Abstract
To elucidate the molecular mechanism by which xanthohumol (XN) alleviates rheumatoid arthritis (RA), this study integrated network pharmacology prediction, proteomic screening, and molecular dynamics simulation to systematically evaluate the inhibitory effects and targets of XN in TNF-α-induced fibroblast-like synoviocytes (MH7A cells). XN dose-dependently reversed the pathological proliferation and invasive behavior of MH7A cells and down-regulated pro-inflammatory cytokines such as IL-1β and IL-6, as well as fibrosis/angiogenesis-related factors including TGF-β1 and VEGF. Cross-validation of network pharmacology and proteomics indicated that these effects were closely associated with the inhibition of the p38/MAPK pathway, modulating inflammatory signal transduction to exert anti-RA effects. Molecular docking revealed that XN bound to the ATP-binding pocket of p38α kinase with an optimal binding energy of −7.623 kcal/mol, forming a stable hydrogen-bond network and hydrophobic stacking with the P-loop, hinge region, and catalytic residue Lys53. Molecular dynamics simulation (100 ns) combined with MM/GBSA analysis (ΔGbind = −104.14 kcal/mol) showed that the complex conformation was highly convergent, with significantly reduced RMSF of key residues in the catalytic domain, suggesting that XN induces rigidification of the kinase catalytic domain by stably occupying the ATP site, thereby blocking its phosphotransfer activity. Collectively, these results demonstrate that XN targets p38α kinase to inhibit the p38/MAPK pathway, intervening in RA pathological progression at multiple levels and providing experimental evidence for its anti-RA molecular mechanism.
1. Introduction
Rheumatoid arthritis (RA) is a chronic systemic autoimmune disease characterized by persistent synovial inflammation and progressive joint destruction, with a global prevalence of 0.5–1.0% [1]. In China, the prevalence of RA continues to rise, with more than 5 million confirmed patients, causing a heavy social and economic burden [2]. The pathological hallmark of RA is the transformation of synovial tissue into an invasive, proliferative pannus that invades and erodes adjacent articular cartilage and subchondral bone, ultimately leading to irreversible joint deformity and dysfunction [3,4]. At the cellular level, fibroblast-like synoviocytes (FLS) derived from RA patients exhibit a tumor-like phenotype, including resistance to apoptosis, abnormal proliferation, enhanced migratory capacity, and excessive secretion of pro-inflammatory cytokines and matrix-degrading enzymes [5,6]. These pathological behaviors are driven by dysregulated intracellular signaling networks, among which the mitogen-activated protein kinase (MAPK) cascade, particularly the p38/MAPK subfamily, serves as a key hub integrating inflammatory stimuli and regulating the development of synovial lesions [7]. Upon activation by pro-inflammatory cytokines such as tumor necrosis factor-α (TNF-α) and interleukin-1β (IL-1β), p38/MAPK phosphorylates downstream transcription factors, thereby amplifying the expression of IL-6, IL-8, and matrix metalloproteinases (MMPs), and promoting FLS-mediated joint destruction [8].
Despite significant advances in RA therapy, current therapeutic strategies still face multidimensional challenges, including heterogeneous efficacy and the balance between route of administration and long-term safety. Biologics represented by TNF-α inhibitors can significantly delay radiographic progression, but approximately 30–40% of patients exhibit primary or secondary non-response, and the injectable route of administration affects long-term compliance [9,10]. JAK inhibitors, represented by tofacitinib and baricitinib, provide an oral small-molecule alternative for patients with inadequate response to biologics [11]. However, studies such as ORAL Surveillance have revealed that in patients over 50 years of age with cardiovascular risk factors, the risks of major adverse cardiovascular events, venous thromboembolism, malignancy, and serious infections are significantly higher with JAK inhibitors than in the TNF-α inhibitor control group [12]; consequently, the U.S. FDA and European EMA have added boxed warnings [13]. These interventions focus on systemic immunity or broad-spectrum kinases, yet lack precise regulation of the pathological activation of the direct executors of RA joint destruction. FLS are not merely responders to inflammation, but active drivers of synovial hyperplasia, cartilage erosion, and pannus formation, and their functional abnormalities are regulated by multiple intersecting signaling pathways [14,15]. Among these, p38/MAPK occupies a central converging position in this network, and its inhibitors exert both anti-inflammatory and bone-protective effects in preclinical models, although early development was hampered by hepatotoxicity and off-target effects [16,17]. In this context, screening natural lead compounds with p38/MAPK-modulating activity and improved safety profiles may offer new directions for overcoming these dilemmas.
Xanthohumol (XN) is an important prenylated chalcone compound in the hop plant (Humulus lupulus L.) [18,19]. Beyond its traditional use in beer brewing, XN has attracted increasing attention as a multi-target natural nutraceutical due to its potent antioxidant, antibacterial, and anti-inflammatory activities [20,21]. In terms of anti-inflammation, XN has been shown to inhibit lipopolysaccharide (LPS)-induced macrophage activation, downregulate the production of IL-1β, IL-6, and TNF-α, and suppress nuclear factor-κB (NF-κB) signaling in various inflammatory models [22,23]. Notably, recent evidence indicates that XN can modulate the MAPK cascade, including inhibition of p38 phosphorylation in a context-dependent manner across different cell types [19,24]. In preclinical studies of RA, XN administration has been reported to alleviate inflammatory pain and reduce arthritis severity by inhibiting mitochondria-mediated inflammation and NLRP3 inflammasome activation [25]. Although XN has well-recognized anti-inflammatory potential, its direct effects on the activation of rheumatoid arthritis fibroblast-like synoviocytes (RA-FLS) and its precise molecular targets remain unclear. More importantly, in RA-FLS, activation of p38/MAPK is a core event driving the overexpression of IL-6, IL-8, and MMPs [26]; moreover, previous studies have demonstrated that inhibition of p38 significantly reduces the invasiveness and inflammatory cytokine secretion of RA-FLS, indicating that p38 is a therapeutically relevant target [26]. However, direct evidence as to whether XN can effectively inhibit the p38/MAPK signaling pathway and its downstream effects in this specific cellular context is still lacking.

Against this background, the present study first evaluated the anti-RA effects of XN in TNF-α-stimulated MH7A cells; network pharmacology was then used to predict the potential targets and signaling pathways of XN at a systemic level, proteomics was employed to profile the changes in protein expression following XN intervention, and molecular dynamics simulation was performed to verify the binding modes and stability of XN with key targets, thereby establishing an evidence chain of “phenotypic evaluation–target prediction–mechanistic dissection”. This study systematically investigates whether XN exerts anti-RA effects through modulating the p38/MAPK signaling pathway in TNF-α-stimulated MH7A cells, with the aim of providing experimental evidence for the development of XN as a candidate adjunctive intervention for RA.
2. Results
2.1. XN Inhibits TNF-α-Induced Proliferation and Inflammatory Responses in MH7A Cells
To establish an in vitro model of RA synovitis, MH7A cells were stimulated with TNF-α to induce an inflammatory model [27], and the anti-inflammatory potential of XN was subsequently evaluated. CCK-8 assay showed that XN at 5, 10, and 20 μmol·L−1 significantly inhibited TNF-α-induced proliferation of MH7A cells in a concentration-dependent manner, with an IC50 value of 14.06 μmol·L−1. The inhibitory effect of 20 μmol·L−1 XN was comparable to that of the positive control methotrexate (MTX) at 10 μmol·L−1 (Figure 1A,B). To clarify the regulatory effect of XN on the inflammatory cytokine secretion profile of MH7A cells, ELISA was used to detect the levels of inflammatory cytokines (IL-6, IL-1β) and growth factors (VEGF, TGF-β1) in the cell culture supernatant. TNF-α stimulation significantly upregulated the secretion of the above factors, in which IL-6, IL-1β, and VEGF increased to 1.26-, 1.27-, and 1.36-fold of the control group, respectively, while TGF-β1 increased to 3.13-fold, suggesting that TNF-α simultaneously activates inflammatory responses and angiogenesis/fibrosis-related programs in MH7A cells. After XN intervention, TGF-β1 levels decreased in a dose-dependent manner and were significantly suppressed at all three concentrations (5, 10, and 20 μmol·L−1). In contrast, the inhibition of IL-6, IL-1β, and VEGF was high-dose-dependent, with significant reductions observed only in the 10 and 20 μmol·L−1 groups, and the 20 μmol·L−1 group approaching control levels (Figure 1C–F). These results indicate that XN can simultaneously block TNF-α-induced release of inflammatory cytokines and secretion of growth factors, ameliorating the RA synovial microenvironment at both the inflammatory and angiogenesis/fibrosis levels. In addition, previous studies have reported that XN showed no obvious cytotoxicity toward normal gastric epithelial cells (GES-1) even at concentrations as high as 100 μmol·L−1 [28], and that at concentrations effective in inhibiting tumor cell proliferation, XN was significantly less toxic to normal fibroblasts (HSF) than to pathological cells [29]. Taken together with our experimental results, these findings suggest that XN may have a relatively wide safety window while effectively intervening in the pathological phenotype of RA-FLS.
