Preamble

Study on the Pharmacological Mechanism of Xihuang Pill and its Treatment of Breast Cancer Based on Nontargeted Metabonomics

Yifan Su¹², Xiaohui Zhao¹², Dehui Li¹*, Jiao Liu¹² & Xu-kuo Liu¹²
¹Hebei Province Hospital of Chinese Medicine; The First Affiliated Hospital of Hebei University of Chinese Medicine, Shijiazhuang, 050011, China
²Graduate School of Hebei University of Chinese Medicine, Shijiazhuang 050091, China

Yi-fan Su and Xiao-hui Zhao contributed equally to this work and share first authorship. Yi-fan Su and Xiao-hui Zhao conducted most of the research, performed statistical analysis and wrote the manuscript. De-hui Li designed the study and reviewed the manuscript. Jiao Liu and Xu-kuo Liu helped in the animal study and recorded the data.

Abstract: [Objective] To investigate the main differential metabolites of Xihuang Pill (XHP) in rat serum and the mechanism of related metabolic pathways in breast cancer treatment. [Methods] Liquid chromatography-mass spectrometry (LC-MS) was used for qualitative and quantitative analysis of metabolites in XHP drug serum and blank serum groups, with sample correlation heat maps and multivariate statistical analysis methods employed to compare metabolic differences between the two groups. Metabolites were analyzed through cluster analysis, Variable Importance in Projection (VIP) analysis, Kyoto Encyclopedia of Genes and Genomes (KEGG) compound classification, and KEGG functional and enrichment topology analysis. [Results] LC-MS identified a total of 765 metabolites in the XHP drug serum group and 697 metabolites in the blank serum group. VIP analysis screened the top 30 serum differential metabolites, which included abscisic acid, quillaic acid, 2,2-Bis(4-hydroxyphenyl)-1-propanol, Corey PG-Lactone Diol, (S)-Naproxen, among others. KEGG compound classification revealed that most metabolites in XHP were categorized as phospholipids and amino acids, steroid hormones, and carboxylic acids. The main KEGG functional pathways involved were lipid metabolism, amino acid metabolism, and cancer overview. KEGG enrichment and topology analysis indicated primary involvement in the steroid hormone biosynthesis pathway and beta-alanine metabolism pathway. [Conclusions] The main differential metabolite of XHP in rat serum may be abscisic acid. XHP may exert its pharmacological effects on breast cancer by regulating the steroid hormone biosynthesis pathway to modulate estrogen and progesterone levels, and by modulating the beta-alanine metabolism pathway to induce cancer cell apoptosis.

Keywords: Xihuang Pill; LC-MS; Metabolomics; Metabolite analysis

1. Introduction

Xihuang Pill (XHP) originates from the Waike Zhengzhi Quansheng Ji compiled by Wang Weide during the Qing Dynasty. Its formulation consists of NIU HUANG, SHE XIANG, RU XIANG, and MO YAO¹. As a classical prescription, XHP has been used clinically to treat various diseases, demonstrating good curative effects on breast cancer, liver cancer, gastric cancer, colon cancer, lung cancer, and other tumors². Modern pharmacological research shows that XHP possesses antitumor, anti-inflammatory, antibacterial, and immune-enhancing effects³. Experiments have demonstrated that XHP can inhibit tumor cell proliferation while promoting apoptosis, prevent tumor invasion and metastasis, inhibit tumor angiogenesis, and regulate the tumor microenvironment⁴,⁵. In our earlier work, we primarily investigated the mechanism of XHP in treating precancerous lesions of breast cancer. Using 7,12-dimethylbenzanthracene (DMBA) combined with estrogen and progesterone to induce a rat model of breast cancer precancerous lesions, we intervened with XHP and found that XHP may promote apoptosis of precancerous cells and reduce hyperplasia of vascular endothelial cells by regulating gene and protein expression related to the PI3K/AKT/mTOR pathway, thereby inhibiting the progression of breast cancer precancerous lesions⁶. Additionally, XHP may treat precancerous lesions by improving microcirculation in affected rats⁷, and it can block and reverse the histopathological changes in breast tissue sequentially triggered by DMBA combined with estrogen and progesterone, with its mechanism potentially related to regulation of NF-κB protein expression in breast tissue⁸. Clinical studies have shown that XHP combined with chemotherapy for breast cancer can enhance clinical efficacy, improve quality of life, and reduce adverse reactions caused by chemotherapy⁹.

Metabolomics is used to detect dynamic and comprehensive changes in metabolites, obtain meaningful biomolecular information, and understand biological processes under drug actions¹⁰. This approach helps elucidate the relationship between traditional Chinese medicine (TCM) and diseases from a molecular biology perspective and study their mechanisms of action, which is conducive to better understanding TCM efficacy in vivo. Metabolomics is now widely applied in screening active components of TCM and studying the pharmacokinetics and mechanisms of action of TCM, providing a new method for TCM modernization¹¹. Due to the multicomponent, multitarget, and multimechanism characteristics of XHP treatment, its complexity and the diversity of in vivo processes make modern research challenging, and the metabolites of XHP remain unclear. To address this problem, this study applied LC-MS technology to conduct nontargeted metabolomic analysis of XHP components to explore its main differential metabolites, classify these metabolites, and evaluate their relationships with diseases. Investigating the differential metabolites of XHP is of great significance for comprehensively understanding its pharmacological mechanisms in breast cancer treatment.

2. Materials and Methods

2.1 Instruments and Equipment

The study utilized a NewClassic MS electronic balance (NewClassic MF MS105DU, METTLER TOLEDO), freezing centrifuge (Centrifuge 5424 R, Eppendorf), Temperature Control Ultrasonic Cleaner-10L (SBL-10TD, Ningbo Xinzhi Biological Technology Co., Ltd.), multi-sample freeze grinder (Wonbio-96c, Shanghai Wanbai Biotechnology Co., Ltd.), benchtop rapid centrifugal instrument (JXDC-20, Shanghai Jingxin Industrial Development Co., Ltd.), nitrogen concentration dryer (LNG-T88, Taicang Huamei Biochemical), UHPLC liquid chromatography system (Vanquish Horizon system, Thermo Scientific), and mass spectrometer (Q-Exactive, Thermo Scientific).

2.2 Drugs and Reagents

XHP (Zhejiang Tianyitang Pharmaceutical Co., Ltd., lot No.: 1703011), methanol (Fisher Chemical), acetonitrile (Fisher Chemical), formic acid (CNW), water (Fisher Chemical), 2-propanol (Merck), and 2-Chloro-L-Phenylalanine (Adamas-beta) were used in this study.

2.3 Animals and Groups

Twelve SPF 6-week-old female Sprague-Dawley (SD) rats (weighing 180 ± 20 g) were purchased from Hebei Experimental Animal Center (Shijiazhuang, Hebei, China, license number: 1705351). Animal experiments were approved by the Medical Ethics Committee of Hebei University of Chinese Medicine and carried out in accordance with the Animal Welfare Guidelines of the Medical Ethics Committee of Hebei University of Chinese Medicine (ethical review number: DWLL2019036). The rats were divided into two groups: an XHP drug serum group and a blank serum group, with six rats per group. All rats were allowed free access to food and water, and experiments were conducted after one week of routine feeding.

