Exploring Salvia miltiorrhiza's Therapeutic Effects on Adenomyosis by Inhibiting TNF-α/HIF-1α/IL-17-Driven Inflammatory Cascade: Mechanistic Insights from Target Prediction and Experimental Validation

Adenomyosis is a common gynecological inflammatory disorder marked by the invasion of endometrial glands and stroma into the uterine myometrium. It is strongly associated with reduced fertility potential. 1 The advent of sophisticated imaging modalities coupled with enhanced clinical expertise has led to both higher detection rates and a demonstrably younger demographic profile in adenomyosis. 2 Research indicates that adenomyosis affects approximately 20% of gynecology patients and may contribute to more than 40% of all hysterectomies performed. 3 Patients with adenomyosis suffer from severe anemia resulting from heavy menstrual bleeding, as well as challenges related to female infertility and adverse obstetric outcomes. 4 This condition significantly burdens patients and society, posing substantial challenges for gynecology and reproductive endocrinology specialists. Hysterectomy remains the gold-standard definitive treatment for symptomatic adenomyosis. 5 However, this approach carries substantial clinical limitations, including permanent fertility loss and potentially significant quality-of-life impairments. Current conservative surgical alternatives demonstrate considerable variability in therapeutic efficacy, with reported recurrence rates exceeding 40%. 6 The suboptimal outcomes frequently necessitate adjunctive pharmacotherapy to achieve acceptable clinical results. 7 Given these constraints, medical therapy functions as the primary first-line therapeutic strategy, especially for younger patients seeking fertility preservation. However, developing precise targeted therapeutic options still presents unresolved challenges. 8 Therefore, the development of effective therapeutic agents for adenomyosis is critically important. Adenomyosis is known as an estrogen-dependent disease, as its growth is typically stimulated by estrogen and suppressed by progestogens. 9 Consequently, hormone-based endocrine therapies are often effective at managing their symptoms. 10 Some studies suggest that oral contraceptives or progesterone can induce decidualization and atrophy of ectopic endometrium, thereby controlling disease progression. Conversely, the basal endometrium in adenomyosis may be progesterone-resistant, making the efficacy of progesterone treatment debatable. 11 Gonadotropin releasing hormone analogs can suppress adenomyosis by inducing menopausal status, but their use is limited to short-term treatment due to side effects and high recurrence rates after discontinuation. 12 These factors pose significant challenges to current pharmacological management of adenomyosis. Therefore, a growing number of patients are seeking complementary therapeutic options, prompting specialists to investigate traditional Chinese medicine (TCM) protocols. Emerging evidence positions TCM as a promising modality that not only modulates systemic homeostasis 13 but also addresses the chronic inflammatory nature of adenomyosis. 14–16 The accumulation of inflammatory mediators in adenomyotic lesions and their role in disease progression and symptom manifestation 17 provide a rationale for TCM’s observed clinical benefits, including symptom relief, improved fertility outcomes, and reduced recurrence. 14 Mechanistic studies, such as those on Qiu’s Neiyi Recipe, demonstrate TCM’s capacity to regulate inflammatory cascades via the MAPK pathway. 15 And, rhein effectively alleviates adenomyosis in a dose-dependent manner by counteracting the IL-1β-induced increase in nuclear translocation of β-catenin through the inhibition of the NF-κB and β-catenin signaling pathways. 16 These findings collectively underscore the therapeutic potential of targeting inflammatory pathways with herbal medicine in adenomyosis. DS is one of the most significant medicinal plants of TCM in Chinese. Numerous studies have demonstrated that DS and its active compounds possess notable anti-inflammatory, therapeutic effects on menstrual dysregulation, antiplatelet properties, etc. 18 , 19 Although DS is widely used to treat various gynecological diseases, including gynecological cancers, 20 , 21 polycystic ovary syndrome, 22 and endometriosis, 19 the specific active compounds in DS responsible for treating adenomyosis and their mechanisms of action remain unexplored. Understanding the pharmacology and mechanisms of action of DS is crucial for elucidating its therapeutic effects in adenomyosis, offering significant potential for developing novel treatment strategies. While the multi-component and multi-target therapeutic characteristics of herbal medicines have gained global recognition for managing complex diseases lacking specific therapies, this polypharmacological profile also poses challenges in systematically evaluating therapeutic outcomes and deciphering underlying mechanisms. 23 To address these challenges, contemporary research has adopted innovative methodologies such as serum pharmacochemistry, which has proven effective in identifying bioactive herbal components and validating their therapeutic benefits in TCM. 24 Additionally, network pharmacology has emerged as a powerful tool for analyzing complex interactions among drugs, diseases, and targets, enabling a holistic assessment of multi-component and multi-pathway effects. 25 This approach is instrumental in modern TCM research, advancing investigations into the safety, efficacy, and mechanistic basis of herbal therapies, thereby enhancing their scientific credibility and clinical acceptance. 26 Despite these advancements, network pharmacology applications in adenomyosis research remain limited, with no studies to date exploring the mechanisms of DS in treating adenomyosis using this approach. This study aims to elucidate the multi-target mechanisms by which DS ameliorates adenomyosis through an integrative pharmacological approach that combines computational prediction methods (network pharmacology, molecular docking, and molecular dynamics simulations) with multimodal experimental validation techniques (serum pharmacochemistry, transcriptomics, and in vivo assays). The results indicate that DS exerts significant anti-inflammatory effects by synergistically inhibiting the TNF-α/IL-17/HIF-1α signaling pathways and TNF-α/IL-1β activation. These findings provide new insights into the use of TCM for treating adenomyosis and offer theoretical guidance for future experimental research and clinical applications. Figure 1 illustrates the study workflow evaluating the therapeutic effects of DS on adenomyosis. Figure 1 Workflow for studying the mechanisms of DS in treating adenomyosis. Workflow for studying the mechanisms of DS in treating adenomyosis.