Figure 1.
Effects of XN on TNF-α-induced proliferation and inflammation in MH7A cells. (A) Effects of different concentrations of XN and MTX on the viability of TNF-α-induced MH7A cells. (B) IC50 curve for the inhibitory effect of XN on MH7A cells. (C–F) Concentrations of IL-6 (C), IL-1β (D), VEGF (E), and TGF-β1 (F) in the culture supernatant of TNF-α-induced MH7A cells, as determined by ELISA (XN treatment for 24 h). Data are presented as mean ± s.e.m. of n = 3 biologically independent replicates. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001 vs. the model group (one-way ANOVA followed by Dunnett’s multiple comparison test); ns, not significant.
2.2. XN Inhibits TNF-α-Induced DNA Synthesis and Migration in MH7A Cells
Abnormal proliferation and invasive migration of synovial fibroblasts are key drivers of articular cartilage destruction in RA [30]. Therefore, we evaluated the effects of XN on MH7A cells from two aspects: DNA synthesis and cell migration capacity. EdU incorporation assay showed that TNF-α stimulation significantly increased the proportion of EdU-positive cells (approximately 28.37% higher than that of the control group), suggesting that the inflammatory microenvironment promotes MH7A cells to enter S phase. The positive control methotrexate (MTX, 10 μmol·L−1) significantly reduced the EdU incorporation rate. XN treatment (5, 10, and 20 μmol·L−1) decreased the EdU incorporation rate in a dose-dependent manner, with the 20 μmol·L−1 group reducing the EdU incorporation rate to a level close to that of the control group (Figure 2A), further supporting the inhibitory effect of XN on TNF-α-induced DNA synthesis in MH7A cells. After 24 h of TNF-α stimulation, the migration rate was significantly increased (reaching 19.87% relative to the control group), suggesting that the inflammatory model induces enhanced cell migration; MTX significantly inhibited this effect. XN treatment dose-dependently decreased the scratch wound healing rate, with the inhibition rate of cell migration in the 20 μmol·L−1 group exceeding that of the control group (Figure 2B), indicating that XN effectively reverses TNF-α-induced enhanced migration of MH7A cells. Notably, the inhibition of synthesis and migration by XN exhibited a similar dose–response relationship, suggesting that XN may simultaneously regulate these two pathogenic processes through a common upstream signaling node (such as the NF-κB or MAPK pathway). These results indicate that XN can block TNF-α-induced abnormal synthesis and migration of MH7A cells, inhibiting the pathogenic behavior of RA-FLS in both the dimensions of cell number expansion and spatial invasion.
Figure 2.
Effects of XN on proliferation and migration of MH7A cells. (A) Effects of different concentrations of XN and MTX on TNF-α-induced proliferation of MH7A cells, as detected by EdU staining: representative fluorescence images of each group are shown on the left, and quantitative analysis of the EdU-positive cell rate is shown on the right. (B) Effects of different concentrations of XN and MTX on TNF-α-induced migration of MH7A cells, as detected by scratch wound healing assay: representative light microscopy images of each group are shown on the left, and quantitative analysis of the cell migration rate is shown on the right. Data are presented as mean ± s.e.m. of n = 3 biologically independent replicates. * p < 0.05, **** p < 0.0001 vs. the model group (one-way ANOVA followed by Dunnett’s multiple comparison test).
2.3. Mechanistic Exploration of XN Against RA via Network Pharmacology
Potential targets of XN were predicted using the SwissTargetPrediction database, yielding 287 related targets. A total of 2584 RA disease-related targets were retrieved and integrated from the GeneCards, DisGeNET, and OMIM databases. Through Venn intersection analysis, 144 potential targets of XN for the treatment of RA were identified (Figure 3A). The intersection targets were imported into the STRING database to construct a PPI network, and eight core hub targets, including IL2, CCL5, IGF1, MAPK14 (p38α), MDM2, JAK2, MMP9, and MMP2, were screened out by CytoNCA topological analysis (Degree ≥ 10) (Figure 3B), covering key pathological processes of RA, including upstream inflammatory regulation and midstream signal transduction to downstream joint destruction. GO functional enrichment analysis of the 144 intersection targets showed that, at the biological process (BP) level, these targets were significantly enriched in processes such as regulation of inflammatory response, protein folding, and energy metabolism; at the molecular function (MF) level, they were mainly associated with protein kinase/phosphatase activity and transmembrane receptor protein tyrosine kinase activity; and at the cellular component (CC) level, they were mainly localized to mitochondria-associated regions, focal adhesions, and cell–matrix junctions (Figure 3C). KEGG pathway enrichment analysis further revealed that the MAPK signaling pathway, p53 signaling pathway, and PI3K–Akt signaling pathway were among the top six significantly enriched signaling pathways (Figure 3D). To further demonstrate the associations between the key pathways (top six) and target genes, a KEGG pathway–gene interaction network was constructed. This network showed that core targets such as MAPK14, JAK2, MMP9, MMP2, and IGF1 simultaneously belonged to both the MAPK signaling pathway and the PI3K–Akt signaling pathway, suggesting that XN may regulate RA-related inflammatory and proliferative signals through multi-pathway synergy (Figure 3E). Notably, MAPK14 (p38α) exhibited the highest degree of connectivity in the PPI network and was simultaneously significantly enriched as a core kinase of the MAPK signaling pathway, suggesting that p38/MAPK signaling may be a key regulatory node in the anti-RA effect of XN.
Figure 3.
Network analysis predicting the potential targets of XN against RA. (A) Venn diagram of the intersection between predicted XN targets and RA-related targets. (B) Core PPI network; darker node color indicates a higher degree value. (C) GO enrichment analysis (BP, CC, MF) of the intersection targets; bar length represents −log10(p value). (D) KEGG pathway enrichment analysis of the intersection targets; bubble size represents the number of enriched genes, and color represents −log10(p value). (E) KEGG pathway–gene interaction network; edge color represents log2FC, and the legend on the right indicates the biological functional grouping of genes.
2.4. Proteomics Reveals Differentially Expressed Proteins and Signaling Pathways in XN-Intervened RA
To validate the network pharmacology predictions at the protein level and elucidate the molecular mechanism of XN against RA, TNF-α-induced MH7A cells were used to simulate the RA synovial inflammatory microenvironment and were treated with XN (10 μmol·L−1) for quantitative proteomic analysis. PCA showed that the cumulative variance contribution of the two groups reached 76.1%, indicating that XN treatment induced global remodeling of the cellular proteome (Figure 4A). A total of 3871 proteins were identified by database searching, of which 3281 were shared between the two groups. Based on the fold change (|log2FC| > 2) and statistical significance (p < 0.05), a total of 522 differentially expressed proteins (DEPs) were obtained, including 255 upregulated and 267 downregulated proteins (Figure 4B,C). GO enrichment analysis showed that, at the biological process (BP) level, the DEPs were mainly enriched in processes such as protein folding, energy metabolism, and regulation of inflammatory response; at the molecular function (MF) level, they were mainly associated with ubiquitin ligase binding, protein phosphatase activity, and ATP binding; and at the cellular component (CC) level, they were mainly localized to mitochondria, focal adhesions, and cell–matrix junctions (Figure 4D). KEGG enrichment analysis showed that the DEPs were significantly enriched in the MAPK signaling pathway, PI3K–Akt signaling pathway, and p53 signaling pathway, while metabolism-related pathways such as carbon metabolism, biosynthesis of amino acids, and peroxisome were also significantly enriched (Figure 4E), suggesting that both inflammatory signaling and metabolic processes in MH7A cells were altered after XN intervention.
Figure 4.