2.4 Preparation of Serum Samples

2.4.1 Preparation of XHP Drug Serum

Rats in the XHP-containing serum group were administered XHP Chinese medicine solution daily. XHP was ground, crushed, and made into a suspension with distilled water at a concentration of 0.25 g/ml, with an administration volume of 5 ml/kg given twice daily, resulting in a cumulative daily dose of 2.5 g/kg. Dose conversion between humans and rats was performed according to the Methodology of Pharmacological Research of Traditional Chinese Medicine edited by Chen Qi. After three days of administration, rats were fasted for 12 hours prior to the final gavage on day 4. One hour after the last administration, rats were anesthetized with pentobarbital sodium (200 mg·kg⁻¹, ip), blood was collected from the abdominal aorta, and serum was separated in sterile tubes. The serum was inactivated at 56 °C for 30 minutes and stored at -80 °C for later use.

2.4.2 Preparation of Blank Serum

Rats in the blank serum group were administered distilled water intragastrically daily (with the same dosage volume as the XHP drug serum group) for three consecutive days. Blood collection and serum separation were performed at the same time points as the XHP drug serum group, using identical serum preservation methods.

2.5 Serum Grouping and Sample Processing

Two hundred microliters of serum from each rat in both the XHP serum and blank serum groups were extracted for LC-MS metabolomics analysis. The drug serum from the XHP-containing group was designated as the experimental group (sy), which was further divided into experimental subgroups 1-6 (sy1-sy6). The serum from the blank group was designated as the control check (ck), divided into control subgroups 1-6 (ck1-ck6). LC-MS nontargeted metabolomics analysis and multivariate statistical analysis were performed to compare metabolic differences between the experimental and control groups.

2.5.1 Sample Treatment

We accurately transferred 100 μL serum from each group into a 1.5 mL centrifuge tube and added 400 μL extraction solvent (methanol:acetonitrile = 1:1 (v:v)) containing 0.02 mg/mL internal standard (L-2-chlorophenylalanine). After vortex mixing for 30 seconds, low-temperature ultrasonic extraction was performed for 30 minutes (5 °C, 40 kHz). The sample was then placed at -20 °C for 30 minutes and centrifuged for 15 minutes (13,000 × g, 4 °C). The supernatant was collected, dried with nitrogen, and reconstituted with 120 μL reconstitution solution (acetonitrile:water = 1:1). After vortex mixing for 30 seconds and low-temperature ultrasonic extraction for 5 minutes (5 °C, 40 kHz), the sample was centrifuged for 10 minutes (13,000 × g, 4 °C). The supernatant was transferred to an injection vial with an inner cannula for analysis. Additionally, 20 µL of supernatant from each sample was pooled to create a quality control sample.

2.5.2 Chromatographic Conditions

Chromatographic separation was performed on an ACQUITY UPLC HSS T3 column (100 mm × 2.1 mm i.d., 1.8 µm; Waters, Milford, USA). Mobile phase A consisted of 95% water + 5% acetonitrile (containing 0.1% formic acid), and mobile phase B consisted of 47.5% acetonitrile + 47.5% isopropanol + 5% water (containing 0.1% formic acid). The flow rate was 0.40 mL/min, injection volume was 2 μL, and column temperature was 40 °C.

2.5.3 Mass Spectrometry Conditions

Samples were analyzed using electrospray ionization with mass spectrometry signals collected in both positive and negative ion scanning modes. The scan range was 70-1050 m/z; sheath gas flow rate was 40 arb; auxiliary gas flow rate was 10 arb; heater temperature was 400 °C; capillary temperature was 320 °C; spray voltage was +3500 V in positive mode and -2800 V in negative mode; s-Lens RF Level was 50; normalized collision energy was 20, 40, and 60 eV; resolution (Full MS) was 70,000; and resolution (MS2) was 17,500.

2.5.4 Preparation of Quality Control Samples

Quality control (QC) samples were prepared by mixing equal volumes of extracts from all samples. Each QC sample volume matched that of analytical samples and was processed and tested identically. QC samples were injected every four samples to monitor the stability of the entire detection process, resulting in three quality control samples for this experiment.

2.6 Data Analysis

Thermo Fisher's Ultra High-Performance Liquid Chromatography tandem Fourier Transform mass spectrometry UHPLC-Q Exactive system was used for chromatographic peak identification. Progenesis QI software (Waters Corporation, Milford, USA) performed baseline filtering, peak identification, integration, retention time correction, and peak alignment to obtain information including mass-to-charge ratio (m/z), retention time, and peak area. The software was then used to search and identify characteristic peaks, matching MS and MS/MS spectral information against metabolic databases with an MS mass error threshold of less than 10 ppm. Metabolites were identified based on secondary MS spectrometry matching scores. The primary databases used were http://www.hmdb.ca/, https://metlin.scripps.edu/, along with other mainstream public databases and in-house databases. The normalized data matrix was imported into the Ropls software package for pattern recognition analysis using PCA and OPLS-DA. The analysis software platforms and versions used in this study are summarized in Table 1.

3. Results

3.1 Venn Analysis of Differential Metabolites

Substances detected by mass spectrometry were obtained through LC-MS analysis (Supplementary Table S1). Using primary and secondary mass spectrometry data, searches were conducted across libraries (in-house database, Metlin database, HMDB database, etc.) to obtain annotated substances. Metabolite information for the experimental and control groups is provided in Supplementary Tables S2 and S3, respectively. A total of 765 metabolites were identified in the serum of the experimental group, while 697 metabolites were identified in the control group.

The metabolite data from both groups were imported into Venn Diagram (R package) software to identify overlapping metabolites. Venn diagram analysis revealed the number and overlapping relationships of metabolites in each group. Among positive ion metabolites, 396 were detected in the control group and 442 in the experimental group, with an intersection of 396 metabolites. Among negative ion metabolites, 301 were detected in the control group and 323 in the experimental group, with an intersection of 301 metabolites. These results are summarized in Figure 1.

3.2 Multivariate Statistical Analysis

3.2.1 Difference Statistics and Volcano Plot

Differential metabolites between the two groups were compared and visualized as shown in Figure 2. The X-axis represents the fold change value of metabolite expression differences between groups (log₂FC), while the Y-axis represents the statistical significance value (-log₁₀(p-value)). Higher values indicate more remarkable expression discrepancies. Each point in the figure represents a metabolite, with point size indicating VIP value. Points on the left represent downregulated metabolites, while points on the right represent upregulated metabolites. Points closer to the left, right, and upper regions show more significant expression differences. The results demonstrated significant differences between experimental and control groups, with 115 differential metabolites in positive ion mode and 87 in negative ion mode, showing more upregulation than downregulation (Figure 2A, B, C).