UHPLC-MS analysis successfully identified and characterized 35 potential IAIBs in DS, including 8-epi-Chlorajapolide F, 21-Deoxyneridienone B, Neochlorogenic acid, 8alpha-Methacryloyloxybalchanin, and 1-O-Acetyl britannilactone et al. The comprehensive analytical data, including compound name, retention time (RT), ionization mode, adduct formation patterns, molecular formula, observed m/z values, and fragmentation scores, were systematically documented and statistically analyzed in Figure 2 and Table 2 . This high-resolution mass spectrometry-based characterization provides the detailed chemical profile of DS-derived bioactive components, establishing a robust foundation for subsequent pharmacological investigations. Table 2 Thirty-Five IAIBs Identified from DS No. Compound Name m/z RT (min) Mode Adducts Formula Fragmentation Score 1 8-epi-Chlorajapolide F 297.1135393 10.21878333 neg M+Na-2H C 16 H 20 O 4 88.6 2 21-Deoxyneridienone B 373.2026137 10.96355 neg M+FA-H C 21 H 28 O 3 86.5 3 Neochlorogenic acid 353.0883516 5.822633333 neg M-H C 16 H 18 O 9 84 4 8alpha-Methacryloyloxybalchanin 313.1445245 10.92606667 neg M-H2O-H C 19 H 24 O 5 83 5 1-O-Acetyl britannilactone 289.1444651 11.21623333 neg M-H2O-H C 17 H 24 O 5 81.7 6 Foetidin 331.1918445 10.25463333 neg M-H, M+Cl C 20 H 28 O 4 81.6 7 13-Deoxy-Isogibberellic acid 329.1398533 9.884816667 neg M-H C 19 H 22 O 5 81.1 8 Lysofungin 595.2898083 14.82545 neg M-H C 27 H 49 O 12 P 80.9 9 Effusanin B 371.1870339 10.5752 neg M-H2O-H C 22 H 30 O 6 79.9 10 Estrone sulfate 253.1582688 10.20326667 pos M+H-H2O C 18 H 22 O 2 78.1 11 Tagitinin F 313.1428926 8.0611 pos M+H-2H 2 O C 19 H 24 O 6 77.5 12 Chlorantholide E 299.0927631 8.079933333 neg M+Na-2H C 15 H 18 O 5 77 13 Daucoidin A 325.1085895 9.884816667 neg M-H2O-H C 19 H 20 O 6 75.5 14 sucrose 341.1094917 1.136366667 neg M-H, M+Cl C 12 H 22 O 11 73.7 15 Ethyl arachidonate 350.3048373 10.8563 pos M+NH4 C 22 H 36 O 2 72.9 16 3,5-Diprenyl-4-hydroxybenzaldehyde 239.1442279 9.479516667 neg M-H2O-H C 17 H 22 O 2 72.8 17 Gluconic acid 231.0278593 0.785466667 neg M+Cl C 6 H 12 O 7 71.7 18 Norpterosin B 203.1080567 8.410266667 neg M-H C 13 H 16 O 2 71 19 (2R,3S)-Pterosin C 199.1114842 10.80656667 pos M+H-2H2O C 14 H 18 O 3 70 20 Monomethyl lithospermate 551.117,9927 6.79815 neg M-H C 28 H 24 O 12 68.8 21 Toddanol 311.092629 8.835316667 neg M+Na-2H C 16 H 18 O 5 68.1 22 Evernic acid 377.0883516 5.128 neg M-H, M+FA-H C 17 H 16 O 7 66.6 23 6′-O-Acetylglycitin 509.1089692 6.132816667 neg M+Na-2H C 24 H 24 O 11 65.1 24 8-Hydroxythymol 331.1917385 11.10555 neg 2M-H C 10 H 14 O 2 64.4 25 Toddalolactone 3-O-methyl ether 343.1193302 7.7395 neg M+Na-2H C 17 H 22 O 6 64.1 26 Stearic acid amide 325.3206714 11.08583333 pos M+ACN+H C 18 H 37 NO 63.1 27 7,8-Dihydrokawain-5-ol 247.0976718 8.410266667 neg M-H C 14 H 16 O 4 61.4 28 Kainic acid 236.0914016 0.892116667 pos M+Na C 10 H 15 NO 4 59.6 29 Syringin 417.1406351 4.68075 neg M+FA-H C 17 H 24 O 9 58.8 30 Chrysin-7-O-glucuronide 475.0885247 5.891816667 neg M+FA-H C 21 H 18 O 10 57.6 31 Methyl isodrimeninol 271.1701485 10.8218 neg M+Na-2H C 16 H 26 O 2 55.2 32 7-Methoxy-4-methylcoumarin 235.0613522 5.250266667 neg M-H, M+FA-H C 11 H 10 O 3 54.6 33 Glycosolone 296.1275976 9.841 pos M+Na C 16 H 19 NO 3 53.4 34 Crebanine 372.1798724 9.886883333 pos M+CH3OH+H C 20 H 21 NO 4 52.1 35 Stepharine 296.1296352 10.06435 neg M-H C 18 H 19 NO 3 50.2 Figure 2 Ingredients of DS containing serum detected by UHPLC-MS. ( A ) DS TCM, Non-medicated serum, and medicated serum samples (negative ion mode). ( B ) DS TCM, Non-medicated serum, and medicated serum samples (positive ion mode). ( C ) Pie chart of serum metabolite classification. Thirty-Five IAIBs Identified from DS Ingredients of DS containing serum detected by UHPLC-MS. ( A ) DS TCM, Non-medicated serum, and medicated serum samples (negative ion mode). ( B ) DS TCM, Non-medicated serum, and medicated serum samples (positive ion mode). ( C ) Pie chart of serum metabolite classification. To construct the compound-putative target network of DS, we first analyzed 35 compounds and 624 hypothesized targets in DS, and the interactions were identified and visualized as the herb-compound-target network, which contained 660 nodes and 1618 edges in Cytoscape 3.7.2 ( Figure 3A ). Potential adenomyosis targets were identified from GeneCards (364 targets), OMIM (1 target), and PharmGKB (4 targets). After duplication, 367 unique targets were retained for further analysis. Using a Venn diagram tool, we intersected the DS and adenomyosis targets and identified 66 common targets ( Figure 3B and Supplementary Table 1 ). Figure 3 DS and adenomyosis cross-core target analysis. ( A ) DS-Compound-Target network diagram. ( B ) Venn diagram of DS and adenomyosis for identification of intersection targets. ( C ) PPI network diagram of intersecting targets. ( D ) MCODE analysis DS-adenomyosis of PPI network for screening of the core targets. ( E ) Top 12 core targets. DS and adenomyosis cross-core target analysis. ( A ) DS-Compound-Target network diagram. ( B ) Venn diagram of DS and adenomyosis for identification of intersection targets. ( C ) PPI network diagram of intersecting targets. ( D ) MCODE analysis DS-adenomyosis of PPI network for screening of the core targets. ( E ) Top 12 core targets. The PPI network comprised 66 nodes and 106 edges, with an average node degree of 3.21. The network demonstrated significant PPI enrichment ( p -value < 1.0e-16), indicating that the interacting genes had substantially more connections among themselves than would be expected for a randomly selected gene set of the same size and degree distribution from the genome. This significant enrichment suggests that these genes are biologically interconnected and potential function as a cohesive group in the context of drug screening DS for the treatment of adenomyosis. Such analysis offers critical mechanistic insights into the therapeutic potential of DS for adenomyosis treatment. 