Effects of XN on the proteome of TNF-α-induced MH7A cells and mechanistic analysis. (A) Principal component analysis (PCA). (B) Intersection of identified proteins between the TNF-α group and the TNF-α + XN group. (C) Volcano plot of differentially expressed proteins; red indicates upregulated proteins (255), blue indicates downregulated proteins (267), and gray indicates proteins with no significant difference. (D) GO enrichment analysis (BP, CC, MF) of differentially expressed proteins. (E) KEGG pathway enrichment analysis of differentially expressed proteins; bubble size represents the number of enriched proteins, and color represents −log10(p value).
2.5. Multi-Dimensional Cross-Validation Identifies the Core Pathway
To further screen the core molecular mechanism of XN against RA, the pathways enriched from the DEPs in the proteomics analysis were intersected with the XN anti-RA-related pathways predicted by network pharmacology, and the top 25 significantly enriched pathways were displayed. A total of two commonly enriched pathways, MAPK and PI3K–Akt, were identified (Figure 5A). Among these common pathways, the MAPK signaling pathway was significantly enriched, suggesting that it may play a hub role in linking the pharmacological effects of XN with the pathological processes of RA. Given that p38/MAPK is an important branch of the MAPK signaling pathway and plays a key regulatory role in inflammatory responses, synovial cell proliferation, and joint destruction [31], this study further focused on this branch for PPI analysis. A total of 79 functionally related molecules were involved in this pathway (Figure 5E), of which 62 were successfully quantified in the proteomics data, and 19 were significantly differentially expressed after XN intervention, including 7 upregulated and 12 downregulated proteins (Figure 5C,D). Pathway mapping results showed that these DPEs were mainly distributed at p38/MAPK-related signaling nodes, involving processes such as inflammatory response, stress response, and regulation of cell fate. Furthermore, a four-way intersection analysis was performed among RA-related targets, potential targets of XN, p38/MAPK pathway members, and proteomics-derived differentially expressed proteins. The results showed that MAPK14 (p38α) was the only core molecule simultaneously satisfying all four conditions (Figure 5B), suggesting that it may be a high-confidence candidate target for XN in regulating the pathological progression of RA.
Figure 5.
Integration of network pharmacology and proteomics to predict the key targets and pathways of XN against RA. (A) Venn diagram of DEG pathways and XN–RA pathways. (B) Venn diagram of RA targets, DEGs, XN targets, and p38/MAPK pathway genes. (C) Fold changes of differentially expressed proteins in the p38/MAPK pathway. (D) PPI network of p38/MAPK pathway-related proteins identified by proteomics. (E) p38/MAPK signaling pathway map; red nodes represent differentially expressed proteins.
2.6. Validation of the Molecular Mechanism of Direct XN Binding to p38 Protein
Molecular docking results showed that the binding energy of the optimal conformation of XN with p38/MAPK (Mode 1) was −7.623 kcal·mol−1 (Figure 6A), which was lower than that of the p38α inhibitor SB203580 under the same docking parameters (−6.741 kcal·mol−1) (Figure 6C,D), suggesting that XN has a relatively high predicted binding affinity for the ATP-binding pocket of p38 kinase. Among the nine candidate conformational modes, the affinities of Modes 1–6 were all superior to ≥−6.98 kcal·mol−1 (Figure 6E), and all were localized to the ATP-binding site, maintaining good consistency with the optimal conformation, indicating that this binding pocket possesses favorable spatial and charge complementarity. Further analysis revealed that the binding site of XN spatially overlaps to a high degree with the ATP catalytic pocket of the p38α crystal structure (PDB: 1A9U), exhibiting the typical characteristics of ATP-competitive binding. Specifically, the core scaffold of the compound occupies the hinge region and hydrophobic pocket where the ATP adenine ring resides, forming key interactions with the hinge-region residues Met109 and Gly110, and extends deeply into the p38α-characteristic hydrophobic cavity formed by Thr106, Leu104, Leu167, and other residues. Meanwhile, the side chain of the compound extends toward the ribose- and phosphate-binding region, interacting with the catalytic Lys53, Glu71, and Asp168 of the DFG motif, and lies close to residues such as Val30, Tyr35, and Val38 near the glycine-rich loop (P-loop) (Figure 6B). XN and the classical ATP-competitive p38 inhibitor SB203580 exhibit highly overlapping binding modes within the ATP-binding pocket of p38α kinase (Figure 6D), with both anchored to the ATP-binding core hotspot composed of Val38, Ala51, Lys53, Thr106, Leu108, Met109, and other residues [32]. This consistency in spatial occupancy and in key residue interactions, together with previous studies demonstrating that XN can directly bind to AKT1/2 [33] and act as an ATP-competitive inhibitor—a mechanism that may be transferable to p38α—suggests that XN may occupy the ATP-binding site of p38α (including both the core anchoring region and the peripheral extension region), thereby sterically hindering ATP binding to p38α and inhibiting its phosphotransferase activity.
Figure 6.
Molecular docking results of XN with the core targets of the p38 pathway. (A) Overall structure of XN bound to p38α and a magnified view of the binding pocket. (B) Two-dimensional schematic diagram of the interaction between XN and p38α. (C) Overall structure of the positive control SB203580 bound to p38α and a magnified view of the binding pocket. (D) Two-dimensional schematic diagram of the interaction between SB203580 and p38α. (E) Binding affinities, lower and upper RMSD bounds, and binding site annotations for different docking conformations of XN (Modes 1–6).
Further 100 ns all-atom molecular dynamics (MD) simulations showed that the RMSD of the backbone of the receptor protein reached a stable plateau after initial relaxation in the first 10 ns, with an average value of 0.335 nm over the entire simulation period. The ligand pose adjustment within the pocket reached equilibrium after 30 ns, with an average RMSD of 0.750 nm, and no dissociation or sliding escape occurred throughout the simulation (Figure 7A). After removing rigid-body translation and rotation, the RMSD of the XN scaffold conformation alone averaged only 0.264 nm, confirming that the small molecule consistently maintained a dominant rigid extended conformation within the binding cavity. These results indicate that XN does not contact p38α via transient association, but rather forms a conformationally restricted, long-residence stable complex. Cα-RMSF analysis showed that the RMSF values of key residues in the core catalytic site and binding pocket of the kinase (Lys53, Thr106, His107, Leu108, Met109, Asp168, etc.) were generally below 0.12 nm (Figure 7B), exhibiting highly rigid constraints, suggesting that XN embedding effectively restricts the thermal fluctuations of residues surrounding the binding pocket, whereas fluctuation peaks were mainly located in solvent-exposed surface loop regions and the N/C termini. Hydrogen bond monitoring showed that an average of 0.99 hydrogen bonds was maintained between the receptor and ligand during the simulation, with at least one hydrogen bond connection maintained for 75.9% of the simulation time and a maximum instantaneous number of 4 hydrogen bonds (Figure 7C). Rg analysis showed that the Rg values of three independent simulations fluctuated within approximately 2.0–2.3 nm, with an overall stable trend, indicating that the complex maintained a compact conformation throughout the 100 ns simulation (Figure 7D). The number of contact atoms at the binding interface averaged as high as 154.4 over the entire simulation period, with a maximum of 243 contacts; The average SASA of the ligand remained at 6.57 nm2 (Figure 7E), indicating that XN maintained dense interfacial packing and complementary contacts with residues lining the binding cavity and was deeply buried within the hydrophobic cleft of the kinase. Together, these dynamic metrics confirm that XN has a relatively long residence time and highly specific interfacial complementarity within the ATP pocket.
Figure 7.
Molecular dynamics simulation analysis of XN binding to p38/MAPK. (A) RMSD of the ligand (red) and protein backbone (black) as a function of time. (B) RMSF profile of protein residues. (C) Number of hydrogen bonds (black) and number of atom pairs within 0.35 nm (red) as a function of time. (D) Radius of gyration (Rg) of the protein as a function of time. (E) The ligand–protein contact number (<0.35 nm) varies with time. (F) Free energy landscape based on PC1 and PC2. (G) Three-dimensional interaction mode of the ligand in the binding pocket; key residues (Thr106, Met109, Asp168, etc.) are shown as stick models. (H) Decomposition of the binding free energy (ΔGbind) into energy components (ΔEvdw, ΔEele, etc.). (I) Energy contribution of key residues (ΔGres).