3.2.2 Details of Intergroup Differences

Intergroup differential metabolism was analyzed to generate histograms of relative expression abundance for metabolites in each sample group. Difference testing was performed using Student's t-test (unpaired, two-tailed). OPLS-DA data conversion employed Pareto transformation with a confidence level of 0.95 and 200 permutations (P < 0.05, VIP > 1, FC > 1 or < 1). The Y-axis represents mass spectrometry intensity values (post-preprocessing), with error bars showing mean ± standard deviation. VIP values were selected in descending order from OPLS-DA, and the metabolite with the highest VIP value was selected for differential metabolism analysis (Figure 2D, E, F). Abscisic acid showed the greatest difference between experimental and control groups in positive ion mode and combined positive-negative ion mode, while (S)-naproxen showed the greatest difference in negative ion mode. Both metabolites had higher concentrations in the experimental group than in the control group.

3.2.3 PCA Analysis

Principal component analysis (PCA) was applied to LC-MS data from experimental and control groups. After dimensionality reduction, PCA results in positive and negative ion modes are shown in Figure 3A and B, respectively. The experimental and control groups exhibited obvious separation trends, indicating significantly different metabolite profiles between the two groups. The confidence ellipse indicates that "true" samples of each group are distributed within this area at a 95% confidence level; samples beyond this area may be considered abnormal. Panels A and B in Figure 3 show that no abnormal samples deviated from the confidence ellipse in the LC-MS analysis (green dots: control group; blue triangles: experimental group).

3.2.4 PLS-DA Analysis

Partial least squares discriminant analysis (PLS-DA) was performed on LC-MS data from both sample groups. The PLS-DA results in positive and negative ion modes are shown in Figure 4A and B, respectively, demonstrating obvious separation trends and significant metabolic differences between groups. To test model reliability and prevent overfitting, model overview and permutation test charts were used for verification. The PLS-DA model is shown in Figure 4C and D. The positive ion model contained two principal components: model interpretation rate R²X was 0.407, predictive ability parameter Q² was 0.953, and Q² > 0.5 indicated good predictive ability and fit. The negative ion model also contained two principal components: R²X was 0.523, Q² was 0.966, and Q² > 0.5 indicated excellent fit.

In the permutation test, the abscissa represents permutation retention, the ordinate represents R² (red dots) and Q² (blue triangles) values from permutation tests, and the two dotted lines represent regression lines for R² and Q², respectively. Permutation tests were performed by randomly disrupting grouping labels (Y variables) of experimental and control groups. Permutation retention on the horizontal axis represents the proportion consistent with the original Y variable order, with retention of 1 corresponding to R² and Q² values of the original PLS-DA model. With 200 random permutations, R² was 0.8818 and Q² was -0.0401 for the negative ion model, while R² was 0.9435 and Q² was -0.0710 for the positive ion model. Negative Q² values indicated no overfitting and successful modeling. As permutation retention decreased, the regression line showed an upward trend, indicating the permutation test was passed (Figure 4E, F).

3.2.5 OPLS-DA Analysis

The OPLS-DA score plot filters out information irrelevant to grouping through orthogonal rotation, thereby better distinguishing between groups and improving model performance. OPLS-DA results in positive and negative ion modes are shown in Figure 5A and B, respectively, with obvious separation trends indicating significant metabolic differences between groups. Model reliability was verified using model overview and permutation test charts to prevent overfitting (Figure 5C, D). The positive ion model contained two principal components: R²X was 0.407, Q² was 0.919, and Q² > 0.5 indicated reliable model construction with good predictive ability and fit. The negative ion model contained two principal components: R²X was 0.523, Q² was 0.96, and Q² > 0.5 indicated excellent fit.

In the permutation test, with 200 random permutations, R² was 0.8811 and Q² was -0.0948 for the negative ion model, while R² was 0.943 and Q² was -0.0607 for the positive ion model. Negative Q² values indicated reliable model construction without overfitting. As permutation retention decreased, the regression line showed an upward trend, indicating the permutation test was passed (Figure 5E, F).

3.3 Cluster Analysis of Metabolites

Hierarchical cluster analysis was performed on differential metabolites between experimental and control groups using hierarchical clustering algorithms. Subclusters were selected from the top ten, with the top 30 most abundant metabolites shown in Figure 6. The left side shows the metabolite clustering dendrogram, with metabolite names on the right. The closer two metabolite branches are, the more similar their expression levels. The sample clustering dendrogram is at the top, with sample names at the bottom. The closer two sample branches are, the more similar the expression patterns of all metabolites between the two samples. Metabolite expression trends differed between control and experimental groups. The expression levels of 2-indolecarboxylic acid, indole-3-carboxylic acid-O-sulfate in subcluster 5, carboxyoprost bromethamine in subcluster 8, and (±)-propionylcarnitine in subcluster 10 were upregulated, while other subclusters were downregulated. The expression levels of remaining metabolites were upregulated in the experimental group.

3.4 Variable Importance in Projection (VIP) Value Analysis

Based on weighting coefficients from the OPLS-DA model, VIP scores were used to rank metabolite contributions to group discrimination. The metabolite clustering algorithm was hierarchical clustering with Euclidean distance and complete linkage methods. VIP values were derived from OPLS-DA with VIP ≥ 1, and the top 30 metabolites by VIP value were selected. As shown in Figure 7, higher VIP values indicated greater differences between groups. The top five metabolites were abscisic acid (VIP = 6.1604), quillaic acid (VIP = 6.0746), 2,2-bis(4-hydroxyphenyl)-1-propanol (VIP = 5.543), Corey PG-lactone diol (VIP = 5.4712), and (S)-naproxen (VIP = 4.7948). These differential metabolites are detailed in Table 2.

3.5 KEGG Compound Classification

Kyoto Encyclopedia of Genes and Genomes (KEGG) compound classification categorizes compounds by biological role. Identified metabolites were compared against the KEGG Compound database to obtain classification profiles and statistical mapping. KEGG compound classification revealed that XHP metabolites consisted primarily of phospholipids (22 metabolites), amino acids (14 metabolites), steroid hormones (6 metabolites), and carboxylic acids (5 metabolites), as shown in Figure 8.

3.6 KEGG Functional Pathway Analysis

By comparing results with the KEGG database, metabolites in the experimental group were sorted according to pathways involved or functions performed. As shown in Figure 9, the main pathways included lipid metabolism (64 metabolites), amino acid metabolism (44 metabolites), cancer overview (41 metabolites), digestive system (28 metabolites), metabolism of other amino acids (16 metabolites), nervous system (15 metabolites), and membrane transport (15 metabolites). XHP exhibited characteristics of multicomponent, multitarget, and multilink effects. To further investigate the mechanism of XHP in treating breast cancer, we identified 15 metabolites that XHP significantly regulated and that were related to breast cancer, along with their functional predictions (see Supplementary Table S4).