38 After eliminating isolated and disconnected genes, the remaining interconnected targets were imported into Cytoscape 3.7.2 for visualization and further topological analysis. This process generated a comprehensive PPI network diagram, facilitating the exploration of interaction networks ( Figure 3C ). The module analysis of the MCODE plugin identified the clustering module with the highest score, with an MCODE score of 24.733. The clustering hub module of the 31 targets is the core target for DS therapy of Adenomosis ( Figure 3D ). Subsequently, through rigorous network topology analysis employing predefined screening criteria (degree centrality (DC) > 28, betweenness centrality (BC) > 9.09, closeness centrality (CC) > 0.93), we systematically identified 12 core therapeutic targets ( Figure 3E ). These pivotal targets include TNF, IL1β, MMP2, ESR1, PTGS2, STAT3, BCL2, AKT1, MMP9, ALB, CTNNB1 , and EGFR , collectively representing 18.18% of the total number of targets. The complete topological parameters for these key targets are presented in Table 3 . Table 3 Information on 12 Core Interaction Targets and Its Topological Parameters NO. Gene Full Name Degree Betweenness Closeness 1 STAT3 Signal transducer and activator of transcription 3 30 13.069386 1 2 BCL2 Apoptosis regulator Bcl-2 30 13.069386 1 3 AKT1 AKT Serine/Threonine Kinase 1 30 13.069386 1 4 TNF Tumor necrosis factor superfamily 30 13.069386 1 5 MMP2 Matrix metalloproteinase-2 29 10.9052 0.9677419 6 IL1β Interleukin-1 beta 29 11.507224 0.9677419 7 ESR1 Estrogen receptor alpha 29 10.959687 0.9677419 8 PTGS2 Prostaglandin G/H synthase 2 29 11.5540495 0.9677419 9 MMP9 Matrix metalloproteinase-9 29 10.9052 0.9677419 10 ALB Albumin 29 10.959687 0.9677419 11 CTNNB1 Catenin beta 1 29 11.534021 0.9677419 12 EGFR Epidermal growth factor receptor 28 9.093922 0.9375 Information on 12 Core Interaction Targets and Its Topological Parameters Functional enrichment analysis of the 12 core therapeutic targets in DS-adenomyosis interaction network systematically performed GO and KEGG pathway analyses. The GO enrichment analysis mainly functions through biological processes (BP), molecular functions (MF), and cellular components (CC). The analysis identified a total of 3,139 GO terms, with 2,458 (78.3%) reaching statistical significance. GO-BP contained most terms (2,736 total, 2,240 significant), representing 91.13% of all significant GO terms; GO-MF comprised 200 terms (118 significant, 4.8% of significant GO terms); GO-CC included 203 terms (100 significant, 4.07% of significant GO terms). DS-adenomyosis core therapeutic targets involved in BP mainly include cellular progress, immune system process, biological regulation, biological adhesion, growth, and response to stimulus ( Figure 4A ). These findings collectively suggest that DS may exert its therapeutic effects on adenomyosis by acting on key targets involved in multi-level regulation of these fundamental BP. Figure 4 Core target bioinformatics analysis. ( A ) Gene ontology functional analysis. ( B ) KEGG Pathway annotation, and ( C ) Top 20 KEGG Pathway Enrichment. Core target bioinformatics analysis. ( A ) Gene ontology functional analysis. ( B ) KEGG Pathway annotation, and ( C ) Top 20 KEGG Pathway Enrichment. KEGG enrichment analysis identified a total of 174 signaling pathways corresponding to12 core targets, of which 109 (62.64%) were statistically significant. The most enriched pathways in Human Diseases were associated with cancer, diabetic complications, and infectious diseases, reflecting the inflammatory and proliferative nature of AM. Pathways in Organismal Systems involved in endocrine and immune regulation were prominent, including: Estrogen signaling pathway (Q=4.34e-07), IL-17 signaling pathway (Q=7.67e-05), and C-type lectin receptor signaling pathway (Q=1.24E-04); Key pathways in Environmental Information Processing regulating inflammatory and stress responses: TNF-α signaling pathway (Q=5.72e-06), HIF-1α signaling pathway (Q=1.58e-04) ( Figure 4B ). The top 20 signal pathways are shown in the bubble diagram ( Figure 4C ). These results strongly suggest that DS exerts therapeutic effects on adenomyosis through multi-pathway modulation, primarily targeting inflammation, hormone signaling, and tissue remodeling processes. Given that adenomyosis is characterized as an estrogen-dependent chronic inflammatory process with abundant inflammatory mediators in adenomyotic lesions, 39–41 our pathway analysis revealed five particularly significant inflammatory-related signaling pathways that are predicted to be key mediators of DS’s therapeutic effects: Estrogen signaling pathway, TNF-α signaling pathway, IL-17 signaling pathway, C-type receptor signaling pathway, and HIF1-α signaling pathway. This comprehensive analysis not only highlights the most relevant pathways involved in DS’s action against adenomyosis but also reveals the multi-target mechanisms underlying its therapeutic potential for adenomyosis. 42 Based on the five identified key DS-adenomyosis pathways, we constructed an integrated “DS- Components- Targets - Adenomyosis -Pathways” network ( Figure 5A ). This analysis provides a systematic demonstration of the therapeutic efficacy of DS, achieved through the synchronized regulation of functional targets, regulatory genes, and signaling pathways. Identification of five key pathways corresponding to core targets: TNF, IL1β, MMP2, ESR1, PTGS2, STAT3, BCL2, AKT1, MMP9 , and EGFR . Furthermore, a sankey diagram was employed to visualize the quantitative relationships among hub genes, core pathways, and active compounds ( Figure 5B ). Notably, both IL-1β and TNF were identified as critical regulatory nodes concurrently involved in the TNF-α, IL-17, and HIF-1α signaling pathways. This finding provides novel insights into the multi-target mechanisms through which DS may treat adenomyosis, particularly highlighting its potential to simultaneously modulate multiple inflammatory pathways and providing a theoretical foundation for subsequent experimental validation and drug development. Figure 5 Major pathway-target network of DS in the treatment of adenomyosis. ( A ) “DS-Components-Targets-Signaling pathways-Adenomyosis” network diagram based on major pathway-target. ( B ) Sankey diagram of major pathway-core targets- key components of DS for adenomyosis. Major pathway-target network of DS in the treatment of adenomyosis. ( A ) “DS-Components-Targets-Signaling pathways-Adenomyosis” network diagram based on major pathway-target. ( B ) Sankey diagram of major pathway-core targets- key components of DS for adenomyosis. Molecular docking analysis was used further to validate to characterize the binding interactions between the bioactive components of DS and key core targets. Based on sankey diagram screening, our analysis revealed robust binding interactions of the 10 core target proteins and corresponding bioactive components of DS ( Figure 6A ), with all binding energies < −5.0 kcal/mol and an average binding affinity approaching −8.0 kcal/mol, and the corresponding bioactive compounds were presented in Table 4 . The compound Chrysin-7-O-glucuronide exhibited the highest binding affinity with PTGS2 (−10.8 kcal/mol). Closely following were 21-Deoxyneridienone B with TNF (−9.6 kcal/mol), Estrone sulfate with ESR1 (−9.4 kcal/mol), and Chrysin-7-O-glucuronide with EGFR (−9.3 kcal/mol). For visualization, the conformation with the lowest binding energy was selected for multi-ligand targets. In the case of single-ligand targets, which included only two, namely BCL2 with Glycosolone and IL1β with 13-Deoxy-Isogibberellic