Untargeted metabolomics analysis combined with network pharmacology reveals differences in chemical profiles and activities of different processed products of Clematis chinensis Osbeck

Clematis chinensis Osbeck (WLX), an herbal medicine, which has been used to treat diseases for thousand years. In clinic, WLX presents four forms, including unprocessed WLX (SZ), WLX mixed with yellow wine (JB), WLX soaked in yellow wine (JJ) and WLX stir-fried with yellow wine (JC). However, few studies have focused on revealing differences in its composition and activity. In this study, metabolomics was used to screen featured metabolites of WLX and its processed products. Subsequently, network pharmacology was applied to reveal the differences in their activity. The results showed that a total of 310 metabolites were identified, and 11 featured metabolites were selected, such as tracheloside, astilbin and genistein. Network pharmacology analysis suggested that SZ samples mainly exerts its specific therapeutic effects by regulating pain pathways, including glutamatergic synapse, spinocerebellar ataxia and retrograde endocannabinoid signaling. JB, JJ, and JC groups predominantly exert their effects through pathways, including calcium signaling pathway, estrogen signaling pathway and cAMP signaling pathway, thereby inhibiting the development of inflammatory and pain transmission. This study will provide a theoretical basis for clinical application of WLX and its processed products, and offer a reference for their quality control. Untargeted metabolomics analysis combined with network pharmacology reveals differences in chemical profiles and activities of different processed products of Clematis chinensis Osbeck Jianlin Ke a,c, Lei Wang c, Lili Huang c, Chongchong Li c, Jingjing Yang b*, Chao Zhang a* a. School of Pharmacy, Shandong University of Traditional Chinese Medicine, Jinan 250300, China b. Department of Occupational disease, The Second Affiliated Hospital of Shandong University of Traditional Chinese Medicine, Jinan, 250024, China c. Department of Special Inspection, The Second Affiliated Hospital of Shandong University of Traditional Chinese Medicine, Jinan, 250024, China *Correspondence author: Chao Zhang & Jingjing Yang E-mail address: [email protected] (C. Zhang) & [email protected] (J.J. Yang)

Clematis chinensis Osbeck (WLX), an herbal medicine, which has been used to treat diseases for thousand years. In clinic, WLX presents four forms, including unprocessed WLX (SZ), WLX mixed with yellow wine (JB), WLX soaked in yellow wine (JJ) and WLX stir-fried with yellow wine (JC). However, few studies have focused on revealing differences in its composition and activity. In this study, metabolomics was used to screen featured metabolites of WLX and its processed products. Subsequently, network pharmacology was applied to reveal the differences in their activity. The results showed that a total of 310 metabolites were identified, and 11 featured metabolites were selected, such as tracheloside, astilbin and genistein. Network pharmacology analysis suggested that SZ samples mainly exerts its specific therapeutic effects by regulating pain pathways, including glutamatergic synapse, spinocerebellar ataxia and retrograde endocannabinoid signaling. JB, JJ, and JC groups predominantly exert their effects through pathways, including calcium signaling pathway, estrogen signaling pathway and cAMP signaling pathway, thereby inhibiting the development of inflammatory and pain transmission. This study will provide a theoretical basis for clinical application of WLX and its processed products, and offer a reference for their quality control.