3.7 KEGG Pathway Enrichment Analysis

Through KEGG enrichment and topological analyses, 13 different metabolic pathways were identified, as shown in Figures 10 and 11. KEGG enrichment analysis revealed significant changes between the two groups in the following pathways: necroptosis, steroid hormone biosynthesis, beta-alanine metabolism, vitamin digestion and absorption, and apoptosis. Topological analysis identified four main metabolic pathways: steroid hormone biosynthesis, beta-alanine metabolism, retinol metabolism, and caffeine metabolism. The metabolic pathways with statistically significant differences (P < 0.05) were steroid hormone biosynthesis and beta-alanine metabolism. The steroid hormone biosynthesis pathway included three metabolites: 21-hydroxypregnenolone (c05485), corticosterone (c05488), and 21-deoxycortisol (c05497). The beta-alanine metabolism pathway included two metabolites: carnosine (c00386) and pantothenic acid (c00864), as shown in Figure 12.

4. Discussion

With the gradual deepening of biotechnology research, various molecular compounds have been found to impact diseases. For example, F. S. Mohammed et al.¹² showed that Sechium edule extract prolongs QT or QTc and RR intervals and increases heart rate in high-fat diet-challenged mice, while S. Taifa et al.¹³ found that copper nanoparticles have excellent therapeutic effects on mastitis. This study investigated the metabolic compounds of XHP and their mechanisms in breast cancer treatment. XHP, composed of NIU HUANG, SHE XIANG, RU XIANG, and MO YAO, is a commonly used Chinese patent medicine for treating tumors, especially breast cancer. Recent years have seen numerous pharmacological studies on these four constituent medicines, demonstrating that NIU HUANG and SHE XIANG inhibit cancer cell proliferation¹⁴,¹⁵. Active ingredients in NIU HUANG can reduce proliferation of human colon adenocarcinoma cells¹⁶, while deoxycholic acid, a NIU HUANG component, can block cells in G2/M phase, inhibit growth of human gastric cancer BGC-823 cells, and induce apoptosis through the mitochondrial pathway¹⁷. Muskone, the active ingredient in SHE XIANG, can significantly inhibit expression of basic fibroblast growth factor and vascular endothelial cell growth factor in nude mouse models of breast cancer MDA231 cell line, thereby inhibiting tumor angiogenesis¹⁸. RU XIANG and MO YAO can promote cancer cell apoptosis¹⁹. The main antitumor components of RU XIANG are boswellic acids, which reduce cancer cell proliferation, promote apoptosis, and attenuate tumor-related signaling pathway conduction²⁰.

Currently, no studies have investigated the pharmacological mechanism of XHP using nontargeted metabolomics. Therefore, this study employed nontargeted metabolomics to analyze XHP metabolites in rat serum. Mass spectrometry (MS) and nuclear magnetic resonance (NMR) are the two most widely used techniques for compound identification and have a complementary relationship in compound analysis. MS provides the atomic formula of analytes, while NMR reveals how these atoms are organized into structural moieties. NMR can distinguish between isobaric compounds and positional isomers, while MS can recognize certain functional groups such as sulfate and nitro groups. These complementary capabilities mean that neither technique alone suffices for comprehensive compound analysis in natural product discovery, metabolomics, and drug metabolite identification²¹. As an analytical platform, NMR is a powerful tool for studying biomolecular structure and dynamics under physiological conditions²² and plays an essential role in drug discovery by providing protein-ligand binding information in solution²³. NMR was the earliest technique for measuring metabolites in biological specimens and has been successfully applied in tumor marker selection, prognosis assessment, pathological typing, and early diagnosis²⁴. Advantages of NMR include selective isotopic detection in complex mixtures, determination of unknown key structural parameter metabolites, inherent quantification, in situ analysis of pathway dynamics from cells to whole organisms, and nondestructive effects²⁵.

However, NMR has significant limitations: spectral resolution is not high, it lacks chromatographic separation capability, and derived spectra suffer from mutual interference between peaks of various compounds, limiting analytical results. Additionally, its sensitivity is lower than LC-MS²⁶. LC-MS can compensate for these deficiencies simultaneously, enabling more accurate qualitative and quantitative analyses of target compounds in samples.

Through nontargeted metabolomic analysis, we identified 765 metabolites in the XHP drug serum group and 697 in the blank serum group. Multivariate statistical analysis revealed metabolite differences between groups. Comparison of differential metabolites identified abscisic acid and (S)-naproxen as significantly different between groups, with higher concentrations in the XHP drug serum group, suggesting they may be the main functional components of XHP in blood metabolism and circulation. VIP analysis identified abscisic acid, quillaic acid, 2,2-bis(4-hydroxyphenyl)-1-propanol, Corey PG-lactone diol, and (S)-naproxen as significantly different between groups. Therefore, abscisic acid and (S)-naproxen may be the primary differential metabolites.

Abscisic acid is a plant hormone found not only in plants but also in humans and animals. In mice, abscisic acid reduces obesity and interferes with insulin resistance. It can inhibit proliferation of various cancer cells and induce apoptosis, promote mesenchymal stem cell growth and hematopoietic progenitor cell proliferation, and exert anti-inflammatory effects in the immune system, making it a potential therapeutic agent for various diseases²⁷. Studies have found that abscisic acid has antitumor effects and inhibits tumor angiogenesis. Recent research indicates that mammalian cells can also produce abscisic acid, which can inhibit cancer cell proliferation and regulate cancer cell dormancy in bone marrow²⁸⁻³⁰. H. W. Zhao et al.²⁸ found that abscisic acid enhances Caspase-3 activity in cancer cell lines, increases cancer cell apoptosis, slows growth of human tongue cancer Tca8113 xenografts, and induces differentiation. D. S. Jang et al.³¹ found that abscisic acid induced increased quinone reductase content, which could inhibit Hepa 1c1c7 hepatoma growth.

(S)-Naproxen is a nonsteroidal anti-inflammatory drug that reduces synthesis of thromboxane, prostacyclin, and prostaglandin from arachidonic acid by inhibiting cyclooxygenase subtypes, exerting anti-inflammatory, analgesic, and antipyretic effects³². It is used to treat chronic arthritis, osteoarthritis, ankylosing spondylitis, gout, and other conditions³³. (S)-Naproxen exerts anticancer adjuvant effects primarily by inhibiting cyclooxygenase-2 activation, while its inhibition of cyclooxygenase-1 can induce prostaglandin E synthesis in gastric mucosa, causing gastrointestinal adverse reactions³⁴. Z. J. Li et al.³⁵ found that cisplatin compounds containing naproxen structures showed special antitumor activity, with in vitro cytotoxicity on breast tumor cell lines MCF-7 and MDA-MB-231 stronger than cisplatin alone. In conclusion, abscisic acid and (S)-naproxen inhibit breast cancer progression to some extent, though specific mechanisms require further investigation.