acid, the complex structures were used, having binding energies of −6.4 kcal/mol and −6.6 kcal/mol, respectively ( Figure 6B ). These findings suggest that the bioactive compounds of DS effectively bind to the active sites of these hub targets, highlighting their potential in treating adenomyosis through these targets. Table 4 Details of the Compounds in the Compound-Core Target-Key Pathway Network of DS-Adenomyosis NO. Component Name PubChem CID Molecular Weight g/mol CAS CHEBI/ChEMBL ID 1 8-epi-Chlorajapolide F 131676071 276.33 863301-69-3 CHEBI:3960896 2 Neochlorogenic acid 5280633 354.31 906-33-2 CHEBI:16384 3 Evernic acid 10829 332.3 537-09-7 CHEBI:111284 4 Tagitinin F 5281501 348.4 59979-57-6 CHEBI:9390 5 Glycosolone 155059 273.33 67879-81-6 – 6 21-Deoxyneridienone B 16104854 328.4 924910-83-8 CHEMBL376854 7 Foetidin 15945065 382.5 89900-57-2 CHEMBL1362201 8 Monomethyl lithospermate 25256837 552.5 933054-33-2 CHEMBL473868 9 8alpha-Methacryloyloxybalchanin 102004556 332.4 104021-39-8 – 10 1-O-Acetyl britannilactone 25018668 308.4 681457-46-5 CHEMBL274543 11 Syringin 5316860 372.37 118-34-3 CHEBI:9380 12 Chrysin-7-O-glucuronide 14135335 430.4 35775-49-6 CHEBI:181485 13 7-Methoxy-4-methylcoumarin 390807 190.19 2555-28-4 CHEBI:107662 14 Crebanine 159999 339.4 25127-29-1 CHEBI:228860 15 Ethyl arachidonate 5367369 332.52 1808-26-0 CHEBI:84873 16 Estrone sulfate 3001028 350.4 481-97-0 CHEBI:17474 17 6′-O-Acetylglycitin 10228095 488.4 134859-96-4 CHEBI:133348 18 13-Deoxy-Isogibberellic acid 51136328 330.4 – – 19 Daucoidin A 6479092 344.4 103629-87-4 – 20 3,5-Diprenyl-4-hydroxybenzaldehyde 71519781 258.35 52275-04-4 – 21 Sucrose 5988 342.3 57-50-1 CHEBI:17992 Figure 6 Validation and screening of molecular docking. ( A ) Molecular docking hotspot map of core targets and key components. Binding energies (kcal/mol) of core targets and active compounds of DS. ( B ) Docking patterns of core targets and active compounds of adenomyosis. Details of the Compounds in the Compound-Core Target-Key Pathway Network of DS-Adenomyosis Validation and screening of molecular docking. ( A ) Molecular docking hotspot map of core targets and key components. Binding energies (kcal/mol) of core targets and active compounds of DS. ( B ) Docking patterns of core targets and active compounds of adenomyosis. Transcriptomic profiling was carried out on adenomyosis-related datasets of GEO database ( GSE193928 and GSE190580 ), Using the “limma” R package to identify potential targets ( Supplementary Tables 2 and 3 ). The differentially expressed genes (DEGs) were visualized with volcano plots, where red dots indicated up-regulated genes and blue dots represented down-regulated ones ( Figure 7A ). A heatmap showcased the distribution of DEGs, emphasizing across different groups ( Figure 7B ). Integration of drug-target and disease-gene analyses identified PTGS2, IL-1β, TNF , and MMP9 as core genes ( Figure 7C ). Combined with the network pharmacology findings, IL-1β and TNF were confirmed as hub nodes regulating the TNF, IL-17, and HIF-1α signaling pathways. Overall, these results confirm that DS exert therapeutic effects on adenomyosis through targeting TNF and IL-1β . This integrative multi-omics approach offers valuable insights into the molecular mechanisms underlying DS’s multi-target therapeutic efficacy. Figure 7 Validation of core targets for DS in treating adenomyosis from GEO datasets. ( A ) Volcano plot illustrating differentially expressed genes. ( B ) Hierarchical clustering heatmap of the DEGs. ( C ) Venn diagram demonstrating the intersection between DS-adenomyosis targets and DEG datasets. Red font highlights IL-1β and TNF as hub node that coregulate the TNF/IL-17/HIF-1α signaling pathways in subsequent studies. Validation of core targets for DS in treating adenomyosis from GEO datasets. ( A ) Volcano plot illustrating differentially expressed genes. ( B ) Hierarchical clustering heatmap of the DEGs. ( C ) Venn diagram demonstrating the intersection between DS-adenomyosis targets and DEG datasets. Red font highlights IL-1β and TNF as hub node that coregulate the TNF/IL-17/HIF-1α signaling pathways in subsequent studies. Since molecular docking only provides a static interaction state between compounds and proteins, 43 we further conducted 100 ns MD simulations to explore the dynamic movement and stability of key targets with the DS active components (TNF-21-deoxyneridienone B, and IL-1β-13-Deoxy-Isogibberellic acid), ensuring the feasibility of investigation into DS-adenomyosis therapeutic potential. The TNF complex ( Figure 8A ) exhibited stable RMSD (1.5–2.0 Å), showing a mild increase compared to unliganded TNF (1.0–1.5 Å) while maintaining an overall low fluctuation threshold (<2.0 Å), indicative of limited conformational adjustments upon ligand binding. The complex exhibited RMSF fluctuations similar to unbound TNF, with the majority of residues displaying modest fluctuations (0.5–2.0 Å), indicating well-maintained global backbone stability. RMSF profiling revealed a characteristic flexibility peak (4.0–4.3 Å) at residues 75–85, likely corresponding to functional loop dynamics. Comparable Rg (1.55–1.58 nm) and SASA (78–83 nm²) with the apo protein confirmed preserved tertiary structure integrity. A stable hydrogen-bonding network (1–2 persistent H-bonds) further supported dynamic equilibrium at the binding interface. This IL-1β complex ( Figure 8B ) demonstrated greater conformational plasticity, with RMSD (3.5–4.5 Å) significantly higher than unbound IL-1β (2.0–3.0 Å), yet maintaining controlled fluctuations (<10% deviation). The complex exhibited RMSF fluctuations similar to unbound IL-1β, with the majority of residues displaying modest fluctuations (0.5–2.5 Å), indicating well-maintained global backbone stability. The prominent RMSF peak (4.5–5.0 Å) at residues 50–60 suggested localized flexibility enhancement, while subtle increases in Rg (1.57–1.60 nm vs 1.53–1.56 nm) and SASA (90–98 nm² vs 85–93 nm²) reflected moderate structural expansion. The hydrogen-bonding pattern (1–2 primary H-bonds) mirrored that of the TNF complex, demonstrating analogous binding stability. Thus, MD simulation results confirm the stable binding of the TNF-21-deoxyneridienone B and IL-1β-13-Deoxy-Isogibberellic acid complexes, demonstrating high consistency with the molecular docking predictions. These findings further confirm that the ligand-induced stabilization of target protein conformations is a crucial molecular mechanism by which DS regulates inflammatory targets in adenomyosis. Figure 8 Molecular dynamics simulations of bioactive compounds and target proteins. ( A ) TNF and 21-Deoxyneridienone B. ( B ) IL-1β and 13-Deoxy-Isogibberellic. Molecular dynamics simulations of bioactive compounds and target proteins. ( A ) TNF and 