Clematis chinensis Osbeck, metabolomics, network pharmacology, chemical components, activity

Traditional Chinese Medicine Processing (TCMP), a distinctive pharmaceutical methodology rooted in millennia-old medical philosophies, has been designated by UNESCO as part of Humanity’s Intangible Cultural Heritage 1 . This specialized practice enhances therapeutic efficacy while minimizing side effects through targeted modifications of medicinal properties, ultimately generating differentiated therapeutic outcomes 2 . The pharmacological divergence between freshly harvested Panax notoginseng and its steam-processed counterpart exemplifies such transformation-induced therapeutic variations 3 . Modern scientific investigations reveal that TCMP-initiated chemical transformations, including hydrolysis, oxidation, displacement reactions, isomerization and molecular decomposition, constitute the primary mechanism underlying these pharmacological differences 4-6 . Therefore, systematically studying the differences in the content of chemical components will be beneficial for revealing the processing mechanism of traditional Chinese medicine. Clematidis Radix et Rhizome, also called “Wei-Ling-Xian (WLX)” in Chinese, is the root of Clematis chinensis Osbeck, which has been used for thousands of years as a traditional Chinese medicine to treat various diseases, such as rheumatic arthralgia, limb numbness, tendon constriction and inconvenience in flexion and extension 7-9 . These pharmacological effects can be attributed to their rich chemical composition, including organic acids, flavonoids, triterpenoid saponins, and other components 10 . For instance, phenolic glycosides and indole alkaloids isolated from WLX could reduce the secretion of inflammatory factors in LPS-induced RAW 264.7 cells, and exhibit anti-inflammatory activity 7 . Wu et al 11 found that saponin-rich fraction from WLX could evidently alleviate experimental osteoarthritis induced by monosodium iodoacetate in rats through protecting articular cartilage and inhibiting local inflammation. Saponin fraction from WLX effectively alleviates joint destruction and cartilage erosion in MIA-induced rat osteoarthritis models. The chondroprotective mechanism of SFC primarily operates by preventing extracellular matrix degradation and mitigating chondrocyte damage, thereby preserving the integrity of articular cartilage 12 . In clinical practice, WLX is often used after processing. After processing, WLX presents four forms, including unprocessed WLX (SZ), WLX mixed with yellow wine (JB), WLX soaked in yellow wine (JJ) and WLX stir-fried with yellow wine (JC). Different processed products of WLX have varying therapeutic effects on pain and inflammation. However, few studies have focused on the chemical composition of different processed products of WLX. Metabolomics, as an innovative approach, has emerged as a powerful tool for revealing differences in chemical constituent content and elucidating the relationship between compounds and pharmacological effects. This methodology enables comprehensive analysis of chemical components present in herbal medicines, offering systematic insights into their complex chemical profiles 13-16 . Network pharmacology serves as a powerful tool for investigating the potential mechanisms of action of traditional Chinese medicine (TCM) and its compound formulas. By constructing interaction networks involving multi-component, multi-target, and multi-pathway relationships, it elucidates the scientific principles underlying the therapeutic effects of TCM in disease treatment, which aligns with the holistic concept inherent in traditional Chinese medical theory 17-19 . However, traditional network pharmacology approaches primarily focus on establishing tripartite interaction relationships while neglecting the differential influence of component concentrations on target regulation intensity and prioritization. To address this limitation, You et al 20 proposed the novel concept of “Target Control Force”, which incorporates quantitative component profiles into analytical frameworks. This advancement partially elucidates the mechanistic basis for the varied therapeutic effects observed when administering identical herbal medicines, thereby enhancing the precision of network pharmacology in interpreting traditional Chinese medicinal formulations. This study aims to utilize ̵ultra-performance liquid chromatography coupled with high-resolution mass spectrometry (UPLC-HRMS) to analyze differences in the content of chemical components among different processed products of WLX. Subsequently, ̵ Target control force ̵ will be employed to calculate target variations induced by these chemical component disparities. Finally, ̵network pharmacology will be integrated to evaluate differences in the pharmacological activities of different processed products of WLX. The results will provide a theoretical basis for the clinical application of different processed products of WLX and serve as a reference for their ̵quality control.