The functional KEGG pathways primarily involve lipid metabolism and amino acid metabolism. Modern studies have found that these pathways inhibit breast cancer progression. Regulation of lipid metabolism, including lipid uptake, synthesis, and hydrolysis, is critical for maintaining cellular homeostasis³⁶. Dysregulation of lipid metabolism is one of the most prominent metabolic alterations in cancer. Cancer cells utilize lipid metabolism for energy, biofilm composition, and signaling molecules required for proliferation, survival, invasion, metastasis, and response to tumor microenvironment influences and cancer therapy³⁷. Cancer cells constantly change nutrient availability during tumor progression, using lipid metabolism to support rapid proliferation, survival, migration, invasion, and metastasis. Lipid metabolism facilitates tumor invasion and metastasis by increasing lipid decomposition. Tumor cell lipids derived from adipocytes produce fatty acid derivatives that enhance breast cancer invasiveness³⁸. Y. Dunneram found that postmenopausal estrogen decline may lead to altered fat metabolism and promote breast cancer³⁹. Significant postmenopausal estrogen reduction likely causes lipid metabolism disorders, making postmenopausal women more prone to visceral lipid accumulation than premenopausal women. Increased fatty acid derivatives from adipocyte decomposition can induce elevated cancer risk⁴⁰. Estrogen receptor-positive breast cancer is more dependent on estrogen and lipid metabolism, suggesting the lipid metabolism pathway may be closely related to breast cancer development and progression. Regulating lipid metabolism and inhibiting adipocyte decomposition may help suppress breast cancer growth and metastasis.

Modern research shows that amino acid metabolism is involved in tumor growth, as amino acids are essential for tumor cell survival and tumor cells have increased nutrient demands to promote proliferation and cancer progression. Greater amino acid intake correlates with stronger tumor proliferation ability. Recent studies demonstrate altered amino acid metabolism pathways in breast cancer, with amino acid transporters influencing tumor growth and progression. In breast cancer, glutamine is a key nutrient, and glutamine metabolism is closely related to amino acid transporters⁴¹. Cancer cell proliferation depends on intracellular nucleotide synthesis and continuous DNA replication. Glutamine, serine, glycine, and aspartic acid are necessary amino acid raw materials for nucleotide synthesis, with serine involved in DNA methylation. Arginine, leucine, and glutamine act as signaling factors to activate the mammalian target of rapamycin signaling pathway, which is closely related to tumor growth regulation. Tumor cell proliferation and maintenance depend on amino acid supply to the intracellular space, regulated by amino acid transporters. Tumor development involves many pathological molecular mechanisms induced by various factors, with amino acid metabolism-related mechanisms including oncogene activation, related protein expression, cancer cell proliferation and metastasis, immune suppression, and immune escape⁴².

KEGG enrichment and topological analysis showed XHP was primarily involved in steroid hormone biosynthesis and beta-alanine metabolism pathways. The three XHP metabolites involved in steroid hormone biosynthesis were 21-hydroxypregnenolone, 21-deoxycortisol, and corticosterone. 21-Hydroxypregnenolone is involved in steroid hormone synthesis and can prevent fatigue and reduce stress⁴³. 21-Deoxycortisol is a steroid hormone precursor, and corticosterone is a steroid hormone and key substance in the steroid hormone biosynthetic pathway. Steroid hormones are fat-soluble hormones divided into two cholesterol derivatives based on their pharmacological effects: adrenal cortex hormones and sex hormones. Steroid hormones can enter target cells by diffusion or carrier transport, bind to receptors to form hormone-receptor complexes, and regulate gene expression. Steroid hormone synthesis typically occurs in the adrenal cortex, gonads, brain, placenta, and adipose tissue, primarily involving hydroxysteroid dehydrogenases⁴⁴. Steroid hormones play important roles in regulating hormone-dependent tumor development, maintaining fluid balance, regulating immunity, and mediating metabolic changes and stress responses⁴⁴. The secretion and growth of SHE XIANG are related to steroid hormone biosynthesis⁴⁵, and RU XIANG and MO YAO participate in steroid hormone biosynthesis in breast hyperplasia treatment⁴⁶. Steroid hormones play prominent roles in breast cancer pathogenesis and progression, with estrogen and progesterone potentially inducing breast cancer progression by increasing breast cancer stem cells⁴⁷. Clinically, hormone replacement therapy (HRT) is used to treat sex hormone-dependent tumors⁴⁸. The steroid hormone biosynthesis pathway synthesizes estrogen and progesterone, and modern studies show estrogen can promote proliferation of estrogen receptor-positive breast cancer cells. Therefore, inhibiting this pathway to reduce estrogen production may suppress proliferation of estrogen-progesterone receptor-positive breast cancer cells. Whether XHP affects estrogen-progesterone receptor-positive breast cancer occurrence and development by regulating the steroid hormone biosynthesis pathway warrants further investigation.

The two XHP metabolites involved in the beta-alanine metabolism pathway were carnosine and pantothenic acid. Modern research shows beta-alanine is a unique β-type amino acid that does not participate in protein synthesis. Beta-alanine is metabolized from cytosine and uracil⁴⁹, and its precursor is involved in cytochrome P450 synthesis: enzymes involved in coenzyme A synthesis combine with histidine to form carnosine and its derivative anserine⁵⁰,⁵¹. Therefore, carnosine, pantothenic acid, and beta-alanine are closely related in the beta-alanine metabolism pathway. Carnosine helps inhibit tumor proliferation, and significant carnosine accumulation markedly inhibits tumor growth, likely by hindering glycolysis and energy production⁵². Beta-alanine can transmit information between neurons, regulate hormones, and serves as an intermediate metabolite of various active substances that regulate physiological functions such as metabolism⁵⁰. Beta-alanine metabolism may be involved in apoptosis pathways. B. Saunders et al.⁵⁴ found that continuous beta-alanine content increase in early apoptosis can cause oxidative stress damage to the brain but can increase muscle carnosine levels and relieve soreness. T. Gemelli et al.⁵⁵ showed that gradual beta-alanine content increase reduces activities of two key glycolytic enzymes, pyruvate kinase and lactate dehydrogenase, thereby disrupting glycolysis and inducing apoptosis. Y. Xu⁵⁶ demonstrated that XHP can regulate beta-alanine, inhibit lactate dehydrogenase expression in breast cancer 4T1 cells, reduce cancer cell increase, and promote cell apoptosis. Therefore, XHP may induce breast cancer cell apoptosis by regulating the beta-alanine metabolism pathway.

5. Conclusion

This study used LC-MS technology to investigate metabolite differences between XHP and blank serum groups. Comparative analysis identified 30 differential metabolites, with abscisic acid and (S)-naproxen as the main differential metabolites. Most metabolites were classified as phospholipids, amino acids, steroid hormones, and carboxylic acids. The major KEGG functional pathways involved were lipid metabolism, amino acid metabolism, and cancer overview. Thirteen KEGG enrichment pathways were identified, with steroid hormone biosynthesis and beta-alanine metabolism as the main pathways involved in tumor proliferation, differentiation, and apoptosis. These metabolic changes of XHP in rats elucidate the nature of the main differential metabolites and key metabolic pathways involved, enabling more in-depth study of XHP in treating breast cancer and other tumors.

Data Availability: Data can be obtained from the author upon reasonable request.

Conflict of Interest: All authors declare no conflicts of interest.