The free energy landscape revealed that the complex was highly convergent to a single global low-energy basin during the 100 ns sampling (ΔG < 2.0 kJ/mol), and the 3D surface exhibited a funnel-shaped deep well without metastable splitting or conformational bimodal drift (Figure 7F). The dominant clustered conformation (Cluster Top 1) occupied 23.0% of the simulation time, and its binding mode showed that the oxygen atom of XN’s phenolic hydroxyl bonds with the side-chain hydroxyl of Thr106 and the backbone nitrogen of Met109 (2.8–3.2 Å), the chalcone core maintained tight electrostatic coordination with Lys53, and the isopentenyl side chain and B ring extended into the hydrophobic cavity, forming strong hydrophobic burial with nonpolar side chains such as Val38, Ala51, and Leu108 (Figure 7G). MM/GBSA calculations showed that the total binding free energy ΔGbind was −104.14 ± 7.82 kcal/mol, among which the van der Waals contribution was the most significant (ΔEvdW = −79.94 ± 6.68 kcal/mol), occupying a dominant driving position; the nonpolar solvation free energy was a favorable contribution (ΔGnonpolar = −25.13 ± 3.08 kcal/mol); and the electrostatic interaction and polar solvation energy were essentially in an energy-balanced compensatory state (Figure 7H). Residue-wise energy decomposition and contact occupancy analysis jointly identified Val38 (−9.25 kcal/mol, 99.6%), Leu108 (−8.38 kcal/mol, 97.8%), Ala51 (−7.11 kcal/mol, 100.0%), Lys53 (−6.72 kcal/mol, 100.0%), Met109 (−6.66 kcal/mol, 93.3%), Val30 (−5.94 kcal/mol, 99.6%), and Thr106 (−5.53 kcal/mol, 100.0%) as the most critical binding anchor residues (Figure 7I), mainly distributed in the ATP-binding P-loop region, the kinase hinge region, and the catalytic triad region [34]. This residue contribution profile highly overlaps with the binding hotspots of the classical ATP-competitive inhibitor SB203580, further supporting the competitive inhibition mechanism of XN occupying the ATP-binding site from an energetic perspective.
In summary, XN binds into the ATP-binding pocket of p38 kinase with an optimal binding energy of −7.623 kcal/mol, forming a stable hydrogen-bond network and hydrophobic stacking with the P-loop region (Val38, Ala51), the hinge region (Thr106, Leu108, Met109), and the catalytic residue Lys53, inducing rigidification of the catalytic domain conformation, thereby directly blocking phosphotransfer activity through an ATP-competitive mechanism similar to that of SB203580.
3. Discussion
RA is a chronic autoimmune disease characterized by persistent synovial inflammation, abnormal proliferation of synovial cells, and destruction of joint structures, and its pathological process involves multiple mechanisms, including amplification of inflammatory signaling, metabolic reprogramming, and cell cycle [35]. Natural source active small molecules have become an important source for the discovery of anti-RA drug candidates owing to their structural diversity and multi-target regulatory potential; however, their direct targets and molecular mechanisms remain to be systematically elucidated [36]. XN, as a natural product with anti-inflammatory and immunomodulatory potential, has shown promise for intervening in the pathological progression of RA, but the key molecular mechanism underlying its anti-RA effect remains unclear [37]. On this basis, the present study integrated network pharmacology, quantitative proteomics evidence, and computational simulation to elucidate the key targets and action pathways of XN against RA.
The results indicate that XN exerts multi-dimensional intervention in RA-related pathological processes, and its effects are not limited to inhibiting abnormal proliferation and invasion of synovial cells, but also involve the downregulation of pro-inflammatory mediators and fibrosis/angiogenesis-related factors. Network pharmacology and proteomics results jointly pointed to the p38/MAPK signaling pathway; through multi-dimensional cross-validation of disease targets, XN targets, differentially expressed proteins, and p38/MAPK pathway members, p38α was further identified as a high-confidence candidate target. The results of molecular docking and 100 ns molecular dynamics simulations suggest that XN can stably bind to the ATP-binding pocket of p38α kinase, thereby potentially blocking its phosphotransferase activity through competitive occupancy of the ATP site. In summary, this study provides evidence that XN targets p38α to intervene in RA at three levels: cellular phenotype, omics association, and structural simulation. The rationale for prioritizing p38α as the target lies in its predominant expression in RA-FLS [38], its central role in inflammatory cytokine production, and its structural accessibility, which make it more amenable to intervention than other p38 isoforms.
Compared with previously reported natural p38 inhibitors, the binding mode of XN has its own characteristics. Although flavonoids such as quercetin and curcumin can inhibit p38 phosphorylation, they mostly act through indirect antioxidant effects or upstream kinase regulation and lack evidence of direct binding to the p38 ATP pocket [39]. In this study, molecular docking and MD simulations suggested that XN can directly embed into the ATP-binding pocket of p38α, and its binding energy was superior to that of the p38α inhibitor SB203580, suggesting that it may be a natural inhibitor that directly occupies the ATP site. In addition, the IC50 of XN was 14.06 μmol·L−1, consistent with the activity range of XN reported in breast cancer and liver cancer cells (10–30 μmol·L−1), suggesting that its anti-RA activity concentration is physiologically relevant [40]. Compared with the ATP-competitive inhibitor SB203580, XN also binds by occupying the ATP pocket of p38α, and its anchoring residues (Val38, Ala51, Lys53, Thr106, Leu108, Met109) highly overlap with the binding hotspot of SB203580, supporting the mechanism of competitive occupancy of the ATP site from an energetic perspective. However, the two differ in their chemical basis of binding: synthetic inhibitors such as SB203580 are mostly high-affinity scaffolds obtained through structural optimization, whereas XN, as a natural prenylated chalcone, may rely more on a combination of hydrophobic and hydrogen-bonding interactions for binding. It should be noted that the binding energy of XN being superior to that of SB203580 in this study only indicates a binding tendency and is not equivalent to true affinity or an advantage in isoform selectivity; moreover, this study did not systematically evaluate the selectivity of XN toward p38α/β/γ/δ or other kinases. Therefore, the proposed selectivity advantage of XN is currently only speculative and still requires validation by kinome profiling and p38 isoform selectivity experiments.
Modulating the p38/MAPK signaling pathway remains an important strategy worthy of further exploration in the field of RA therapy. Unlike interventions targeting a single inflammatory factor, p38/MAPK occupies a key node in the integration of intracellular inflammatory signals, and its inhibition can simultaneously reduce the release of multiple pro-inflammatory mediators, block abnormal proliferation and migration of FLS, and potentially intervene in their metabolic reprogramming, thereby conferring multiple potential effects encompassing anti-inflammation, anti-proliferation, and joint protection [41,42]. Early clinical studies have preliminarily validated the anti-inflammatory activity of p38/MAPK inhibitors in RA patients, and multiple phase II trials have shown that oral p38/MAPK inhibitors can reduce C-reactive protein and matrix metalloproteinase levels in the short term and improve joint swelling scores [43,44,45]. However, first-generation ATP-competitive inhibitors such as VX-745, BIRB 796, and SCIO-469 still have notable limitations, including hepatotoxicity, rash, and central nervous system adverse reactions due to the lack of cell-type selectivity, as well as adaptive tolerance indicated by the rebound of inflammatory markers after long-term administration [46]. These challenges do not negate the target value of p38/MAPK, but rather reflect the inadequacy of the traditional occupancy-driven inhibition mode, suggesting an urgent need for intervention strategies with novel mechanisms of action, higher selectivity, or allosteric regulatory properties. In this context, XN, as a natural small molecule that can stably bind to p38α and modulate p38/MAPK-related signaling, may provide a new choice for optimizing intervention approaches targeting this pathway.
An increasing number of studies have shown that natural products with prenylated chalcone structures similar to XN exhibit favorable biological activities in inflammatory diseases and osteoarticular diseases. It has been reported that such compounds can exert anti-arthritic effects by modulating classical inflammatory signaling pathways such as NF-κB, MAPK, and PI3K–Akt, thereby inhibiting synovial cell inflammatory responses, cartilage degradation, and bone destruction [47]. Unlike compounds that mainly act on a single inflammatory signaling node, XN in this study showed regulatory effects on the p38/MAPK signaling pathway. Notably, the prenylated chalcone skeleton of XN and its phenolic hydroxyl substitution pattern may be closely related to its target recognition capability and signaling regulatory characteristics; whether the anti-inflammatory activity of such compounds is also mainly mediated through the p38/MAPK signaling pathway, and how XN differs from reported natural p38α inhibitors in terms of binding mode, affinity, and selectivity, still lack experimental validation, which also provides an important basis for further structural optimization and derivative design. Therefore, this study not only expands the understanding of the anti-RA mechanism of XN, but also provides new strategies for the development of novel anti-RA candidate molecules based on the prenylated chalcone skeleton.
In rheumatoid arthritis, RA-FLS signal crosstalk exists between the p38 pathway and the PI3K–Akt pathway, and inhibition of p38 can induce compensatory activation of the Akt pathway. This is consistent with previous observations that selective p38 inhibitors have shown limited efficacy in RA clinical trials, as blockade of a single pathway is readily compensated by the redundancy of signaling networks. In the present study, in addition to the p38 MAPK signaling pathway, the PI3K–Akt signaling and metabolism-related pathways were also enriched in the network pharmacology and proteomics results, suggesting that XN is more likely to act through multi-pathway synergy with p38α as the core. These enrichment findings are correlative predictions, and their p38α-dependent versus p38α-independent effects remain to be distinguished by pathway inhibitor or gene knockdown experiments. Furthermore, the potential effects of XN on metabolic reprogramming in RA-FLS (e.g., changes in glycolysis- and oxidative stress-related proteins) suggest that metabolic regulation may be a component of its anti-RA effects; however, whether this is upstream/downstream of, or parallel to, p38α inhibition cannot currently be determined and warrants further validation by metabolomics and Seahorse assays.
This study still has certain limitations. Although molecular docking and MD simulations suggest that XN binds p38α in an ATP-competitive manner, and this inference is supported by cellular functional experiments and corroborating literature evidence, confirmation of direct binding and ATP competitiveness still requires future enzyme kinetic or biophysical experiments; the selectivity of XN for p38 isoforms and the specific regulatory mechanisms of its downstream inflammatory, proliferative, and metabolic signaling also remain to be further elucidated. In addition, the MH7A cell model cannot completely replace the complex pathological features of primary synoviocytes and animal models, and the in vivo pharmacokinetic characteristics, synovial tissue distribution, and long-term administration safety of XN still require systematic evaluation.
4. Materials and Methods
4.1. Materials
Xanthohumol (XN, purity ≥ 98%, HPLC) and methotrexate (MTX, purity ≥ 98%, HPLC) were obtained from Shanghai Yuanye Bio-Technology Co., Ltd. (Shanghai, China). Dulbecco’s Modified Eagle Medium (DMEM, high glucose) and penicillin–streptomycin (100×) were purchased from Gibco (Thermo Fisher Scientific, Waltham, MA, USA). Fetal bovine serum (FBS) was from Gibco (USA). Phosphate-buffered saline (PBS) was obtained from Biosharp Life Sciences (Beijing, China), and 0.25% trypsin–EDTA was sourced from Gibco. The Cell Counting Kit-8 (CCK-8) was purchased from Dojindo Laboratories (Kumamoto, Japan), and the EdU (Cy3) imaging kit was from APExBIO (Houston, TX, USA). Enzyme-linked immunosorbent assay (ELISA) kits for interleukin-6 (IL-6) and interleukin-1β (IL-1β) were obtained from Solarbio Life Sciences (Beijing, China), while VEGF and TGF-β1 ELISA kits were purchased from Neobioscience Technology Co., Ltd. (Shenzhen, China). The BCA protein assay kit (P0012) was acquired from Beyotime Biotechnology (Shanghai, China). All reagents were of analytical grade and used according to the manufacturers’ instructions.
4.2. Cell Culture and Cell Viability Assay
The human rheumatoid arthritis fibroblast-like synoviocyte cell line (MH7A) was purchased from Wuhan Procell Life Science & Technology Co., Ltd. (Wuhan, China). Cells were cultured in DMEM complete medium without sodium pyruvate (containing 10% fetal bovine serum and 1% penicillin–streptomycin) at 37 °C in a 5% CO2 incubator. MH7A cells at approximately 90% confluence were digested with trypsin and seeded into 96-well plates at 1 × 104 cells per well, followed by adherence culture for 24 h. The groups were treated as follows: (1) control group (serum-free basal medium); (2) TNF-α group (containing 20 ng·mL−1 TNF-α); (3) methotrexate (MTX) positive control group (containing 20 ng·mL−1 TNF-α + 10 μM MTX); and (4) XN treatment groups (containing 20 ng·mL−1 TNF-α + XN at 5, 10, and 20 μmol·L−1). After further culture for 24 h, the absorbance (OD value) of each well was measured at 450 nm according to the instructions of the CCK-8 kit, and the cell viability was calculated. Three replicate wells were set for each group, and independent experiments were repeated three times.
4.3. ELISA Analysis
MH7A cells in the logarithmic growth phase were seeded into 96-well plates at a density of 1 × 105 cells per well, with 1 mL of complete medium (containing 10% PBS and 1% penicillin–streptomycin) added to each well. The plates were incubated overnight at 37 °C in a 5% CO2 incubator to allow the cells to adhere fully. After adherence, the original medium in each well was carefully aspirated. Subsequently, grouping and drug treatment were performed as described in Section 4.2. After drug administration, the cells were further incubated under the same culture conditions for 24 h. Then, strictly following the instructions of the corresponding ELISA kits, the concentrations of IL-1β, IL-6, VEGF, and TGF-β1 in the cell supernatant were measured.
4.4. EdU Assay for Cell Proliferation
MH7A cells in the logarithmic growth phase were seeded into 96-well plates at 5 × 103 cells per well and cultured for adherence in a 37 °C, 5% CO2 incubator for 12 h. The original medium was discarded, and the grouping and treatment were performed as described in Section 4.2. After further culture for 24 h, 5-ethynyl-2′-deoxyuridine (EdU) working solution was added to each well at a final concentration of 50 μmol·L−1 and incubated for 2 h. Subsequently, cell fixation, permeabilization, and staining were performed strictly according to the instructions of the EdU imaging kit. Images were acquired using an EVOS M7000 fully automated fluorescence imaging system (Thermo Fisher Scientific, Waltham, MA, USA).
EdU-positive cell quantification: ImageJ (v1.54g, NIH, Bethesda, MD, USA) was used for cell counting. Images were converted to 8-bit grayscale, and the DAPI and EdU channels were segmented using the Threshold function with Otsu’s method. Particles were then analyzed using the Analyze Particles function with a minimum size cutoff of 300 pixels2 to exclude background debris. The EdU-positive rate was calculated as: (number of EdU-positive cells/total number of DAPI-positive cells) × 100%. At least three independent fields per group were randomly selected for image acquisition and quantification. All measurements were performed independently by two investigators blinded to group allocation. The cell proliferation rate was expressed as the percentage of positive cells relative to the total number of cells.
4.5. Scratch Wound Healing Assay
According to a previously reported method [48], MH7A cells in the logarithmic growth phase were seeded into 6-well plates at 5 × 105 cells per well. After the cells grew to a confluent monolayer, a sterile 200 μL pipette tip was used to create a vertical scratch at the bottom of each well. The cells were gently washed with PBS to remove detached cells, and the cells were then allocated to different treatment groups as described above. After 30 min of treatment, the cells were further incubated for 24 h. Cell migration at 0 and 24 h after scratching was observed using an inverted microscope at 100× magnification.
Scratch migration assay quantification: The wound area was quantitatively measured using ImageJ (v1.54g, NIH, USA). To eliminate differences in image scale, spatial calibration was first performed for each image using the “Set Scale” function. The wound area was then manually outlined along the border of the cell-free region using Polygon Selections. Migration rate was calculated as: [(wound area at 0 h − wound area at 24 h)/wound area at 0 h] × 100%. At least three independent fields per group were randomly selected for image acquisition and quantification. All measurements were performed independently by two investigators blinded to group allocation.
4.6. Screening of Common Targets of XN and RA
Potential targets of XN were predicted and collected from the SwissTargetPrediction and PharmMapper databases using “Xanthohumol” as the search term, with the GeneCards database used as a supplement, and the searched species was restricted to Homo sapiens. RA-related target genes were retrieved from the GeneCards, DisGeNET, and OMIM databases using “rheumatoid arthritis” as the search term. After merging and deduplication of the results, the XN potential target set and the RA-related target gene set were obtained. Meanwhile, the RA transcriptome dataset (GSE89408) was downloaded from the GEO database. Differential expression analysis was performed with |log2FC| > 2 and p < 0.05 as the thresholds for screening DEGs. The XN potential targets were intersected with RA-related target genes, and the resulting intersection was further intersected with DEGs to obtain candidate targets simultaneously satisfying the three conditions of drug prediction, disease association, and differential expression, which were used for subsequent PPI network construction and enrichment analysis.