21-Deoxyneridienone B. ( B ) IL-1β and 13-Deoxy-Isogibberellic. A detailed flowchart of the in vivo assays is presented in Figure 9A . H&E staining results indicated that compared to the control group, the adenomyosis model group exhibited disordered endometrial architecture with glandular and stromal cell invasion into the myometrium. Intervention with Gestrinone and DS effectively ameliorated the structural disorganization of the endometrium to varying extents, with the therapeutic effects of DS demonstrating a dose-dependent manner ( Figure 9B ). Figure 9 DS effectively ameliorated the pathological changes associated with adenomyosis. ( A ) The diagram of animal experiments. ( B ) H&E staining of Control, Model, Gestrinone, DS-H, DS-M, and DS-L (Scale bar: 100 μM and 200 μM, n = 3). DS effectively ameliorated the pathological changes associated with adenomyosis. ( A ) The diagram of animal experiments. ( B ) H&E staining of Control, Model, Gestrinone, DS-H, DS-M, and DS-L (Scale bar: 100 μM and 200 μM, n = 3). The core targets within the TNF-α, IL-17, and HIF-1 signaling pathways include TNF-α, IL-17A, IL-1β, and HIF-1α. qRT-PCR and WB analyses demonstrated that the expression levels of TNF-α, IL-17A, IL-1β, and HIF-1α at the mRNA and protein levels were significantly elevated in the adenomyosis model group. Treatment with DS resulted in a concentration-dependent reduction in the expression of these targets ( Figure 10A, B and Supplementary Figure 1 ). IF analysis of TNF-α and IL-17A expression ( Figure 10C ) further validated these findings, illustrating that DS ameliorates adenomyosis, which aligns with the outcomes observed in qRT-PCR and WB analyses. These findings suggest that DS effectively mitigates adenomyosis through modulation of the TNF-α, IL-17, and HIF-1 signaling pathways. Figure 10 qRT-PCR, WB, and IF method verification for key targets in signaling pathways. ( A ) qRT-PCR of TNF-α, IL-17A, IL-1β, and HIF-1α mRNA expression. ( B ) Western Blot detection of TNF-α, IL-17A, IL-1β, and HIF-1α protein expression. ( C ) Immunofluorescence staining shows the fluorescent expression of TNF-α and IL-17A (Scale bar: 50 μM, n = 3). Data were expressed as mean ± SD (n = 3) * p <0.05, ** p <0.01, *** p <0.001, **** p <0.0001 VS Model. qRT-PCR, WB, and IF method verification for key targets in signaling pathways. ( A ) qRT-PCR of TNF-α, IL-17A, IL-1β, and HIF-1α mRNA expression. ( B ) Western Blot detection of TNF-α, IL-17A, IL-1β, and HIF-1α protein expression. ( C ) Immunofluorescence staining shows the fluorescent expression of TNF-α and IL-17A (Scale bar: 50 μM, n = 3). Data were expressed as mean ± SD (n = 3) * p <0.05, ** p <0.01, *** p <0.001, **** p <0.0001 VS Model.

DS was purchased from Kangmei Pharmaceutical Co., Ltd (Guangdong, China). Ultra-high liquid chromatography (Thermo Fisher Scientific, Model: Vanquish Flex UHPLC). High-resolution mass spectrometer (Q Exactive HFX, Thermo, USA). Chromatographic column (Waters HSS T3 (100×2.1 mm, 1.8 μm)). High-speed freeze centrifuge (Hettich, Germany, model: Mikro 220R), Ultrasonic extractor (Kunshan Ultrasonic Instrument Co., Ltd., model: KQ3200D). Antibodies mentioned in the experiments included tumor necrosis factor-α (TNF-α, Affinity, AF7014), interleukin-17A (IL-17A, Affinity, DF6127), interleukin-1β (IL-1β, Affinity, AF5103), hypoxia-inducible factor-1α, (HIF-1α, Affinity, AF1009), glyceraldehyde-3-phosphate dehydrogenase (GAPDH Affinity, AF7021), β-actin (Affinity, AF7018). The extraction of DS was performed according to a previously established method. 27 Briefly, 90% ethanol was added to DS and refluxed for 1.5 hours, followed by filtration and recovering the ethanol from the filtrate to a thick paste. Boil the medicinal residue in water for 1 hour, filter, and combine the filtration with the aforementioned thick paste. Concentrate under reduced pressure to 1g/mL. Take 200 μL of DS extract sample, add 600 μL of methanol, sonicate for 30 minutes, and centrifuge at 4°C and 12000 rpm for 10 minutes. Take 100 μL of solution and place it in an injection bottle for testing. Thirty ICR mice were maintained in an SPF environment of 22 ± 2°C and 40–70% humidity. All experimental procedures involving animals were reviewed and approved by the Experimental Animal Ethics Committee of Fujian Medical University (Approval NO. IACUC-FJMU-2023-Y-0804) in strict accordance with the national standard “Guidelines for Ethical Review of Animal Welfare” (GB/T 35892–2018) of China. After one week of acclimatization, all mice were randomly divided into six groups (n = 6): control (Saline solution daily), model, positive drug (Gestrinone, 0.325mg/kg, Twice a week), and drug intervention groups (7.8 g/kg for the high-dose group [DS-H], 3.9 g/kg for the medium-dose group [DS-M], and 1.3 g/kg for the low-dose group [DS-L]). The doses of DS were determined through literature review and clinical evidence, 28 showing: 1) Clinical danshen dosage ranges 10–37.3 g/day; 2) Function-dependent variations exist: blood circulation (12–30 g), pain relief (10–37.3 g), abscess treatment (15–30 g), and sedation (10–15 g). Considering adenomyosis’ “blood stasis and internal obstruction” pathology, we selected three human-equivalent doses (15, 30, 60 g) converted to mouse doses (1.3, 3.9, 7.8 g/kg) using body surface area normalization (70 kg human × 9.1 conversion factor). These doses, proven effective in other disease models, 29 were well below safety thresholds (80 mg/kg) and toxicity levels (320 mg/kg), ensuring intervention safety. 30 Select age-matched ICR male mice, euthanize them, and perform a craniotomy to extract the pituitary gland, creating a pituitary saline suspension; for female mice, administer propofol injection at a dose of 0.1 mg/g via intraperitoneal injection for anesthesia, sterilize the lower abdomen, make a 2 cm longitudinal incision slightly to the right of the midline, isolate the right uterus, and use a catheter to inject the pituitary saline suspension intrauterine, followed by the application of gentamicin solution (0.25 mL, 10000 U) to prevent infection. In the identification IAIBs of DS group, the gastric dose of DS was 100 mg/kg/d, the blank group was gavage for 7 days, fasting for 12 hours before the last time, anesthesia for 1 hour, cardiac blood was collected, packaged in a centrifuge tube without sodium heparin, stand for 1 hour, 4°C, 3000 rpm centrifugation for 10minutes, the upper serum was divided and stored in −80 °C. Chromatography conditions: Column: Waters HSS T3 (1002.1 mm, 1.8 μm); Mobile phase A (0.1% formic acid in water) and B (0.1% formic acid in acetonitrile); Flow rate 0.3 mL/min, column temp 40°C, injection volume 2 μL. Elution gradient: 0 min A phase/B phase (100:0, v/v), 1 min Phase A/B phase (100:0, v/v), 12 min A/B (5:95, v/v), 13 min