Preparation of different processed products of WLX samples Unprocessed WLX samples (SZ) were purchased from Kangmei Pharmaceutical Company, Bozhou, Anhui Province, China. The unprocessed WLX sample (1.0 kg) was added to an appropriate amount of yellow wine, stirred for 10 min, collected, sun-dried, and designated as WLX mixed with wine (JB). The unprocessed WLX sample (10 kg) was added to 1.5 kg of yellow wine, and placed for 2.5 h, collected, sun-dried, and designated as WLX soaked in wine (JJ). The unprocessed WLX sample (12.5 kg) was added to 2.5 kg of yellow wine, and stir-fried for 30 min, collected, sun-dried, and designated as WLX stir-fried with wine (JC). Finally, these samples were crushed into powder and sieved through a 40-mesh sieve for subsequent analysis. UPLC-HRMS analysis Samples (0.5 g) were mixed with 5 mL of 80% methanol and subjected to ultrasonic-assisted extraction at 30°C for 10 min using an ultrasonic constant temperature cleaning machine. Subsequently, the sample sequence employed sequential injection analysis with quality control (QC) samples interspersed at every twelve needles throughout the analytical process. Three additional QC injections were implemented both as system stability analysis before testing. Chemical characterization was performed using a Vanquish UHPLC system (ThermoFisher, Germany) coupled with an Orbitrap Q ExactiveTM HF mass spectrometer, Orbitrap Exploris 480 or Orbitrap Q ExactiveTMHF-X mass spectrometer (Thermo Fisher, Germany) in Novogene Co., Ltd. (Beijing, China). Separation was achieved on a Hypersil Goldcolumn (100×2.1 mm, 1.9μm) maintained at 25°C with a 0.2 mL/min flow rate. The binary mobile phase system consisted of (A) 0.1% formic acid in water and (B) methanol, applied with the following gradient profile: 0-1.5 min (2-2% B), 1.5-3 min (2-85% B), 3-10 min (85-100% B), 10-10.1 min (100-2% B), and 10.1-12 min (2-2% B). The analyses were conducted on a mass spectrometer equipped with an electrospray ionization (ESI) source. The ion source parameters were as follows: Spray Voltage: 3.5 kV; Sheath Gas Flow Rate: 35 psi; Aux Gas Flow Rate: 10 L/min; Capillary Temp: 320°C; S-lens RF Level: 60; Aux Gas Heater Temp: 350°C; Polarity: Positive, Negative; MS/MS Scan Mode: Data-Dependent Scans. The mass range for scanning was m/z 100-1500. To process the raw data (.raw) files in CD 3.3 database searching software, the files were first imported and subjected to simple screening of each metabolite based on parameters such as retention time and mass-to-charge ratio (m/z). Peak area correction was then performed using the first quality control (QC) sample to enhance the accuracy of identification. Subsequently, peak extraction was conducted with settings including a mass deviation of 5 ppm, a signal intensity deviation of 30%, a defined minimum signal intensity, and considerations for adduct ions, while simultaneously quantifying the peak areas. Target ions were integrated, and molecular formulas were predicted based on molecular ion peaks and fragment ions. These predicted formulas were compared against databases such as mzCloud (https://www.mzcloud.org/), mzVault, and Masslist. Background ions were removed using blank samples. The original quantitative results were then standardized using the formula: (original quantitative value of the sample) / (total quantitative value of metabolites in the sample / total quantitative value of metabolites in the QC1 sample), resulting in relative peak areas. Compounds with a coefficient of variation (CV) of relative peak areas in QC samples greater than 30% were excluded. Finally, the identification and relative quantification results of the metabolites were obtained. Following preprocessing steps including peak detection, feature alignment, and compound identification, a K × N matrix was generated, with K denoting WLX sample quantities and N representing detected chemical constituents. Then, multivariate data analysis was performed. Specifically, dimensionality reduction via PCA (“FactoMineR” package) visualized inter-grade sample distribution patterns. Featured metabolites with VIP > 1.5 were identified through PLS-DA modeling (“DiscriMiner” package). Network pharmacology analysis Collection of medicine targets The SMILES formats of featured metabolites were obtained from the PubChem database (https://pubchem.ncbi.nlm.nih.gov/). The SwissTargetPrediction database was employed to predict the related targets of these featured metabolites. Targets were further filtered using the criterion ̵Probability > 0 to refine medicine targets. Calculation of target control force Based on a published method 20, the peak area and Probability were used to represent component content and the component-target relationship, respectively. These parameters were utilized to calculate the target control force and evaluate the contribution order of component content to efficacy. Targets with control forces greater than the average of all target control forces were selected as important medicine targets for WLX and its processed products to reflect their respective therapeutic effects. Collection of disease targets Based on the GeneCards database (https://www.genecards.org/), disease targets of WLX and its different processed products were obtained by combining keywords such as “inflammation” and “pain”. Targets were further filtered using the criterion ̵Probability > 5 to refine disease targets. Subsequently, after integration and deduplication, a disease target library was constructed. Screening of key targets The disease targets and important medicine targets of WLX and its processed products were imported into the Venny platform (https://bioinfogp.cnb.csic.es/tools/venny/) to identify key targets. ̵ Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) Enrichment Analysis Key targets of SZ, JB, JJ, and JC were imported into the DAVID database (https://davidbioinformatics.nih.gov/) for GO functional enrichment and KEGG pathway enrichment analysis, with “Homo sapiens” as the species. The top 5 GO terms (ranked by ̵Gene count and ̵P-value) and relevant KEGG pathways were selected for further analysis.