Author Contributions: Yi-fan Su and Xiao-hui Zhao contributed equally to this work and share first authorship. Yi-fan Su and Xiao-hui Zhao conducted most of the research, performed statistical analysis and wrote the manuscript. De-hui Li designed the study and reviewed the manuscript. Jiao Liu and Xu-kuo Liu helped in the animal study and recorded the data. All authors have read and agreed to the published version of the manuscript.

Funding: This research was supported by the National Natural Science Foundation of China (Grant No. 81603412); Key R&D Projects of Hebei Province (Grant No. 18277731D); Scientific Research Project of Hebei Administration of Traditional Chinese Medicine (Grants No. 2017163, 2019008, 2020014, 2023045); General Projects for Improving Scientific Research Capacity of Hebei College of Traditional Chinese Medicine (Grant No. KTY2019009); Hebei Key Laboratory of Chinese Medicine Research on Cardio-Cerebrovascular Disease; Key Laboratory of Integrated Traditional Chinese and Western Medicine Hepatonephrosis in Hebei Province (Grant No. A201902); Hebei Province "Three Three Three Talent Project" funded project (Grant No. A202002008); Scientific Research Project of Health Commission of Hebei Province (Grant No. 20220962).

Acknowledgment: We would like to thank Majorbio Cloud Platform (www.majorbio.com) for assistance with data analysis.

Supplementary Materials: Supplementary Table 1: Substances detected by the mass spectrometer. Supplementary Table 2: Summary of metabolite information for experimental groups. Supplementary Table 3: Summary of metabolite information for the control group. Supplementary Table 4: The 15 metabolites that XHP significantly regulated that are related to breast cancer and their function predictions.