Phase A/Phase B. (5:95, v/v), 13.1 min Phase A/Phase B (100:0, v/v), 17 min Phase A/B Phase (100:0, v/v). Mass spectrometry conditions: the primary and secondary maps were collected using the Thermo Q Exactive HFX high-resolution mass spectrometry system, equipped with electric spray ion source (ESI), sheath gas 40 arb, auxiliary gas 10 arb, ion spray voltage+3000 V/-2800 V, temperature 350°C, ion transport tube temperature 320°C. The scanning mode was Full-ms-ddMS 2 and scanned in positive/negative ions. The scanning range of primary mass spectrometry is 70–1050 Da, with a primary resolution of 70000 and a secondary resolution of 17500, collision energy 10/30/60 V. Thirty-five compounds were identified, and 2D structures were sourced from PubChem ( https://pubchem.ncbi.nlm.nih.gov/ ). Targets were gathered via Swiss target prediction ( http://www.swisstargetprediction.ch/ ). GeneCards ( https://www.genecards.org/ ), OMIM ( https://www.omim.org/ ) and PharmgKB ( https://www.pharmgkb.org/ ) searched for “Adenomyosis” targets. Venn tool identified intersections of compound targets. The shared target genes between DS components and adenomyosis were analyzed using the STRING 10.5 database ( https://cn.string-db.org/ ) for conducting PPI network analysis, with the species parameter set to “Homo sapiens.” The interaction threshold was configured to “medium confidence (> 0.4)”, while other parameters were maintained at their default settings. The resulting interaction data were exported in TSV format and imported into Cytoscape 3.7.2 for network construction. During this process, isolated nodes were systematically removed to refine the network. Module analysis using the MCODE plugin identified densely connected regions, with the highest-scoring module chosen for further investigation. 31 For the identification of key molecular hubs, the CytoNCA plugin was employed to extract the top core interacting targets based on network centrality measures. Analysis of these core interacting targets through Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis was conducted in DENOVO ( https://www.omicshare.com ), 32 setting p < 0.05 as the threshold to exclude cellular components (CC), biological processes (BP), molecular functions (MF), and enrichment pathways. To elucidate the intricate relationships among “herbs-compounds-targets-pathways”, we incorporated the common core target genes from DS active components and adenomyosis into Cytoscape 3.7.2 for detailed analysis. Through KEGG enrichment analysis, we initially identified key pathways involved in the efficacy of DS against adenomyosis and subsequently mapped the core targets associated with these pathways. We then pinpointed the DS active compounds that correspond to the identified core targets. Leveraging the analytical capabilities of Cytoscape 3.7.2, we established the “DS Components-Targets-Adenomyosis-Pathways” network, aiming to comprehensively elucidate the underlying mechanisms of DS’s therapeutic action in adenomyosis. The crystal structure of the core proteins was obtained from the Protein Database (PDB) database ( http://www.rcsb.org/ ). The core compounds in the network diagram of “DS-Active Compounds-Intersection Targets” were obtained in SDF file format from the PubChem database ( https://pubchem.ncbi.nlm.nih.gov/ ) to download. OpenBabel software was used to convert the SDF file into mol2 format. AutoDock Tool software 33 performed hydrogenation, ligand detection, and other operations before saving it in pdbqt format. Pymol software 34 removed its ligands and water molecules. The binding energy was calculated in Autodock-1.5.7 software. PyMol-2.5.2 software visualized the docking results. Binding energy values indicate the affinity of the active compound for its protein targets. The stability of ligand-receptor binding was inversely correlated with binding energy. 35 Transcriptomic analysis of two adenomyosis datasets, GSE193928 and GSE190580 , was conducted using differential expression analysis from the Gene Expression Omnibus database (GEO: https://www.ncbi.nlm.nih.gov/geo/ ), following established research methods. 36 The “limma” R package was used, setting threshold values at |logFC| > 2.0 and P < 0.05 to identify significant differential expression. The results were visualized using volcano plots and heatmaps, created with the “ggplot2” and “pheatmap” R packages, respectively. Additionally, a Venn diagram was employed to identify common targets between the potential targets of DS-adenomyosis and those found in the two adenomyosis datasets. This study performed all-atom Molecular dynamics (MD) simulations on the protein-ligand complexes derived from molecular docking using GROMACS 2022.4. The AMBER14SB force field was applied to parameterize the protein, while the topology files for small molecules were generated using ACPYPE and Antechamber tools. The system was solvated in a cubic water box with a 1 nm buffer distance between the complex and box edges, utilizing the TIP3P water model. Sodium and chloride ions were added to neutralize the system charge. The energy minimization was conducted via the steepest descent algorithm, followed by temperature equilibration (300 K) under the NVT ensemble and pressure equilibration (101.325 kPa) under the NPT ensemble. Subsequently, production MD simulations were carried out for 100 ns at 300 K, generating 10,000 trajectory frames. Key structural and interaction parameters, including root-mean-square deviation (RMSD), root-mean-square fluctuation (RMSF), solvent-accessible surface area (SASA), radius of gyration (Rg), and intermolecular hydrogen bond dynamics, were systematically analyzed from the trajectories. Higher RMSD values and large-amplitude fluctuations indicate increased conformational flexibility, while lower RMSD values suggest improved structural stability throughout the simulation trajectory. The RMSF is a key metric for assessing local flexibility in molecular dynamics simulations. 