and discussion Metabolomics analysis The metabolites in WLX and its processed products were analyzed using UPLC-HRMS based metabolomics approach. As summarized in Table S1, a total of 310 metabolites were detected, with 176 and 134 compounds identified in positive and negative ion modes, respectively. Metabolite classification revealed a compositional profile containing 171 organic acids and derivatives, 85 flavonoids and derivatives, 36 alkaloids and derivatives, along with 18 lignans, neolignans and related compounds (Table S1). Notably, organic acids and flavonoids emerged as the predominant metabolite groups in this system. Organic acids are acidic organic compounds containing one or more carboxyl groups. They play a significant role in maintaining the nutritional value and sensory quality of food products and are widely found in various foods and herbs 21 . Organic acids have numerous functional benefits, including antioxidant activity, anti-inflammatory activity, immunomodulatory activity, and other activities 22 . In the present study, a total of 171 organic acids and its derivatives were detected, such as succinic acid (Com_41_neg), malic acid (Com_86_neg), and hydroxycitric acid (Com_21582_pos). Among them, succinic acid is a dicarboxylic acid, which has antiplatelet activity, anxiolytic-like effect, and anticancer activity 23-25 . As shown in Fig. 1A, compared with JC and JJ groups, the content of succinic acid in SZ and JB groups is relatively low, which can be attributed to the alcohol solubility of succinic acid. As a naturally occurring carboxylic acid predominantly present in various fruits, malic acid serves dual functions in the food industry due to its organoleptic properties-acting as a mild-flavored acidulant while demonstrating effective antimicrobial preservation capabilities 26 . As shown in Fig. 1B, the content of malic acid is relatively higher than other groups. Additionally, hydroxycitric acid is a natural organic acid, and has good effects in inhibiting fat synthesis, suppressing appetite, and reducing weight 27 . The content of hydroxycitric acid in JC group is relatively low, which could be attributed to the high temperature caused its degradation. Flavonoids are the main bioactive compounds known to be beneficial for many chronic diseases, widely present in fruits and vegetables, with antioxidant activity, prevention of coronary heart disease, and anti-cancer activity 28 . In this study, a total of 85 flavonoids and derivatives were detected, such as kaempferol (Com_1656_neg), luteolin (Com_6987_neg) and quercetin (Com_7924_neg). Meanwhile, kaempferol is a natural flavanol found in different plant species, which has numerous activities, including antioxidant, anti-inflammatory, and anticancer 29, 30 . As shown in Fig. 1D, the content of kaempferol decreased with the prolong of processing time of WLX processed by yellow wine. Simultaneously, the content of luteolin and quercetin also shows regular changes (Fig. 1E and 1F). Therefore, the differences in the content of these metabolites synergistically result in differences in the chemical profiles of WLX and its different processed products. Fig. 1 The relative content of metabolites in WLX and its processed products: succinic acid (A), malic acid (B), hydroxycitric acid (C), kaempferol (D), luteolin (E), and quercetin (F) Principal Component Analysis (PCA), a technique in dimensionality reduction and pattern recognition, is extensively utilized across disciplines ranging from chemical sciences and pharmaceutical research to biological studies 31 . Therefore, PCA was used to present the differences between WLX and its different processed products. As illustrated in Fig. 2A, different groups of WLX samples could be clearly distinguished. Principal component 1 (Dim1) could explain 57.3% of the variance, and principal component 2 (Dim2) could explain 14.6% of the variance. The PCA model could retain 71.9% of information of original data. Based on Dim1 axis, WLX and its different processed products could be divided into two parts, which could be attributed to processing time of WLX processed by yellow wine. As shown in Fig. 2B, based on identified metabolites, the heatmap was established. It