References

[1] Li DH, Fan HF, Dong JF, et al. Based on BATMAN-TCM to Explore the Molecular Mechanism of Xihuang Pill Regulating Immune Function to Treat Breast Precancerous Lesions. Breast cancer (Dove Medical Press) 2021;13:725-742.
[2] Yang YT, Zeng J, Chen P, et al. Anti-tumor Application and Pharmacological Mechanism of Xihuangwan: A Review. Chinese Journal of Experimental Traditional Medical Formulae 2022;28(3):250-258.
[3] Xia SL, Wang X, Zhai YC, Liu JN, and Han T. Anti-tumor Mechanism of Xihuangwan and Research Progress on Its Combined Use in Anti-tumor Treatment. Chinese Journal of Experimental Traditional Medical Formulae 2021;27(6):217-225.
[4] Su L, Jiang YM, Xu Y, et al. Xihuang pill promotes apoptosis of Treg cells in the tumor microenvironment in 4T1 mouse breast cancer by upregulating MEKK1/SEK1/JNK1/AP-1 pathway. Biomedicine & pharmacotherapy 2018;102:1111-1119.
[5] Li XY, Su L, Jiang YM, et al. The Antitumor Effect of Xihuang Pill on Treg Cells Decreased in Tumor Microenvironment of 4T1 Breast Tumor-Bearing Mice by PI3K/AKT~AP-1 Signaling Pathway. Evidence-based complementary and alternative medicine 2018;2018:Article ID 6714829, 13 pages.
[6] Fan HF, Li DH, Guo N, et al. Xihuang Pill inhibits the development of DMBA combined estrogen and progesterone induced breast precancerous lesions rats by PI3K/AKT/mTOR signaling pathway. ResearchGate
[7] Li DH, Su YF, Fan HF, et al. Effect of Xihuang Pill on microcirculation in DMBA combined estrogen and progesterone induced breast precancerous lesions rats. IOP Conf. Ser.: Earth Environ. 2020;474(5):52053–52055.
[8] Li DH, Su YF, Fan HF, et al. Xihuang pill reversing precancerous lesions of breast cancer in rats and regulating the expression of NF-kappa B protein. Chinese Journal of Gerontology 2021;41(1):131-134.
[9] Mao D, Feng L, Huang SQ, Zhang SF, Peng WJ, and Zhang SF. Meta-Analysis of Xihuang Pill Efficacy When Combined with Chemotherapy for Treatment of Breast Cancer. Evidence-based complementary and alternative medicine 2019;2019:Article ID 3502460, 14 pages.
[10] Wu YP, Zhou Y, Pu J, Li H, Zhang JG, and Zhu YP. Application Progress of Metabolomics in Tumor Drug Target Screening. Biotechnology Bulletin 2022:1-8.
[11] Pang B, Liu S, Liu ZQ, Liu SY, and Song FR. Application of Microdialysis Combined with Liquid Chromatography-Mass Spectrometry Technique on Chinese Medicine Research. Journal of Chinese Mass Spectrometry Society 2021;42(5):553-562.
[12] Mohammed FS, Ghosh A, Pal S, et al. Hydroalcoholic Extract of Sechium edule Fruits Attenuates QT Prolongation in High Fat Diet-Induced Hyperlipidemic Mice. Evidence-Based Complementary and Alternative Medicine 2022;2022:1-7.
[13] Taifa S, Muhee A, Bhat RA, et al. Evaluation of Therapeutic Efficacy of Copper Nanoparticles in Staphylococcus aureus-Induced Rat Mastitis Model. Journal of Nanomaterials 2022;2022:1-12.
[14] Khan A, Ahmed QU, Narayanamurthy V, et al. Anticancer activity of grassy Hystrix brachyura bezoar and its mechanisms of action: An in vitro and in vivo based study. Biomedicine & pharmacotherapy= Biomedecine & pharmacotherapie 2019:114:108841.
[15] Xu L, and Cao Y. Native musk and synthetic musk ketone strongly induced the growth repression and the apoptosis of cancer cells. BMC complementary and alternative medicine 2016:16(1):511.
[16] Wang C, and Lou YN. Clinical and experimental research progress of Xihuang Pill in the treatment of breast cancer. Journal of China-Japan Friendship Hospital 2020;34(3):170-172.
[17] Wu EB, Yang HB, Lu YF, et al. Effects of deoxycholic acid on apoptosis of human gastric cancer BGC-823 cells. Guangdong Medical Journal 2016;37(12):1795-1798.
[18] Guo SX, Liu YH, Chang J, et al. Expressions of Angiogenic Factors in Breast Cancer with Blood Stasis By Musk Ketone Intervention. Journal of Hebei Traditional Chinese Medicine and Pharmacology 2011;26(2):40-41.
[19] Chen YL, Zhou CL, Ge ZD,et al. Composition and potential anticancer activities of essential oils obtained from myrrh and frankincense. Oncology letters 2013;6(4):1140-1146.
[20] Cai HD, Su SL, Zhou W, et al. Research progress in boswellic acids from medicinal plants in Boswellia Roxb. ex Colebr. and their pharmacological action as well as mechanisms. Chinese Traditional and Herbal Drugs 2016;47(12):2175-2181.
[21] Gathungu RM, Kautz R, Kristal BS, et al. The integration of LC-MS and NMR for the analysis of low molecular weight trace analytes in complex matrices. Mass Spectrom Rev 2020;39(1-2):35-54.
[22] Li Q, and Kang C. A Practical Perspective on the Roles of Solution NMR Spectroscopy in Drug Discovery. Molecules (Basel, Switzerland) 2020;25(13):2974.
[23] Li Y, and Kang C. Solution NMR Spectroscopy in Target-Based Drug Discovery. Molecules (Basel, Switzerland) 2017;22(9):1399.
[24] Kumar D, Gupta A, Mandhani A, and Sankhwar SN. NMR spectroscopy of filtered serum of prostate cancer: A new frontier in metabolomics. The Prostate 2016;76(12):1106-1119.
[25] Li SB, Tian YF, Jiang PYZ, et al. Recent advances in the application of metabolomics for food safety control and food quality analyses. Critical reviews in food science and nutrition 2021;61(9):1448-1469.
[26] Yan XY, and Gao JZ, Application of MALDI-TOF MS in the early diagnosis of non-small cell lung cancer. The Journal of Medical Theory and Practice 2022;35(9):1478-1481.
[27] Guo TY. Research Progress on the pharmacological action of abscisic acid. Electronic Journal of Clinical Medical Literature 2017;4(77):15239+15242.
[28] Zhao HW, Li LJ, Pan J, Han B, and Xia H. Apoptosis of Tca8113 Squamous Cell Carcinoma Cells Induced by Abscisic Acid in Animal Models. Chinese Journal of Clinical Oncology 2008;35(9):523-526.
[29] Zhang H, Lao XZ, Li HQ, and Zheng H. Research advances in pharmacological function and molecular target of abscisic acid. Biotic Resources 2017;39(5):322-327.
[30] Jung YH, Cackowski FC, Yumoto KJ, et al. Abscisic acid regulates dormancy of prostate cancer disseminated tumor cells in the bone marrow. Neoplasia 2021;23(1):009.
[31] Jang DS, Cuendet M, Pawlus AD, et al. Potential cancer chemo-preventive constituents of the leaves of Macaranga triloba. Phytochemistry 2004;65(3):345-350.
[32] Wojcieszyńska D, and Guzik U. Naproxen in the environment: its occurrence, toxicity to nontarget organisms and biodegradation. Appl Microbiol Biotechnol 2020;104(5):1849-1857.
[33] Hao LF, Chen Y, Huang YP, and Liu ZS. Study of release on the moleculary imprinted polymer for S-naproxen. Journal of Tianjin Medical University 2009;15(3):345-347.
[34] Patrignani P, and Patrono C. Cyclooxygenase inhibitors: From pharmacology to clinical read-outs. Biochimica et biophysica acta 2015;1851(4):422-432.
[35] Li ZJ, Chen Y, Wang QP, et al. Research Progress of Anti-Inflammatory Platinum(IV) Complexes as Antitumor Drugs. Journal of Liaocheng University(Natural Science Edition) 2020;33(5):97-103.
[36] Liu TT, and Yu SP. Research progress in the regulation of glucose and lipid metabolism by sirtuin2. Journal of Jining Medical University 2017;40(1):59-62.
[37] Bian XL, Liu R, Meng Y, et al. Lipid metabolism and cancer. The Journal of experimental medicine 2021;218(1): e20201606.
[38] Wang AM, and Chen SZ. Effect of adipose tissue microenvironment on tumor biology. Chinese Medicinal Biotechnology 2021;16(6):539-545.
[39] Dunneram Y, Greenwood DC, and Cade JE. Diet, menopause and the risk of ovarian, endometrial and breast cancer. The Proceedings of the Nutrition Society 2019;78(3):438-448.
[40] Ko SH, and Kim HS. Menopause-Associated Lipid Metabolic Disorders and Foods Beneficial Postmenopausal Women. Nutrients 2020;12(1):202.
[41] Cha YJ, Kim ES, and Koo JS. Amino Acid Transporters and Glutamine Metabolism in Breast Cancer. International journal of molecular sciences 2018;19(3):907.
[42] Gao HM, and Yang SY. Research progress of tumor amino acid metabolism. Medical & Pharmaceutical Journal of Chinese People's Liberation Army 2021;33(8):112-116.
[43] Chi AP, Kang YJ, Zhang Y, et al. Urinary metabolomics study of Epimedium polysaccharide for the treatment of chronic fatigue rats. Journal of Shaanxi Normal University(Natural Science Edition) 2017;45(2):116-124.
[44] Lina S, Lise B, Elizabeth SB, et al. Human steroid biosynthesis, metabolism and excretion are differentially reflected by serum and urine steroid metabolomes: A comprehensive review. J Steroid Biochem Mol Biol 2019;194:105439.
[45] Zhou XN, Yang LH, Li ZM, Duan Y, Zhou T, and Hu KW. Discussion on the Clinical Application of Xingxiao Pills in the Treatment of Malignant Tumor. Journal of Guangzhou University of Traditional Chinese Medicine 2021;38(8):1745-1749.
[46] Ge AQ, Yang KL, Liu LF, Zeng LT, and Fan HQ. A Network Pharmacology Approach to Explore the Molecular Mechanism of RU XIANG-MO YAO Pair on Breast Hyperplasia. Modernization of Traditional Chinese Medicine and Materia Medica-World Science and Technology 2020;22(4):914-925.
[47] Wang Y., Qiao JJ, Zhang L, Liu SW, and Yang SC. The role of steroid hormones in breast cancer and its effect on cancer stem cells. Practical Oncology Journal 2019;33(5):466-470.
[48] Paulsen O, Klepstad P, Rosland JH, et al. Efficacy of methylprednisolone on pain, fatigue, and appetite loss in patients with advanced cancer using opioids: a randomized, placebo-controlled, double-blind trial. J Clin Oncol 2014;32(29):3221-3228.
[49] Wu ZG, Yin WX, Luo HL, Si YK, Sun MQ, and Liao LC. Metabolic abnormalities associated with ketamine-associated bladder toxicity based on metabolomics. Acta Pharmaceutica Sinica 2020;55(8):1849-1854.
[50] Zhong SJ, Li J, Li L, et al. Study on The Biological Mechanism of Diseases due to Salty Taste Based on Urine Metabonomics. Journal of Basic Chinese Medicine 2021;27(6):950-953+1003.
[51] Qi B, Wu SG, Wang J, Qi GH, and Zhang HJ. Physiological Function and Its Application in Animal Feeding of β-Alanine. Chinese Journal of Animal Nutrition 2016;28(4):1027-1034.
[52] Gaunitz F, and Hipkiss AR. Carnosine and cancer: a perspective. Amino acids 2012;43(1):135-142.
[53] Chen T, Xu SY, and Feng Y. Inducing Conditions of Recombined L-aspartate-α-decarboxylase in Fermentor. Journal of Jinling Institute of Technology 2016;32(3):80-83.
[54] Saunders B, Elliott-Sale K, Artioli G. G, et al. β-alanine supplementation to improve exercise capacity and performance: a systematic review and meta-analysis. Br J Sports Med 2017;51(8):658-669.
[55] Gemelli T, De Andrade RB, Rojas DB, et al. Chronic exposure to beta-alanine generates oxidative stress and alters energy metabolism in cerebral cortex and cerebellum of Wistar rats. Mol Neurobiol 2018;55(6):5101-5110.
[56] Xu Y. The Mechanism research of Xihuang Pills Inhibiting Tumor Growth in Breast Tumor-bearing Mice by Regulating Amino Acid Metabolism. Dalian University 2018.

Figures

Figure 1 Venn diagram of differential metabolites.
Note: A. Positive ion mode; B. Negative ion mode.

Figure 2 Two groups of difference volcano plots and abundance histograms.
Note: A. Positive ion mode; B. Negative ion mode; C. Positive-negative ion mode; D. Positive ion mode; E. Negative ion mode; F. Positive-negative ion mode.

Figure 3 Two sets of PCA score chart.
Note: A. Positive ion mode; B. Negative ion mode.

Figure 4 Two groups of PLS-DA score diagrams, model overview diagrams, and permutation test diagrams.
Note: A. Positive ion mode score diagram; B. Negative ion mode score diagram; C. Positive ion mode model overview diagram; D. Negative ion mode model overview diagram; E. Positive ion mode replacement inspection diagram; F. Negative ion mode replacement inspection diagram.

Figure 5 Two groups of OPLS-DA score diagrams, model overview diagrams and permutation test diagrams.
Note: A. Positive ion mode score diagram; B. Negative ion mode score diagram; C. Positive ion mode model overview diagram; D. Negative ion mode model overview diagram; E. Positive ion mode replacement inspection diagram; F. Negative ion mode replacement inspection diagram.

Figure 6 Metabolite cluster analysis chart.
Note: The color represents the relative expression level of the metabolite in the sample group. The specific expression level change trend is shown in the number label under the color bar at the bottom right.

Figure 7 VIP analysis chart.
Note: The right side shows the heat map of metabolite expression, where each column represents a sample and each row represents a metabolite. The color indicates the relative expression level of the metabolite in the sample group.

Figure 8 KEGG compound classification diagram.

Figure 9 KEGG functional pathway diagram.

Figure 10 KEGG metabolic pathway analysis diagram.

Figure 11 KEGG topology analysis path diagram.

Figure 12 Network diagram of KEGG pathway in steroid hormone biosynthesis and beta-alanine metabolism.
Note: Metabolites on a white background indicate metabolites in this pathway; pathways on a red background indicate metabolites in this metabolic concentration.

Tables

Table 1 Analysis software platforms and related versions involved in this paper.

Software/URL Version Metlin database (https://metlin.scripps.edu/) Version 5.0 HMDB database (http://www.hmdb.ca/) Version 1.6.20 Venn Diagram (R packages) Version 1.0.0 Expression data preprocessing Majorbio's own software Correlation analysis Scipy (Python) Version 1.6.2 Differential Metabolite Analysis - Multivariate Statistics Scipy (Python) Version 1.6.2 PCA analysis Ropls (R packages) Version 1.0.0 Cluster analysis Scipy (Python) Version 1.0.0 Heatmap Scipy (Python) Version 1.6.2 VIP analysis Ropls (R); scipy (Python) Version 1.6.2; Version 1.0.0 KEGG compound classification https://www.kegg.jp/kegg/compound/ kegg_v 94.2 KEGG pathway enrichment Scipy (Python) Version 1.0.0

Table 2 Serum Differential Metabolism Markers of XHP

Identifier Metabolite Formula VIP_value P_value pos_1413 Abscisic acid C₁₅H₂₀O₄ 6.1604 1.659E-11 neg_3106 11,12,13-Trinor-1(10)-spirovetivene-2,7-dione C₁₂H₁₆O₂ 6.0746 1.438E-08 neg_2250 8-Acetylegelolide C₁₆H₂₀O₆ 5.5430 6.684E-07 pos_2847 21-Deoxycortisol C₂₁H₃₀O₄ 5.4712 6.039E-14 pos_1437 2,2-Bis(4-hydroxyphenyl)-1-propanol C₁₅H₁₆O₃ 4.7948 2.675E-10 neg_7754 {3-[2-(3,4-dihydroxy-5-methoxyphenyl)ethyl]phenyl}oxidanesulfonic acid C₁₅H₁₆O₇S 4.1112 2.493E-13 pos_1615 Quillaic acid C₃₀H₄₆O₅ 3.9630 1.900E-10 pos_646 Corey PG-Lactone Diol C₁₅H₂₄O₄ 3.8541 8.277E-11 pos_1583 {[3-(2-hydroxyphenyl)oxiran-2-yl]methoxy}sulfonic acid C₂₀H₃₀O₅ 3.7620 4.111E-12 neg_2304 Citreovirenone C₉H₁₀O₆S 3.6541 1.063E-09 neg_7701 Helenalin C₁₄H₁₆O₄ 3.5420 1.963E-07 pos_1381 Glandulone B C₁₅H₁₈O₄ 3.4412 1.663E-08 pos_643 AUSTRICINE C₁₅H₁₆O₃ 3.3841 4.830E-09 pos_673 3-(1,1-Dimethylallyl)scopoletin C₁₅H₁₈O₄ 3.2841 2.653E-08 pos_2977 1alpha-hydroxy-22-oxo-23,24,25,26,27-pentanorvitamin D3 C₁₅H₁₆O₄ 3.1841 1.584E-09 pos_1604 Pterosin E C₂₂H₃₂O₃ 3.0841 1.988E-16 pos_1490 24-Norursa-3,12-dien-11-one C₁₄H₁₆O₃ 2.9841 5.626E-08 pos_2481 Isoglabrolide C₂₉H₄₄O 2.8841 3.241E-12 pos_2740 Glyyunnansapogenin B C₃₀H₄₄O₄ 2.7841 3.315E-11 neg_3797 Formononetin 7-(6''-malonylglucoside) C₃₀H₄₈O₅ 2.6841 2.035E-08 neg_447 Fusidic Acid C₂₅H₂₄O₁₂ 2.5841 3.238E-08 neg_3897 Lactapiperanol D C₃₁H₄₈O₆ 2.4841 3.116E-16 neg_7071 8-Deoxy-11,13-dihydroxygrosheimin C₁₈H₂₈O₅ 2.3841 7.869E-13 neg_584 Achimilic acid C₁₅H₂₀O₅ 2.2841 8.945E-08 pos_1426 (+)-cis-5,6-Dihydro-5-hydroxy-4-methoxy-6-(2-phenylethyl)-2H-pyran-2-one C₁₅H₁₈O₅ 2.1841 2.113E-10 pos_3000 2-Butanone, 4-(6-hydroxy-2-naphthalenyl)- C₁₄H₁₆O₄ 2.0841 2.690E-08 pos_3055 (4S,6R)-p-Mentha-1,8-diene-6,7-diol 7-glucoside C₁₄H₁₄O₂ 1.9841 6.039E-14 pos_3212 (S)-Naproxen C₁₆H₂₆O₇ 1.8841 1.438E-08 neg_7818 (E)-Suberenol C₁₄H₁₄O₃ 1.7841 6.684E-07 neg_2906 7-(6-hydroxy-2-naphthalenyl)- C₁₅H₁₆O₄ 1.6841 1.659E-11