shown in Table 3, a total of 16 final differential metabolites were screened and identified.

To investigate the correlation between these metabolites and the efficacy of KFXYS, Pearson correlation analysis was performed using the inflammatory indexes in uterine tissue homogenate including IL-6, IL-10, IL-1, and TNF-α, and 16 differential metabolites to further screen metabolites (Figure 5G). The Pearson correlation analysis revealed that inflammatory indexes were significantly correlation to five metabolites, including Leukotriene A4, 5-Hydroxyindoleacetic acid, Ornithine, Arginine, and PC (20:1 (11Z)/20:4 (8Z,11Z,14Z, 17Z)), which indicated that KFXYS may play a role in the treatment of CPID by regulating these directly related metabolites.

To explore the metabolic pathways related to the efficacy of KFXYS, the five metabolites were imported into MetaboAnalyst. Nine corresponding pathways were constructed, including Arginine biosynthesis, Arachidonic acid metabolism, Arginine and proline metabolism, Linoleic acid metabolism, alpha-Linolenic acid metabolism, Glutathione metabolism, Glycerophospholipid metabolism, Tryptophan metabolism, and Aminoacyl-tRNA biosynthesis (Figure 6).

Based on the ingredients absorbed into blood with good ADME behavior, 422 drug targets were predicted from the PharmMapper database and TCMSP database. Combined with GeneCards and OMIM database, 1973 disease targets of CPID were obtained. The 185 overlapping targets of disease targets and drug targets are potential targets for KFXYS in treating CPID (Figure 7A). GO enrichment analysis was used to discover the underlying biological process (BP), cellular component (CC) and molecular function (MF) of the 185 potential targets. BP analysis showed that the treatment of KFXYS was mainly related to response to negative regulation of apoptotic signaling pathway. CC analysis showed that the treatment process of KFXYS was related to cellular components such as membrane raft, membrane microdomain and vesicle lumen. MF analysis showed that the treatment of KFXYS was mainly related to molecular functions such as immunoglobulin binding and cytokine receptor binding (Figure 7B). In KEGG enrichment analysis of potential targets, 176 pathways were obtained with the critical value of p < 0.05. The first 10 pathways were taken as shown in Figure 7C.

The shared pathways of metabolomics and network pharmacology were arginine and proline metabolism, glutathione metabolism, which may be the key pathways of KFXYS in the treatment of CPID. The targets obtained by network pharmacology were screened by the proteins in these two key pathways, and the obtained targets were regarded as key targets. Finally, nine key targets were identified. The specific target information is shown in Table 4. According to the previously obtained components absorbed into blood which had good behavior, nine key targets, and two key metabolic pathways, a “TCM prescription-single herb-compound-key targets-key pathways-disease” network of KFXYS in the treatment of CPID was constructed (Figure 8). In this network, it can be seen that KFXYS is characterized by multi-components, multi-targets and multi-pathways in the treatment of diseases. For example, both aloin and formononetin in KFXYS can treat CPID by acting on ARG1, NOS2, and NOS3 proteins in the arginine and proline metabolism.