could be found that JC group showed similar metabolites content with JJ group. Similarly, SZ group also showed similar metabolites content with JB group. Additionally, the results of Pearson correlation analysis also proved this result (Fig. 2C). These results were consistent with the results of PCA analysis. PLS-DA (Partial Least Squares Discriminant Analysis), a multivariate statistical method integrating the strengths of Partial Least Squares Regression and Discriminant Analysis, holds significant application value in pattern recognition. By constructing a model that links independent variables with class variables, it not only addresses modeling challenges in high-dimensional datasets but also enables accurate identification of critical feature variables driving classification outcomes 32 . Therefore, PLS-DA was used to screen featured metabolites to remove redundant information. As depicted in Fig. 2D, a total of eleven featured metabolites with VIP > 1.5 were selected. For instance, tracheloside is an active component with evident effect in treating cancer 33, 34 . Astilbin is a natural flavanoid compound, which can treat various diseases, such as hyperuricemia, chronic renal failure, and immunosuppressant 35 . Furthermore, genistein, a known endocrine disruptor, exhibits diverse therapeutic properties spanning multiple disease domains. Its pharmacological profile encompasses anticancer effects, glucose metabolism modulation in diabetes, lipid regulation for hyperlipidemia, antiviral mechanisms, and free radical scavenging capabilities 36 . Thus, these featured components synergistically cause the differences in effects of WLX and its different processed products. Fig. 2 PCA (A); Featured metabolites and their VIP value (B); The heatmap of Pearson correlation analysis result (C); The heatmap of metabolites content in WLX and its different processed products (D) Network pharmacology analysis Collection of medicine targets The SwissTargetPrediction database was used to predict medicine targets for featured metabolites, a total of 343 medicine targets with Probability > 0.5 were detected. Among these, 3-chloro-L-tyrosine could act on 11 medicine targets, 4-hydroxyisoleucine could act on 21 medicine targets, astilbin could act on 36 medicine targets, astragaloside I could act on 33 medicine targets, coumestrol could act on 79 medicine targets, fumaric Acid could act on 6 medicine targets, genistein could act on 89 medicine targets, kuwanon A could act on 36 medicine targets, oenin chloride could act on 19 medicine targets, osajin could act on 6 medicine targets, and tracheloside could act on 7 medicine targets. Following integration and deduplication, 227 unique medicine targets were ultimately obtained. Calculation of target control force The calculation of target control force facilitates the elucidation of the main metabolites responsible for therapeutic effects. After calculating target control force, important medicine targets with target control force over the average value of target control force were respectively screened. Specifically, SZ primarily regulates 29 important medicine targets including INMT, CACNA2D1, GRIA1, GRIA4, and GRIA2; JB mainly modulates 58 important medicine targets such as INMT, TBXAS1, MAOA, EGFR, and ESR1; JJ predominantly influences 47 important medicine targets including INMT, CACNA2D1, GRIA1, GRIA4, and GRIA2; while JC primarily governs 53 important medicine targets comprising INMT, CA2, CA1, CA12, and CA9 (Table S2). Collection of disease targets Using the GeneCards database, disease targets were obtained, including 7160 inflammation-related targets and 6581 pain-related targets. Among these, 730 and 1126 disease targets with Relevance > 5 were acquired, respectively. Following integration and deduplication processes, a total of 1475 unique disease targets were ultimately obtained. Screening of key targets The analysis of WLX and its different processed products using the Venny platform revealed key targets associated with their pharmacological effects (Table S3). As shown in Fig. 3A, the SZ group exerts its effects through 4 key targets including CACNA2D1, GRIA1, OAT, and GRM1; the JB group mediates its pharmacological actions via 