4.7. Proteomics Analysis
Cells from the model group and the treatment group were collected (n = 3 biological replicates per group), the medium was discarded, and the cells were washed three times with PBS. Protein extraction and enzymatic digestion were performed according to a previously reported method [49]. Samples were analyzed by Nano-LC-MS/MS, and the raw data were processed using Proteome Discoverer software (version 2.2.0.388, Thermo Fisher Scientific, Waltham, MA, USA). The database search parameters were set as follows: trypsin digestion with cleavage only at the C-terminal side of arginine and lysine, except when followed by proline; a maximum of 2 missed cleavage sites; peptide length of 6–144 amino acids; carbamidomethylation of cysteine (Carbamidomethyl@C, mass shift of +57.021 Da) as a fixed modification; oxidation of methionine (Oxidation@M, mass shift of +15.995 Da) as a variable modification; a precursor ion mass tolerance of 10 ppm and a fragment ion mass tolerance of 0.02 Da; and at least one unique peptide identified per protein. All results were filtered at an FDR of <1%. All analyses were based on the Homo sapiens UniProt database (March 2026 version). This software generated peptide sequences, protein names/accession numbers, and fold change values, and data reliability was validated by FDR filtering (p value). Gene Ontology (GO) analysis was performed using the corresponding website (https://www.geneontology.org), and KEGG pathway enrichment analysis was performed using the KEGG database (https://www.kegg.jp/).
4.8. Construction of Protein–Protein Interaction Network
The candidate targets obtained in Section 4.6 were imported into the STRING database (v12.0; https://string-db.org/) with the species restricted to Homo sapiens and the minimum interaction confidence set to high confidence (interaction score ≥ 0.7), and the PPI network data were exported in TSV format. Network visualization was performed using Cytoscape software (v3.10.3), and the top 10 core targets were screened using the cytoHubba plugin based on the MCC (Maximal Clique Centrality) algorithm, while topological parameters such as Degree, Betweenness, and Closeness were calculated using the NetworkAnalyzer tool. Subsequently, GO and KEGG enrichment analyses of the candidate targets were performed using KOBAS 3.0 with the species restricted to Homo sapiens, using p < 0.05 and an enriched gene count ≥ 3 as the screening criteria, and the Benjamini–Hochberg-corrected FDR values were reported. Bubble plots of the top 20 KEGG pathways ranked by significance and bar plots of the top 7 GO terms for BP, CC, and MF were generated using the Bioinformatics platform or the R package ggplot2 4.0.0. Finally, by integrating node topological ranking, Gene Ontology (GO) functional enrichment analysis, KEGG enrichment analysis, and reported literature on RA inflammation, the core targets were determined for subsequent validation.
4.9. Molecular Docking
Molecular docking was performed using AutoDock Vina (v1.2.3). The human p38α receptor was based on a full-length, high-accuracy three-dimensional model (UniProt: Q16539), which is highly consistent with the catalytic pocket of the classical co-crystal structure (PDB ID: 1A9U) and avoids the missing activation loop in the crystal structure. After removal of solvent molecules, hydrogen atoms were added, and Kollman charges were assigned. The ligand XN was energy-minimized using the MMFF94 force field, assigned Gasteiger charges, and converted to pdbqt format.
The docking grid box was centered on the ATP-binding cavity (center coordinates: X = 4.961, Y = 1.428, Z = −6.792; size: 26 × 26 × 26 Å), fully covering the ATP catalytic cleft containing Lys53, Thr106, Leu108, and Met109. This pocket was selected because small-molecule kinase inhibitors generally follow an ATP-competitive mechanism, and the XN scaffold is highly compatible with the hinge region and hydrophobic vestibule. The docking results were finally visualized for three-dimensional spatial conformation and two-dimensional interactions using PyMOL (Version 3.0.3, Schrödinger, LLC, New York, NY, USA) and Discovery Studio software(Version 2019, Dassault Systèmes, Vélizy-Villacoublay, France).
4.10. Molecular Dynamics Simulation
Molecular dynamics simulations were performed in GROMACS 2022.2 according to a previously reported method [50]. The CHARMM36m force field was used for the protein, and ligand parameters were generated by CGenFF. The complex was placed in a 7 nm cubic periodic box, solvated with the TIP3P water model, and neutralized by adding Na+ and Cl− ions. After energy minimization by the steepest descent method, a three-stage heating and pre-equilibration was performed under the NPT ensemble (100, 200, and 300 K, 1 ns each), with positional restraints of 400 and 40 kJ mol−1 nm−2 applied to the protein backbone and side chains, respectively. The integration time step was 2.0 fs, and covalent bonds involving hydrogen were constrained by LINCS. After pre-equilibration, all restraints were removed, and three independent 100 ns unrestrained NPT production simulations were carried out. The temperature was maintained at 300 K using the Nosé–Hoover method, and the pressure was controlled at 1 bar using the Parrinello–Rahman method. Long-range electrostatics were treated with the PME method with a cutoff of 1.2 nm. After the simulations, trajectories were analyzed using built-in GROMACS tools, including root mean square deviation (RMSD), root mean square fluctuation (RMSF), radius of gyration (Rg), number of hydrogen bonds, and protein–ligand contacts, and combined with free energy landscape (FEL) analysis to comprehensively evaluate the conformational stability and interactions of the complex.
4.11. Statistical Analysis
Statistical analyses were performed using GraphPad Prism 10.0. All experiments were conducted with at least three independent biological replicates (n = 3 per group), and results are presented as mean ± standard error of the mean (mean ± s.e.m.). Comparisons among multiple groups were performed using one-way ANOVA followed by Dunnett’s multiple comparison test (each treatment group vs. the model group). * p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001 indicated statistically significant differences, while ns indicated no statistically significant difference.
This study demonstrates that XN, a representative bioactive constituent of hops, can inhibit the p38/MAPK pathway, alleviate TNF-α-induced pathological proliferation and invasion of MH7A cells, and downregulate pro-inflammatory factors such as IL-1β and IL-6 as well as matrix remodeling/angiogenesis-related factors such as TGF-β1 and VEGF, thereby exhibiting anti-RA activity in vitro. Cross-validation of network pharmacology and proteomics identified p38α as a high-confidence candidate target for the anti-RA effects of XN. Structural simulations suggest that XN may bind to the ATP-binding pocket of the p38α catalytic domain, with its anchoring residues (Val38, Ala51, Lys53, Thr106, Leu108, Met109) overlapping with key sites of the p38α ATP pocket, and that its stable occupancy may induce rigidification of the catalytic domain and restrict the structural dynamics associated with phosphotransfer. Evidence from three levels—cellular phenotype, omics association, and structural simulation—collectively supports XN as a candidate natural ligand targeting p38α and a potential ATP-site modulator, and provides a testable basis for subsequent validation of the XN–p38α interaction through site-directed mutagenesis of candidate residues and biophysical binding experiments, as well as for determining whether XN inhibits p38 MAPK via an ATP-competitive mechanism.
Author Contributions
Conceptualization, J.W. (Jianchao Wang) and H.L.; supervision, H.P. and D.Y.; investigation, data analysis, and writing—original draft preparation, X.M. and J.W. (Jinlong Wang); network pharmacology and molecular dynamics simulations, L.Z. and K.F.; writing—review and editing, J.W. (Jianchao Wang), H.P., D.Y. and H.L. All authors have read and agreed to the published version of the manuscript.
Funding
This research was funded by the National Key Research and Development Program of China, grant number 2022YFD1600304; the Natural Science Foundation of Hunan Province, grant number 2025JJ70606; the Scientific Research Project of the Hunan Provincial Department of Education, grant number 25C0336; and the Graduate Scientific Research Innovation Projects of Jishou University, grant numbers Jdy25075 and JDY2026045.
Institutional Review Board Statement
Not applicable.
Informed Consent Statement
Not applicable.
Data Availability Statement
The original contributions presented in this study are included in the article. Further inquiries can be directed to the corresponding authors.
Acknowledgments
The authors thank the National Key Research and Development Program of China for its support.
Conflicts of Interest
The authors declare no conflicts of interest.