37 Mice uterine tissues were collected and fixed in 4% paraformaldehyde, dehydrated through a graded series of ethanol, cleared with xylene, and embedded using an embedding machine. Sections of 5μm thickness were cut and stained with hematoxylin and eosin (H&E). The histological changes were observed under a light microscope at 200× and 400× magnifications. Cellular RNA was extracted with Trizol (Invitrogen, USA) per the manufacturer protocol. cDNA was synthesized using the PrimeScript RT kit (Biosharp, China). mRNA levels of GAPDH, TNF-α, IL-17A, IL-1β, and HIF-1α were quantified by qPCR with SYBR Premix Ex Taq (Biosharp, China). Primer sequences are in Table 1 , verified by BLAST for specificity. Data were normalized to GAPDH and analyzed using the 2 (−ΔΔCT) method. Table 1 Primer Sequence Gene Primer TNF-α Forward 5′-CAGGCGGTGCCTATGTCTC-3′ Reverse 5′-CGATCACCCCGAAGTTCAGTAG-3′ IL-17 Forward 5′-TCAGCGTGTCCAAACACTGAG-3′ Reverse 5′-CGCCAAGGGAGTTAAAGACTT-3′ IL-1β Forward 5′-GAAATGCCACCTTTTGACAGTG-3′ Reverse 5′-TGGATGCTCTCATCAGGACAG-3′ HIF-1A Forward 5′-TCTCGGCGAAGCAAAGAGTC-3′ Reverse 5′-AGCCATCTAGGGCTTTCAGATAA-3′ GAPDH Forward 5′-AGGTCGGTGTGAACGGATTTG-3′ Reverse 5′-GGGGTCGTTGATGGCAACA-3′ Primer Sequence Proteins from uterine tissue were extracted using RIPA buffer (Beyotime, China). They were separated via SDS-PAGE, transferred to PVDF membranes, and blocked with 5% milk at room temp. Incubate the cell membrane overnight with primary antibodies including TNF-α (1:500), IL-17A (1:500), IL-1β (1:1000), and HIF-1α (1:1000) at 4°C. Subsequently, the membrane was incubated with HRP conjugated secondary antibody at room temperature for 1 hour. ECL Plus (Thermoscience) was used for detection, and Image J for densitometry analysis. For IF analysis, paraffin-embedded sections of Uterine tissue were incubated with primary detection antibody TNF-α (1:200) and IL-17A (1:30) at 4°C, followed by a fluorescent secondary antibody for 1 h in the dark. Sections were counter-stained with DAPI and viewed under a fluorescent microscope (Leica DMi8, Solms, Germany). SPSS 26.0 statistical software was used to analyze the data. All data plots were processed using Graphpad Prism 9.0.0 (Graphpad Software, La Jolla, CA) software. For comparisons involving more than two groups, ANOVA followed by LSD test was applied. All data were represented as mean ± standard deviation (SD) and p < 0.05 was considered statistically different. Statistical analysis was performed by one-way analysis of variance (ANOVA) tests. Each experiment was repeated at least three times independently to confirm the consistency of the findings. The level of statistical significance was set at p < 0.05 for all analyses. Significance levels in the figures were denoted as follows: p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), and p < 0.0001 (****).

This study aimed to preliminarily investigate the potential active components of DS and their mechanisms of action against adenomyosis by integrating various pharmacology methodologies with experiment confirmed. The UPLC-QE-MS identified 35 potential bioactive from IAIBs of DS. Network pharmacology predicted that DS might exert its therapeutic effects on adenomyosis by modulating TNF-α, IL-17, and HIF-1α signaling pathways associated with inflammation. Molecular docking results demonstrated that the active components have strong affinities with core targets including TNF, IL1β, MMP2, ESR1, PTGS2, STAT3, BCL2, AKT1, MMP9, and EGFR . Transcriptomic validation identified TNF and IL-1β as key therapeutic targets, which was further corroborated by MD simulations demonstrating stable binding conformations between DS bioactive compounds and these pivotal inflammatory targets. In Vivo studies confirmed that DS may exert therapeutic effects against adenomyosis by modulating TNF-α, IL-17, and HIF-1α singing pathways of inflammatory, consistent with network pharmacology predictions. These findings offer compelling evidence of DS’s therapeutic potential in modulating the inflammatory pathogenesis and support DS as a promising candidate for adenomyosis. Adenomyosis is a chronic inflammatory disorder characterized by dysregulated cytokine expression with activating inflammatory pathways, particularly the overproduction of pro-inflammatory mediators (eg, TNF, IL-1β, IL-17) and impaired anti-inflammatory responses in adenomyotic lesions. 39–41 , 44 These imbalances sustain a pathological microenvironment that promotes ectopic endometrial invasion and disease progression. 45 Current hormonal therapies remain limited by side effects and high recurrence rates, 46 necessitating safer alternatives targeting inflammatory pathways. Furthermore, our transcriptomic profiling confirm dysregulated IL-1β and TNF expression patterns in adenomyotic ectopic lesions, establishing these inflammatory mediators as mechanistically grounded therapeutic targets. This aligns with the “multi-component, multi-target, and multi-pathway” paradigm of TCM, which has shown efficacy in complex gynecological diseases. 47 Specifically, DS is a widely used TCM for multiple diseases, which exerts pleiotropic effects by coordinately modulating inflammation, coagulation, and fibrosis. 48 Its clinical benefits for adenomyosis may stem from anti-inflammatory effects in multiple organs, including selective downregulation of key cytokines (IL-1β and TNF-α), 49 , 50 modulation of TNF-α/IL-17/HIF-1α pathways, 51 , 52 and synergistic actions across multiple pathological axes. 53 This multi-target engagement profile of anti-inflammatory pathways provides a mechanistic basis for DS’s anti-adenomyosis effects, distinguishing it from single-target hormonal therapies. Further investigation of critical inflammatory pathways may provide deeper insights into DS’s efficacy and applications for adenomyosis. Our network pharmacology and KEGG enrichment analysis identifies the TNF-α and IL-17 signaling pathways as primary inflammatory pathways of DS-adenomyosis. These pathways are pivotal in adenomyosis-associated inflammation, where aberrant TNF and IL-17 expression recruits immune cells, amplifies inflammatory cascades, and exacerbates tissue damage. 44 By suppressing these pathways, DS may restore cytokine equilibrium, attenuate lesion inflammation, and impede disease advancement. This mechanistic rationale is supported by DS’s ability to alleviate dysmenorrhea 54 and pelvic inflammation, 55 offering a theoretical foundation for its repurposing in adenomyosis therapy. Furthermore, DS’s therapeutic potential extends to mitigating hypoxia-driven pathology via the HIF-1α signaling pathway. 56 Adenomyosis lesions often exhibit hypoxic and oxidative stress conditions, which activate HIF-1 to promote angiogenesis, metabolic reprogramming, and neuroinflammation-processes linked to pain and disease aggravation. 57 , 58 DS’s putative inhibition of HIF-1 could disrupt this vicious cycle, reducing hypoxia-induced tissue damage and inflammatory responses. 59 Together, these findings position DS as a promising candidate for adenomyosis treatment, leveraging its multi-pathway anti-inflammatory actions to address both symptoms and underlying pathogenesis. Future research should validate these targets experimentally and explore DS’s synergies with conventional therapies to optimize clinical outcomes. Additionally, an integrated computational analysis including molecular docking and MD simulations were used to evaluate drug-target stability to validate bioactive compound screening. 43 This state-of-the-art computational framework provides a technological platform for exploiting the drug repurposing potential of phytochemical constituents. 60 Our quantitative assessment revealed that DS-derived phytochemicals exhibited potent binding interactions with key inflammatory mediators of TNF-α and IL-1β, achieving a mean binding energy of −8 kcal/mol. This significantly exceeds the established thresholds for biological activity (−5.0 kcal/mol) and strong binding affinity (−7.0 kcal/mol). 61 MD simulations are a powerful computational tool for investigating protein interactions with drug-like small molecules at atomic resolution. 62 They have become indispensable in modern drug design, playing a pivotal role in elucidating the dynamic behavior and conformational changes of protein-ligand complexes at atomic resolution. 63 Through 100-ns MD simulations, we observed a mild increase in the RMSD of the TNF-21-Deoxyneridienone B complex (1.5–2.0 Å) compared to unliganded TNF (1.0–1.5 Å), yet it remained significantly below 2 Å. Importantly, the complex maintained a notably low mean RMSF, indicating that this controlled enhancement of structural flexibility might facilitate signal transduction while preserving the overall structural integrity of the protein-ligand complex. These results are consistent with the stability data by reported McMillan et al 64 for TNF-α/small-molecule complexes. In contrast, the IL-1β-13-Deoxy-Isogibberellic acid complex exhibited higher RMSD values (3.5–4.5 Å) compared to unbound IL-1β (2.0–3.0 Å) while maintaining fluctuations within 10% of the mean. Notably, the average RMSF (1.34 Å) for the complex remained predominantly below 2 Å, suggesting a well-maintained structural framework despite increased global flexibility, aligning with previous studies. 65 These results confirmed that DS active compounds directly binds to TNF and IL1β to regulate inflammation and anti-adenomyosis. This comprehensive computational validation not only elucidates the therapeutic mechanisms of DS at the atomic level but also establishes a theoretical framework for developing multi-target anti-inflammatory therapies for adenomyosis. In summary, inflammation plays a crucial role in the pathogenesis of adenomyosis, 66 characterized by chronic inflammatory responses that trigger local immune cell activation and sustained release of inflammatory mediators. 39 , 67–69 Proinflammatory cytokines such as TNF-α, IL-1β, IL-6, and IL-17 promote the proliferation and migration of endometrial cells, contributing to adenomyosis progression. 39 , 70 Additionally, inflammation-induced angiogenesis provides nutritional support for ectopic lesions, exacerbating disease advancement. 71 Using an established adenomyosis mouse model, we demonstrated that DS can restore endometrial tissue architecture and significantly reduce glandular-stromal invasion into the myometrium, as demonstrated by H&E staining. Further molecular analyses, including qRT-PCR, WB, and IF, confirmed that DS mediates the downregulation of TNF-α, IL1β, IL-17, and HIF-1α expression at both transcriptional and translational levels showing statistical significance. These findings suggest that DS exerts therapeutic effects of anti-inflammatory through coordinated modulation of TNF-α/IL-17/HIF-1α signaling pathways, transcending the limitations of single-target drugs. The observed efficacy arises from DS’s unique polypharmacological capacity to simultaneously regulate multiple signaling cascades and molecular targets, resulting in synergistic therapeutic outcomes. 72 Our study provides compelling evidence that DS ameliorates adenomyosis by targeting TNF-α/IL-17/HIF-1α singing pathways, thereby establishing a solid foundation for the clinical translation and development of novel multi-target therapeutics for adenomyosis. The multi-component and multi-target nature of TCM offers therapeutic advantages for DS-adenomyosis but also poses challenges due to inherent variability in bioactivity and mechanisms. Consequently, the current study has limitations. First, research on how these compounds collectively contribute to the treatment of adenomyosis remains insufficiently explored, particularly regarding the mechanisms of action of key bioactive components in DS and the specific details of their synergistic effects, which require further investigation. Second, only representative molecules from key biological functional modules were selected for analysis in the molecular validation experiments. This selection may limit a comprehensive and in-depth understanding of the mechanistic actions of DS. Future studies should focus on the monomeric compounds of DS identified validate this study’s findings, employing gene knockout rats and rescue experiments targeting potential pathways to strengthen the evidence base. Additionally, while our target screening covered major genetic and pharmacological databases, the rapidly evolving nature of disease-gene annotations warrants periodic updates. Future studies could incorporate emerging resources like Open Targets when more adenomyosis data become available. Although the safety and stability of DS, a plant-based TCM with a long history, have been established, well-designed clinical studies are essential. This includes randomized controlled trials with standardized protocols and rigorous outcome assessments to validate the therapeutic efficacy and safety of DS in managing adenomyosis. While our study demonstrates DS’s therapeutic potential in adenomyosis through modulation of TNF-α/IL-17/HIF-1α signaling pathways, certain limitations should be acknowledged. The current findings, primarily derived from animal models and bioinformatics analyses, would benefit from: (1) mechanistic validation using genetic perturbations (eg, knockout/knockin) or pharmacological interventions; (2) in vitro confirmation of its dual regulatory (agonist/antagonist) effects; and (3) translational studies including dose-response assessments, PK/PD analyses, and validation in patient-derived models (eg, organoids/xenografts) to strengthen clinical relevance. These aspects will be prioritized in future investigations.

This study systematically elucidates the multi-target synergistic mechanisms of DS bioactive compounds against adenomyosis using an integrated approach that combines pharmacologic, transcriptomic, and in vivo analyses. These findings suggest that DS exerts its therapeutic effects on adenomyosis by modulating inflammatory factors through multiple pathways and targets, including the TNF-α, IL-17, and HIF-1α signaling pathways. This study indicates that DS is a promising option within TCM for treating adenomyosis. Our elucidation of the specific molecular mechanism of DS anti-adenomyosis and its effective active ingredients provide a theoretical foundation for enhancing the clinical application of TCM, thereby supporting future clinical research and drug development.