18 key targets such as MAOA, EGFR, ESR1, HTR2A, and ADORA1 (Fig. 3B); the JJ group operates through 11 key targets including CACNA2D1, GRIA1, OAT, GRM1, and MAOA (Fig. 3C); while the JC group exerts its effects by targeting 15 key genes such as CACNA2D1, MAOA, EGFR, ESR1, and HTR2A (Fig. 3D). Fig. 3 Venn plot: SZ (A); JB (B); JJ (C); JC (D) ̵ Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) Enrichment Analysis GO and KEGG enrichment analysis were performed using DAVID database. The GO functional enrichment analysis primarily involved Biological Processes (BP), Cellular Components (CC), and Molecular Functions (MF). As shown in Fig. 4A, the key targets of SZ samples predominantly regulated BP such as regulation of postsynaptic cytosolic calcium ion concentration, regulation of postsynaptic membrane potential, and chemical synaptic transmission; CC including postsynaptic density membrane, glutamatergic synapse, and dendrite; MF such as neurotransmitter receptor activity involved in regulation of postsynaptic cytosolic calcium ion concentration and glutamate receptor activity. As depicted in Fig. S1A, the key targets of JB samples predominantly regulated BP such as vasodilation, positive regulation of ERK1 and ERK2 cascade, and positive regulation of fibroblast proliferation; CC including dendrite, postsynaptic membrane, and neuronal cell body; MF such as identical protein binding, neurotransmitter receptor activity involved in regulation of postsynaptic cytosolic calcium ion concentration, and G protein-coupled adenosine receptor activity. As illustrated in Fig. S2A, the key targets of JJ samples predominantly regulated BP such as vasodilation, response to purine-containing compound, and G protein-coupled adenosine receptor signaling pathway; CC including dendrite, postsynaptic membrane, and neuronal cell body; MF such as identical protein binding, neurotransmitter receptor activity involved in regulation of postsynaptic cytosolic calcium ion concentration, and G protein-coupled adenosine receptor activity. As shown in Fig. S3A, the key targets of JC samples predominantly regulated BP such as vasodilation, positive regulation of fibroblast proliferation, and response to estradiol; CC including dendrite, postsynaptic membrane, and neuronal cell body; MF such as identical protein binding, neurotransmitter receptor activity involved in regulation of postsynaptic cytosolic calcium ion concentration, and G protein-coupled adenosine receptor activity. As shown in Fig. 4B, key targets of SZ were mainly enriched in pathways such as glutamatergic synapse, spinocerebellar ataxia, and retrograde endocannabinoid signaling. As shown in Fig. S1B, key targets of JB were mainly enriched in pathways including arginine and proline metabolism, calcium signaling pathway, and estrogen signaling pathway. As shown in Fig. S2B, key targets of JJ were mainly enriched in pathways including calcium signaling pathway, estrogen signaling pathway, and cAMP signaling pathway. As shown in Fig. S3B, key targets of JC were mainly enriched in pathways including arginine and proline metabolism, calcium signaling pathway, and estrogen signaling pathway. Interestingly, KEGG enrichment pathways of SZ differs significantly from other groups, suggesting the necessity and scientificity of processing. Meanwhile, glutamatergic synapses crucially mediate nociceptive signaling pathways while driving pain sensitization mechanisms across multiple brain regions, with their neurotransmitter glutamate demonstrating particular importance in osteoarthritis-induced chronic pain pathogenesis 37, 38 . Endocannabinoid signaling influences diverse processes including brain development, cognitive functions (memory formation, learning), emotional regulation (mood, anxiety, depression), feeding behavior, pain modulation, and addictive pathways 39 . Therefore, SZ samples showed specific therapeutic effect. Furthermore, JB, JJ, and JC groups showed similar KEGG enrichment pathways. Among them, arginine is a semi-essential cationic amino acid that serves as a precursor to nitric oxide (NO), polyamines, proline, glutamate, creatine, and agmatine 40 . Proline is a cyclic subamino acid, which can alleviate inflammatory response, oxidative