References
- Klingemann, L.E.; Mikuls, T.R.; England, B.R. Epidemiology and outcomes of rheumatoid arthritis. Curr. Opin. Rheumatol. 2026, 38, 69–75. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Cao, M.Z.; Li, C.Q.; Li, M.C.; Lu, K.; Wu, C.Y.; Wang, J.X.; Wei, C.; Zhao, J.L.; Wang, Q.; Tian, X.P.; et al. Prevalence and Incidence of Rheumatoid Arthritis in Urban China: A National Population-Based Study. Int. J. Rheum. Dis. 2025, 28, e70184. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Feng, M.Z.; Chen, H.T.; Liao, L.; Huang, D.H.; Shen, J.; Xiao, L.B.; Lu, Q.G.; Xie, P.J. Cellular crosstalk and signaling networks in the rheumatoid arthritis synovial microenvironment. J. Transl. Med. 2026, 24, 874. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Yue, S.R.; Fan, J.Y.; Xie, D.L.; Cao, C.H.; Wang, Z.Q.; Huang, J.; Qiu, F.; Yang, X.; He, D.Y.; Lu, A.P.; et al. Unveiling the Therapeutic Potential: Targeting Fibroblast-like Synoviocytes in Rheumatoid Arthritis. Expert Rev. Mol. Med. 2025, 27, e18. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Xu, L.X.; Chang, C.; Zhou, X.P.; Jiang, P.; Lin, Z.M.; Lv, X.L.; Jin, Y.H.; He, S.J.; He, D.Y. (5R)-5-hydroxytriptolide inhibits proliferation, invasion, and migration of rheumatoid arthritis fibroblast-like synoviocytes through MIF-CD74 signaling pathway. Int. Immunopharmacol. 2026, 171, 116154. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Jeon, H.M.; Noh, H.S.; Jeon, M.G.; Park, J.H.; Lee, Y.S.; Seo, G.; Cheon, Y.H.; Kim, M.; Han, M.K.; Park, J.Y.; et al. The HRAS-binding C2 domain of PLCη2 suppresses tumor-like synoviocytes and experimental arthritis in rheumatoid arthritis. Exp. Mol. Med. 2025, 57, 335–348. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Zhong, H.; He, X.; Cao, W.Q.; Ban, Z.N. BUB1 modulates inflammatory and destructive phenotypes in rheumatoid arthritis synovial fibroblasts via the p38/ATF-2 signaling pathway. Eur. J. Med. Res. 2026, 31, 944. [Google Scholar] [CrossRef] [Scilit]
- Yang, F.W.; Zhao, L.J.; Xu, Q.L.; Zhao, J.H. The journey of p38 MAP kinase inhibitors: From bench to bedside in treating inflammatory diseases. Eur. J. Med. Chem. 2024, 280, 116950. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Jan, Z.N.; El Assadi, F.; Velayutham, D.; Mifsud, B.; Jithesh, P.V. Pharmacogenomics of TNF inhibitors. Front. Immunol. 2025, 16, 1521794. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Stolshek, B.S.; Wade, S.; Mutebi, A.; De, A.P.; Yeaw, J. Two-year adherence and costs for biologic therapy for rheumatoid arthritis. Am. J. Manag. Care 2018, 24, SP315–SP321. [Google Scholar] [PubMed]
- Smolen, J.S.; Landewé, R.B.M.; Bergstra, S.A.; Kerschbaumer, A.; Sepriano, A.; Aletaha, D.; Caporali, R.; Edwards, C.J.; Hyrich, K.L.; Pope, J.E.; et al. EULAR recommendations for the management of rheumatoid arthritis with synthetic and biological disease-modifying antirheumatic drugs: 2022 update. Ann. Rheum. Dis. 2023, 82, 3–18, Correction in Ann. Rheum. Dis. 2023, 82, e76. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Rolon, S.B.; Echevarria, A.N.O.; Claudio, M.B.; Esquilin, E.A.S.; Ramirez, P.J.R.; Bou, S.; Echevarria, A.O.; Claudio, M.I.B.; Rodriguez-Ramirez, P.J. Long-Term Cardiovascular, Thromboembolic, and Malignancy Safety of Janus Kinase (JAK) Inhibitors in Dermatologic and Rheumatologic Populations: A Narrative Comparative Review. Cureus 2026, 18, e116911. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Mongin, D.; Aymon, R.; Choquette, D.; Coupal, L.; Codreanu, C.; Iannone, F.; Provan, S.; Kvien, T.K.; Fritsch-Stork, R.; Nordström, D. Evolution of Janus Kinase inhibitors (JAKi) prescriptions since 2015 in an international collaboration of rheumatoid arthritis registers (the ‘JAK-pot’study): Effect of regulatory warnings. In Seminars in Arthritis and Rheumatism; Elsevier: Amsterdam, The Netherlands, 2026; p. 152946. [Google Scholar]
- Bartok, B.; Firestein, G.S. Fibroblast-like synoviocytes: Key effector cells in rheumatoid arthritis. Immunol. Rev. 2010, 233, 233–255. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Chae, H.Y.; Yi, K.; Lim, S.G.; Park, J.Y.; Hyung, H.; Kim, S.Y.; Kim, W.; Lee, S.I.; Choi, D.K.; Kim, M.O.; et al. LIN28A-let-7b axis drives the aggressive and proinflammatory phenotype of rheumatoid arthritis fibroblast-like synoviocytes. Arthritis Res. Ther. 2026, 28, 113. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Li, Y.; Liu, J.; Sun, Y.; Hu, Y.D.; Cong, C.Z.; Chen, Y.M.; Fang, Y.Y. Targeting p38 MAPK signaling pathway and neutrophil extracellular traps: An important anti-inflammatory mechanism of Huangqin Qingre Chubi Capsule in rheumatoid arthritis. Int. Immunopharmacol. 2025, 148, 114112. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Mbalaviele, G.; Anderson, G.; Jones, A.; De Ciechi, P.; Settle, S.; Mnich, S.; Thiede, M.; Abu-Amer, Y.; Portanova, J.; Monahan, J. Inhibition of p38 mitogen-activated protein kinase prevents inflammatory bone destruction. J. Pharmacol. Exp. Ther. 2006, 317, 1044–1053. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Protsenko, L.; Liashenko, M.; Vlasenko, A.; Hryniuk, T.; Dobrovolny, O. Investigation of properties of biologically active substances and their content in cones of ukrainian hop varieties. Agric. Sci. Pract. 2018, 5, 52–63. [Google Scholar] [CrossRef] [Scilit]
- Oledzka, E. Xanthohumol-A Miracle Molecule with Biological Activities: A Review of Biodegradable Polymeric Carriers and Naturally Derived Compounds for Its Delivery. Int. J. Mol. Sci. 2024, 25, 3398. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Liu, M.; Hansen, P.; Wang, G.; Qiu, L.; Dong, J. Pharmacological Profile of Xanthohumol, a Prenylated Flavonoid from Hops (Humulus lupulus). Molecules 2015, 20, 754–779. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Azuero, M.; Wenceslau, C.F.; Tan, W.B. Xanthohumol: Mechanistic Actions and Emerging Evidence as a Multi-Target Natural Nutraceutical. Nutrients 2026, 18, 520. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Han, N.; Li, J.; Li, Y.Y.; Zhao, F.E.; Wang, J.; Ye, P.Y.; Zeng, Z.L. Xanthohumol ameliorates dextran sodium sulfate-induced colitis in mice by inhibiting of NF-κB signaling pathways and modulating intestinal microbiota. Eur. J. Nutr. 2024, 64, 21. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Longo, F.; Massaro, A.; Mauro, M.; Allegra, M.; Arizza, V.; Tesoriere, L.; Restivo, I. Cooperative Interaction of Hyaluronic Acid with Epigallocatechin-3-O-gallate and Xanthohumol in Targeting the NF-κB Signaling Pathway in a Cellular Model of Rheumatoid Arthritis. Antioxidants 2025, 14, 713. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Lee, Y.M.; Hsieh, K.H.; Lu, W.J.; Chou, H.C.; Chou, D.S.; Lien, L.M.; Sheu, J.R.; Lin, K.H. Xanthohumol, a Prenylated Flavonoid from Hops (Humulus lupulus), Prevents Platelet Activation in Human Platelets. Evid.-Based Complement. Altern. Med. 2012, 2012, 852362. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Wang, Q.; Chen, T.; Shuqing, Z.; Yu, L.; Chen, S.; Lu, H.; Zhu, H.; Min, X.; Li, X.; Liu, L. Xanthohumol relieves arthritis pain in mice by suppressing mitochondrial-mediated inflammation. Mol. Pain 2023, 19, 17448069231204051. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Suzuki, M.; Tetsuka, T.; Yoshida, S.; Watanabe, N.; Kobayashi, M.; Matsui, N.; Okamoto, T. The role of p38 mitogen-activated protein kinase in IL-6 and IL-8 production from the TNF-α- or IL-1β-stimulated rheumatoid synovial fibroblasts. FEBS Lett. 2000, 465, 23–27. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Bao, X.; Ma, L.; He, C. MicroRNA-23a-5p regulates cell proliferation, migration and inflammation of TNF-α-stimulated human fibroblast-like MH7A synoviocytes by targeting TLR4 in rheumatoid arthritis. Exp. Ther. Med. 2021, 21, 479. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Wei, S.S.; Sun, T.L.; Du, J.; Zhang, B.K.; Xiang, D.X.; Li, W.Q. Xanthohumol, a prenylated flavonoid from Hops, exerts anticancer effects against gastric cancer in vitro. Oncol. Rep. 2018, 40, 3213–3222. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Slawinska-Brych, A.; Król, S.K.; Dmoszynska-Graniczka, M.; Zdzisinska, B.; Stepulak, A.; Gagos, M. Xanthohumol inhibits cell cycle progression and proliferation of larynx cancer cells in vitro. Chem.-Biol. Interact. 2015, 240, 110–118. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Kong, H.R.; Qiao, Y.S.; Qi, D.H.; Zhao, S.X.; Cao, Z.M.; Feng, J.H.; Li, Y.T.; Liu, Y.K.; Liu, T. circ_0002970 promotes fibroblast-like synoviocytes invasion and the inflammatory response through Hippo/YAP signaling to induce CTGF/CCN1 expression in rheumatoid arthritis. Arthritis Res. Ther. 2025, 27, 97. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Zhu, S.; Dang, J.; Shi, Y.; Feng, X.; Hu, Y.; Lin, L.; Huang, J. Sonic hedgehog promotes synovial inflammation and articular damage through p38 mitogen-activated protein kinase signaling in experimental arthritis. J. Autoimmun. 2022, 132, 102902. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Kumar, S.; Boehm, J.; Lee, J.C. p38 MAP kinases: Key signalling molecules as therapeutic targets for inflammatory diseases. Nat. Rev. Drug Discov. 2003, 2, 717–726. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Liu, X.J.; Song, M.Q.; Wang, P.L.; Zhao, R.; Chen, H.Y.; Zhang, M.; Shi, Y.Y.; Liu, K.D.; Liu, F.F.; Yang, R.; et al. Targeted therapy of the AKT kinase inhibits esophageal squamous cell carcinoma growth in vitro and in vivo. Int. J. Cancer 2019, 145, 1007–1019. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Yang, Y.; Shen, Y.; Liu, H.-X.; Yao, X.-J. Molecular Dynamics Simulation and Free Energy Calculation Studies of the Binding Mechanism of Allosteric Inhibitors with p38α MAP Kinase. J. Chem. Inf. Model. 2011, 51, 3235–3246. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Xie, R.; Chen, Z.P.; Deng, S.F.; Jiang, X.F.; Feng, Y.; Zhao, W. Roles of immune cell metabolism in rheumatoid arthritis. Front. Immunol. 2026, 17, 1763130. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Gao, Z.C.; He, C.X.; Zhang, N.; Zhou, X.; Zhang, Y.; Fu, J.; Li, P.Q.; Chu, H.Q.; Zhao, D.; Li, P.; et al. Research progresses on natural product monomers in the treatment of rheumatoid arthritis. Curr. Pharm. Anal. 2026, 22, 544–555. [Google Scholar] [CrossRef] [Scilit]
- Altalbawy, F.M.A.; Zwamel, A.H.; Bishoyi, A.K.; Ballal, S.; Kalia, R.; Joshi, K.K.; Ray, S.; Pramanik, A.; Rizaev, J.; Muzammil, K. Unlocking potentials of a prenylated flavonoid: Xanthohumol’s role in combating inflammatory conditions and Rhematic diseases. Cytokine 2026, 198, 157092. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Korb, A.; Tohidast-Akrad, M.; Cetin, E.; Axmann, R.; Smolen, J.; Schett, G. Differential tissue expression and activation of p38 MAPK α, β, γ, and δ isoforms in rheumatoid arthritis. Arthritis Rheum. 2006, 54, 2745–2756. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Min, Y.-D.; Choi, C.-H.; Bark, H.; Son, H.-Y.; Park, H.-H.; Lee, S.; Park, J.-W.; Park, E.-K.; Shin, H.-I.; Kim, S.-H. Quercetin inhibits expression of inflammatory cytokines through attenuation of NF-κB and p38 MAPK in HMC-1 human mast cell line. Inflamm. Res. 2007, 56, 210–215. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Zhao, X.; Jiang, K.; Liang, B.; Huang, X. Anticancer effect of xanthohumol induces growth inhibition and apoptosis of human liver cancer through NF-κB/p53-apoptosis signaling pathway. Oncol. Rep. 2016, 35, 669–675. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Kim, R.H.; Choi, S.U.; Song, Y.W. Epigenetic suppression of synovial inflammation and osteoclast differentiation in rheumatoid arthritis by I-BET762. Sci. Rep. 2026, 16, 6042. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Hu, Z.; Li, Y.; Zhang, L.; Jiang, Y.; Long, C.; Yang, Q.; Yang, M. Metabolic changes in fibroblast-like synoviocytes in rheumatoid arthritis: State of the art review. Front. Immunol. 2024, 15, 1250884. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Cohen, S.B.; Cheng, T.; Chindalore, V.; Damjanov, N.; Burgos-Vargas, R.; DeLora, P.; Zimany, K.; Travers, H.; Caulfield, J.P. Evaluation of the efficacy and safety of pamapimod, a p38 MAP kinase inhibitor, in a double-blind, methotrexate-controlled study of patients with active rheumatoid arthritis. Arthritis Rheumatol. 2010, 60, 335–344. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Damjanov, N.; Kauffman, R.S.; Spencer-Green, G.T. Efficacy, pharmacodynamics, and safety of VX-702, a novel p38 MAPK inhibitor, in rheumatoid arthritis: Results of two randomized, double-blind, placebo-controlled clinical studies. Arthritis Rheumatol. 2009, 60, 1232–1241. [Google Scholar] [CrossRef] [Scilit]
- Genovese, M.C.; Cohen, S.B.; Wofsy, D.; Weinblatt, M.E.; Firestein, G.S.; Brahn, E.; Strand, V.; Baker, D.G.; Tong, S.E. A 24-week, randomized, double-blind, placebo-controlled, parallel group study of the efficacy of oral SCIO-469, a p38 mitogen-activated protein kinase inhibitor, in patients with active rheumatoid arthritis. J. Rheumatol. 2011, 38, 846–854. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Hammaker, D.; Firestein, G. “Go upstream, young man”: Lessons learned from the p38 saga. Ann. Rheum. Dis. 2010, 69, i77–i82. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Mahapatra, D.K.; Bharti, S.K.; Asati, V. Chalcone derivatives: Anti-inflammatory potential and molecular targets perspectives. Curr. Top. Med. Chem. 2017, 17, 3146–3169. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Grada, A.; Otero-Vinas, M.; Prieto-Castrillo, F.; Obagi, Z.; Falanga, V. Research Techniques Made Simple: Analysis of Collective Cell Migration Using the Wound Healing Assay. J. Investig. Dermatol. 2017, 137, e11–e16. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Shu, X.; Shi, Y.; Huang, Y.; Yu, D.; Sun, B. Transcription tuned by S-nitrosylation underlies a mechanism for Staphylococcus aureus to circumvent vancomycin killing. Nat. Commun. 2023, 14, 2318. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Shan, B.; Zhao, S.; Zhang, Z.; Cao, H.; Jiang, Y.; Du, Y.; Zhang, Q.; Zhao, P.; Li, J. Zaoshi Huatan Xiefei formula improved severe pneumonia by regulating MAPK/NF-κB/VCAM1 signaling pathway mediated leukocyte transendothelial migration. Phytomedicine 2026, 161, 158742. [Google Scholar] [CrossRef] [Scilit] [PubMed]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2026 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license.