stress caused by LPS 41 . Therefore, by regulating the metabolism of arginine and glutamate, inflammation and pain can be alleviated. The calcium signaling pathway is associated with inflammation and pain 42 . In inflammatory contexts, calcium ions signaling associates with both pathological and chronic pain mechanisms, where inflammatory mediators directly induce extracellular calcium ions influx 43 . Notably, macrophage calcium channel activation promotes TNF release, establishing a relationship between calcium ions dynamics and inflammatory progression 44 . In addition, estrogen plays a significant role in extensive neurological functions and can reverse allodynia and hyperalgesia manifested under chemical, physical, and thermal stimuli 45, 46 . Overall, SZ group mainly exerts its specific therapeutic effects by regulating pain pathways, while JB, JJ, and JC groups predominantly exert their effects through inhibiting the development of inflammatory and pain transmission. These results indicated that WLX processed with yellow wine promotes the generation of specific effects. Different processed products of WLX showed different effects, which might be caused by the difference in their strength of action. Fig. 4 GO (A) and KEGG (B) enrichment analysis

In this study, metabolomics was used to analyze the differences in metabolite content of WLX and its different processed products, and network pharmacology was applied to reveal the differences in their pharmacological actions. Our findings illustrated that a total of 310 metabolites were identified in WLX and its different processed products. Among them, 11 featured metabolites were screened by PLS-DA. These featured metabolites might be key compounds causing the differences of effects between WLX and its processed products. Network pharmacology analysis suggested that SZ samples mainly exerts its specific therapeutic effects by regulating pain pathways, while JB, JJ, and JC groups predominantly exert their effects through inhibiting the development of inflammatory and pain transmission. This study will provide a theoretical basis for clinical application of WLX and its processed products, and offer a reference for their quality control. CRediT authorship contribution statement Jianlin Ke: Conceptualization, Methodology, Formal analysis, Writing - Original Draft, Lei Wang & Lili Huang: Validation, Writing - Review & Editing, Chongchong Li: Software, Resource, Jingjing Yang & Chao Zhang: Project administration, Funding acquisition, Supervision, Writing - Review & Editing. 11pt, fleqn, a4paper, ]LegrandOrangeBook Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Data Availability Data will be made available on request. 11pt, fleqn, a4paper, ]LegrandOrangeBook Acknowledgments Thank you to Beijing Novogene Technology Co., Ltd for providing metabolomics analysis services. Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at

1. X. Zhao, R. Li, C. Sun, Z. Liu, X. Xiao and T. Chen, Medicinal Plant, 2021, 12, 5551-5559.2. S. Liu, H. Dong, W. Ji, M. Zhang, W. Duan and X. Wang, Food Chemistry Advances, 2023, 3, 100340.3. C. Sun, S. Ma, L. Li, D. Wang, W. Liu, F. Liu, L. Guo and X. Wang, Journal of Ginseng Research, 2021, 45, 726-733.4. B. Cai, K. Qin, H. Wu, H. Cai, T. Lu and X. Zhang, Progress in Chemistry, 2012, 24, 637-649.5. S. Liu, L. Chen, W. Duan, Z. Meng, H. Dong and X. Wang, ACS Omega, 2022, 7, 46833–46842.6. L. Sun, M. Wang and X. Ren, Chinese Traditional and Herbal Drugs, 2017, 52, 5369-5377.7. H. Jiang, S. Shao, Y. Wei, M. Lin and S. Li, Phytochemistry, 2023, 215, 113832.8. N. P. Commission.9. S. Sun, Y. Li, W. Fang, P. Cheng, L. Liu and F. Li, Clin. Exp. Med., 2010, 10, 113-121.10. T. Lin, L. Wang, Y. Zhang, J. Zhang, D. Zhou, F. Fang, L. Liu, B. Liu and Y. Jiang, Journal of Ethnopharmacology, 2021, 270, 113831.11. W. Wu, X. Gao, X. Xu, Y. Luo, M. Liu, Y. Xia and Y. Dai, Cytotechnology, 2013, 65, 287-295.12. W. Wu, X. Xu, Y. Dai and L. Xia, Phytotherapy Research, 2010, 24, 538-546.13. C. H. Johnson, J. Ivanisevic and G. Siuzdak, Nature Reviews Molecular Cell Biology, 2016, 17, 451-459.14. T. F. Jorge, J. A. Rodrigues, C. Caldana, R. Schmidt, J. T. v. Dongen, J. Thomas-Oates and C. António, Mass Spectrometry Reviews, 2016, 35, 620-649.15. X. Yu, J. Nai, H. Guo, X. Yang, X. Deng, X. Yuan, Y. Hua, Y. Tian, F. Xu, Z. Zhang and Y. Huang, Journal of Pharmaceutical Analysis, 2021, 11, 611-616.16. S. Liu, H. Dong, Y. Geng, Y. Mi, Q. Wang and X. Wang, Applied Food Research, 2025, 7, 100952.17. X. Li, Z. Liu, J. Liao, Q. Chen, X. Lu and X. Fan, Chin. J. Nat. Med., 2023, 21, 323-332.18. C. Wang, H. Chen, S. Song, B. Chen, R. Li, Z. Fu, Z. Zhang, Q. Wang and L. Han, Rapid Commun. Mass Spectrom., 2022, 36, e9411.19. A. Liu, M. Zhong, Z. Han, Y. Yan, D. Zhang, X. Wang, M. Wang, Y. Zou and J. Zhang, Rapid Commun. Mass Spectrom., 2025, 39, e9976.20. H. You, X. Yang, Y. Luo, G. Zeng, F. He and C. Ding, Chinese Traditional and Herbal Drugs, 2024, 55, 7021-7032.21. Y. Shi, D. Pu, X. Zhou and Y. Zhang, Foods, 2022, 11, 3408.22. S. Liu, H. Dong, H. Lu, M. Li and X. Wang, Applied Food Research, 2025, 5, 100760.23. Q. Zhang, Y. Zhao and H. Bian, Phytotherapy Research, 2013, 27, 1894-1896.24. S. Chen, Q. Xin, W. Kong, L. Min and J. Li, Life Sci., 2003, 73, 3257-3264.25. G. Kasarci, B. Ertugrul, E. S. Iplik and B. Cakmakoglu, Med. Oncol., 2021, 38, 144.26. C. Marques, A. R. Sotiles, F. O. Farias, G. Oliveira, M. L. Mitterer-Daltoé and M. L. Masson, Arabian Journal of Chemistry, 2020, 13, 9118-9129.27. E. Yuan, Sichuan Food and Fermentation, 2006, 42, 19-23.28. L. H. Yao, Y. M. Jiang, J. Shi, F. A. Tomás-Barberán, N. Datta, R. Singanusong and S. S. Chen, Plant Foods for Human Nutrition, 2004, 59, 113-122.29. W. Alam, H. Khan, M. A. Shah, O. Cauli and L. Saso, Molecules, 2020, 25, 4073.30. M. R. Paravati, G. G. M. Scarlata, M. Milanović and N. M. Abenavoli, Hepatoma Research, 2024, 10, 43.31. B. Ulewicz-Magulska and M. Wesolowski, Antioxidants, 2023, 12 .32. H. Lu, Y. Wang, J. Zhu, J. Huang and F. Li, Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 2024, 313 .33. M.-K. Shin, Y.-D. Jeon, S.-H. Hong, S.-H. Kang, J.-Y. Kee and J.-S. Jin, Antioxidants, 2021, 10, 513.34. S. Guo, X. Bai, S. Shi, S. Li, X. Liu, H. An and X. Kang, Biomedicine & Pharmacotherapy, 2022, 153, 113392.35. A. Sharma, S. Gupta, S. Chauhan, A. Nair and P. Sharma, Pharmacol. Res., 2020, 158, 104894.36. J. Xiang, R. Mlambo, P. Dube, O. Machona, I. Shaw, Y. Seid, Y. He, M. Luo, T. Hong, B. He, W. Zhou and S. Tan, Frontiers in Endocrinology, 2023, 14, 1308341.37. M. Zhuo, Neuropharmacology, 2017, 112, 228-234.38. H. Deng, P. Zhou, J. Wang, J. Zeng and C. Yu, Gene, 2025, 934, 149024.39. L. A. Kasatkina, S. Rittchen and E. M. Sturm, International Journal of Molecular Sciences, 2021, 22, 5431.40. N. Bakshi and C. R. Morris, Journal of Pain Research, 2016, 9, 167-175.41. V. S. Andrade, D. B. Rojas, R. B. d. Andrade, T. D. H. Kim, A. F. Vizuete, Â. Zanatta, M. Wajner, C.-A. S. Gonçalves and C. M. D. Wannmacher, Mol. Neurobiol., 2018, 55, 4068-4077.42. S. Jin, Y. Li, C. Luo, X. Cheng, W. Tao, H. Li, W. Wang, M. Qin, G. Xie and F. Han, Journal of Ethnopharmacology, 2024, 321, 117499.43. P. J. Dalal, W. A. Muller and D. P. Sullivan, The American Journal of Pathology, 2020, 190, 535-542.44. L. M. Montaño, E. Flores-Soto, B. Sommer, H. Solís-Chagoyán and M. Perusquía, Steroids, 2020, 153, 108509.45. Y. Zhang, X. Xiao, X. Zhang, Z. Zhao and Y. Zhang, Journal of Biological Chemistry, 2012, 287, 33268-33281.46. Q. Jia, H. Zhang, B. Xia, Y. Liu, X. Dai, T. Liu, S. Wang, Y. Zhang, M. Lu, B. Chen, R. Zhu, S. Gao, D. Brὂmme, X. Wang, L. Wang and D. Zhang, Journal of Functional Foods, 2022, 95, 105162. Information & Authors Information Version history Peer review timeline Published Rapid Communications in Mass Spectrometry Version of Record25 Sep 2025Published Copyright This work is licensed under a Non Exclusive No Reuse License. Collection

Authors Metrics & Citations Metrics Article Usage 300views 224downloads Citations Download citation Jianlin Ke, Lei Wang, Lili Huang, et al. Untargeted metabolomics analysis combined with network pharmacology reveals differences in chemical profiles and activities of different processed products of Clematis chinensis Osbeck. Authorea. 25 June 2025. DOI: https://doi.org/10.22541/au.175082959.90030310/v1 DOI: https://doi.org/10.22541/au.175082959.90030310/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu.