Flavonoids attenuate inflammation of HGF and HBMSC while modulating the osteogenic differentiation based on microfluidic chip Authors: Sa Du, Zhongyu Wang, Huilin Zhu, Zhihui Tang, Qing Li Published: 2024 ABSTRACT BackgroundWhen inflammation occurs in periodontal tissues, a dynamic cellular crosstalk interacts between gingival fibroblasts and bone marrow mesenchymal stem cells (BMSCs), which plays a crucial role in the biological behaviour and differentiation of the cells. Recently, flavonoids are increasingly recognized for their therapeutic potential in modulating inflammation and osteogenic differentiation. Owing to their varied molecular structures and mechanisms, there are more needs that flavonoid compounds should be identified by extensive screening. However, current drug research mostly relies on static, single-type cell cultures. In this study, an innovative bionic microfluidic chip system tailored for both soft and hard tissues was developed to screen for flavonoids suitable for treating periodontitis.MethodsThis study developed a microfluidic system that bionically simulates the soft and hard structures of periodontal tissues. Live/dead staining, reactive oxygen species (ROS) staining, and RT-qPCR analysis were employed. These techniques evaluated the effects of flavonoid compounds on the levels of inflammatory factors and ROS contents in HGF and HBMSC under LPS stimulation. Additionally, the impact of these compounds on osteogenic induction in HBMSC and the exploration of the underlying mechanisms were assessed.ResultsThe microfluidic chip used in this study features dual chambers separated by a porous membrane, allowing cellular signal communication via bioactive factors secreted by cells in both layers under perfusion. The inflammatory response within the chip under LPS stimulation was lower compared to individual static cultures of HGF and HBMSC. The selected flavonoids-myricetin, catechin, and quercetin-significantly reduced cellular inflammation, decreased ROS levels, and enhanced osteogenic differentiation of BMSCs. Additionally, fisetin, silybin, and icariside II also demonstrated favorable outcomes in reducing inflammation, lowering ROS levels, and promoting osteogenic differentiation through the Wnt/β-catenin pathway.ConclusionsThe bionic microfluidic chip system provides enhanced capabilities for drug screening and evaluation, delivering a more precise assessment of drug efficacy and safety compared to traditional in vitro methods. This study demonstrates the efficacy of flavonoids in influencing osteogenic processes in BMSCs primarily through the Wnt/β-catenin pathway. These results uncover the potential of flavonoids as therapeutic medicine for treating periodontitis, meriting further research and development.Graphical Abstract Supplementary InformationThe online version contains supplementary material available at 10.1186/s12967-024-05808-1. Background Periodontitis is a chronic inflammatory disease that is mainly characterized by inflammation, bleeding of the gingiva and resorption of the alveolar bone [1]. In the process of periodontitis, secretory factors and other signals participate in the crosstalk between gingival soft tissues and periodontal bone tissues. Human gingival fibroblasts (HGFs), the predominant periodontal cells in gingival connective tissue, produce matrix metalloproteinases that degrade collagen and fibronectin in the extracellular matrix, exacerbating the degeneration of gingival tissue. Concurrently, various inflammatory factors, chemokines and inflammatory mediators, such as interleukin 6 (IL6), interleukin 8 (IL8) and prostaglandin E2 (PGE2), are released and inhibit the osteogenic function of human bone marrow stem cells (HBMSCs) [2–4]. HBMSCs secrete various cytokines and active molecules, such as interleukin 10 (IL10), vascular endothelial growth factor (VEGF) and bone morphogenetic protein 2 (BMP2), to reduce inflammation in gingival tissues [5, 6]. To date, traditional in vitro culture methods for HGFs and HBMSCs mostly involve the use of culture dishes for the static culture of individual types of cells. However, traditional static culture methods cannot mimic the actual in vivo microenvironment [7, 8]. Individual cell culture cannot emulate the crosstalk between HGFs and HBMSCs. Therefore, there is an urgent need for the development of a dynamic, hybrid culture system as a whole soft-hard tissue bionic unit that can be used for research on inflammation control and tissue regeneration. The organ-on-chip method has shown promising potential for investigating pathophysiological processes [9]. In this system, a small amount of fluid can be manipulated within a microsized channel that continuously supplies fresh media. Furthermore, this system simulates the multilayered structure of physiological tissues [10]. Standard treatment for periodontitis consists of plaque removal and control of plaque accumulation. However, it is challenging to achieve optimal outcomes for inflammation control, periodontal tissue regeneration, and immune modulation [11]. Drug-assisted therapy can supplement basic periodontal therapy and effectively prolong therapeutic efficacy [12]. Herbal medicines, as natural products, have been used for the prevention and treatment of diseases since ancient times. Various natural flavonoid compounds, such as myricetin, catechin and quercetin, are found in many medicinal plants and are abundant. They have been shown to have antioxidant, anti-inflammatory, and immunomodulating effects and have been widely explored as periodontal therapeutic medicine [13]. Current research into the pharmacological effects of these drugs primarily uses cell culture methods in vitro and has yet to utilize microfluidic chips as a novel therapeutic approach. Different flavonoid substances may have different biological activities and mechanisms for disease treatment. The identification of novel flavonoid compounds may provide more drug options, more personalized treatment regimens, and various treatment efficacy. Fisetin, silybin and icariside II have shown various pharmacological effects, including good antiaging, antihepatitis and antitumour effects. Nevertheless, there is yet lack of exploration of potential therapeutic options for managing inflammation and osteogenesis. Thus, it is really urgent to develop rapid and efficient detection devices for screening effective drugs and their concentrations [14]. In this study, we aimed to construct a soft and hard tissue bionic chip model that would reproduce the soft and hard tissue crosstalk. Then, we explored the abilities of flavonoid compounds to alleviate inflammation in HGFs and HBMSCs and to modulate the osteogenic differentiation of mesenchymal stem cells under lipopolysaccharide (LPS) stimulation. With the aid of microfluidic chip technology, there is promising application of flavonoids in controlling periodontal inflammation and facilitating tissue regeneration in the future. Materials and methods Cell culture HBMSCs (ScienCell Research Laboratory, Carlsbad, CA, USA) and HGFs (Stomatology School and Hospital of Peking University) were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 IU/mL penicillin (P), and 100 μg/mL streptomycin (S) at 37 °C in a humidified incubator with 5% CO2. Our previous studies confirmed the presence of CD90 on the surface of HBMSCs and the positive expression of vimentin and negative expression of CD31 on the surface of HGFs [15]. In subsequent experiments, cells between 3 and 5 passages were used. For osteogenic differentiation of HBMSCs, the cells were induced in osteogenic induction medium according to the instructions of the OriCell osteogenic differentiation kit (OriCell, China). After the cells adhered, Pg. LPS (5 μg/mL) Ultrapure (InvivoGen, San Diego, CA, USA) was added to create an inflammatory environment for 24 h. The TLR4 inhibitor TAK-242 (5 μM) and the Wnt/β-catenin signalling pathway inhibitor ICG-001 (10 μM) were purchased from Selleck Chemicals (Houston, TX, USA). Microfluidic assays To construct a model of the hybrid structure of oral soft and hard tissues, BEOnChip cell culture chips (BEOnChip, Spain) were used in different settings. To prevent the formation of air bubbles in the pipeline and chamber during perfusion, the perfusate was connected to a reservoir bottle and catheter in advance and was placed in the incubator overnight before cell seeding. HBMSCs were suspended at a concentration of 4 × 105 cells/mL in the culture medium and seeded onto the lower layer of the chip. After the HBMSCs had adhered to the lower layer of the chip, HGFs were seeded in the upper chamber at a concentration of 4 × 105 cells/mL. Cells within the chip were cultured overnight at 37 °C in a humidified CO2 incubator for attachment. The composite chip was connected to a perfusion system comprising microfluidic pumps (PUSH–PULL, Fluigent, France), flow sensors (FLOW UNIT, Fluigent, France), reservoir bottles, and collection tubes. The flow rate of the microfluidic system was determined and controlled in real time by the experimental software Oxygen (version 2.3.0.0) on a computer. The attached cells were further cultured at a constant flow rate (120 μL/h, Fluigent, France). Pretreatment with flavonoid compounds Flavonoid compounds, including myricetin (Myr; PubChem CID: 5281672), catechin (Cat; PubChem CID: 1203), quercetin (Que; PubChem CID: 5280343), fisetin (Fis; PubChem CID: 5281614), silybin (Sil; PubChem CID: 31553), and icariside II (Ica; PubChem CID: 5488822), were dissolved in DMSO and stored at − 20 °C. When flavonoids were added to the experimental cells, the final concentration of DMSO was < 0.1% (v/v). The cells were treated with flavonoids for 2 h before stimulation with 5 μg/mL LPS. Cells in the control group were treated with DMEM supplemented with an equivalent amount of DMSO. Cell viability assay A fluorescent live/dead cell survival assay kit (KGAF001, KeyGEN Bio TECH, China) was used to evaluate the survival of cells within the two layers of the chip. For the preparation of the staining buffer, propidium iodide (PI) and calcein AM were dissolved in 10 mL of PBS at concentrations of 8 μM and 2 μM, respectively. After incubation at 37 °C for 30 min, the chip was washed three times with PBS. Under a confocal laser scanning microscope (Zeiss) or fluorescence microscope (Olympus, Japan), live cells were stained green, and dead cells were stained red. Determination of intracellular reactive oxygen species (ROS) levels Intracellular ROS accumulation in cells was measured in vitro with 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA). The nonfluorescent DCFH-DA can penetrate the cell membrane and is hydrolysed by intracellular esterases to produce DCFH, which cannot penetrate the cell membrane. Accumulated intracellular ROS oxidize nonfluorescent DCFH to produce fluorescent 2′,7′-dichlorofluorescein (DCF). The cells were incubated with 10 μM DCFH-DA (Beyotime, Shanghai, China) at 37 °C for 20 min in the dark and then rinsed with PBS. Images were taken with a fluorescence microscope (Olympus, Japan). Real-time quantitative polymerase chain reaction (RT‒qPCR) Total RNA was obtained using the RaPure total RNA micro kit (Magen, China) following the manufacturer's protocol. Subsequently, cDNA was obtained by reverse transcription using the PrimeScript™ RT master mix (TaKaRa, Japan). The cDNA, primers, and TB Green® Premix (TaKaRa, Japan) were mixed together. The experiments were carried out in an Mx3000P PCR system (Agilent, USA) under the following thermocycler conditions: 95 °C for 30 s, followed by 40 cycles of 15 s at 95 °C and 34 s at 60 °C. The expression of target genes was normalized to that of GAPDH using the comparative 2-ΔΔCT method. The primer sequences are listed in Table 1.Table 1Sequences of the primers used for RT‒qPCRGene namePrimer sequence (5' to 3')IL6Forward: ATGAGGAGACTTGCCTGGTGAAReverse: CTCTGGCTTGTTCCTCACTACTCTCIL8Forward: ACTGAGAGTGATTGAGAGTGGACReverse: AACCCTCTGCACCCAGTTTTCRUNX2Forward: TTCAACGATCTGAGATTTGTGGGReverse: GGATGAGGAATGCGCCCTAALPForward: CTCCTCGGAAGACACTCTGACCReverse: CTGCGCCTGGTAGTTGTTGTGOSXForward: ATGGGCTCCTTTCACCTGReverse: GGGAAAAGGGAGGGTAATCOPNForward: CTCCATTGACTCGAACGACTCReverse: CAGGTCTGCGAAACTTCTTAGATWNTForward: CTCGCTGGCTACCCAATTTGReverse: AGGCTGTCATCTATGGTGGTGCTNNB1Forward: CATCTACACAGTTTGATGCTGCTReverse: GCAGTTTTGTCAGTTCAGGGAGAPDHForward: GTCTCCTCTGACTTCAACAGCGReverse: ACCACCCTGTTGCTGTAGCCAA Table 1Sequences of the primers used for RT‒qPCRGene namePrimer sequence (5' to 3')IL6Forward: ATGAGGAGACTTGCCTGGTGAAReverse: CTCTGGCTTGTTCCTCACTACTCTCIL8Forward: ACTGAGAGTGATTGAGAGTGGACReverse: AACCCTCTGCACCCAGTTTTCRUNX2Forward: TTCAACGATCTGAGATTTGTGGGReverse: GGATGAGGAATGCGCCCTAALPForward: CTCCTCGGAAGACACTCTGACCReverse: CTGCGCCTGGTAGTTGTTGTGOSXForward: ATGGGCTCCTTTCACCTGReverse: GGGAAAAGGGAGGGTAATCOPNForward: CTCCATTGACTCGAACGACTCReverse: CAGGTCTGCGAAACTTCTTAGATWNTForward: CTCGCTGGCTACCCAATTTGReverse: AGGCTGTCATCTATGGTGGTGCTNNB1Forward: CATCTACACAGTTTGATGCTGCTReverse: GCAGTTTTGTCAGTTCAGGGAGAPDHForward: GTCTCCTCTGACTTCAACAGCGReverse: ACCACCCTGTTGCTGTAGCCAA Sequences of the primers used for RT‒qPCR Immunofluorescence staining After cell stimulation, the chip was separated from the perfusion system. The cells were rinsed with PBS and then fixed with 4% paraformaldehyde for 15 min at room temperature. Then, the cells were treated with 0.2% Triton (Sigma‒Aldrich, St. Louis, MO, USA) and blocked with 10% goat serum (ZSGB-BIO, Beijing, China). Primary antibodies against human β-catenin (Proteintech) and RUNX2 (Abcam) were diluted to 1:500 and incubated with the cells at 4 °C overnight. Subsequently, the cells were washed 3 times with PBS. The cells were then incubated with secondary antibodies, including multi-rAb CoraLite® Plus 555 (Proteintech) and multi-rAb CoraLite® Plus 488 (Proteintech), at room temperature for 1 h. The nuclei were stained with DAPI. Images were acquired using a fluorescence microscope (Olympus, Japan). Statistical analysis Data analysis was performed using SPSS (version 26.0), and all charts were generated using GraphPad Prism 9.0. One-way analysis of variance (ANOVA) was used to determine statistical significance of the results. The results are presented as the mean ± standard deviation. Significance at three levels is indicated as follows: *p < 0.05, **p < 0.01, and ***p < 0.001. Results Characterization of the soft-hard tissue bionic hybrid chip The soft and hard tissue bionic hybrid chip developed in this study effectively mimicked the physiological environment, facilitating dynamic cellular interactions. A schematic of the microfluidic chip used for screening flavonoid drugs is shown in Fig. 1a–c. The upper and lower chambers were separated by a porous membrane with a pore diameter of 1 µm. The dimensions of the chip are shown in Fig. 1c. The soft and hard tissue bionic hybrid chip perfusion system established in this study consisted of microfluidic air pumps, microfluidic chips, liquid flow meters, catheters, reservoir bottles, and a computer control program (Fig. 1d). The staining of HGFs and HBMSCs with calcein AM and PI revealed live and dead cells, respectively. After 24 h of incubation under flow conditions with a medium containing 5 μg/mL LPS, the chip had a high number of live cells that were uniformly distributed in both the upper and lower layers, as shown by confocal microscopy (Fig. S1). These results indicated that both HGFs and HBMSCs maintained high survival rates, even in the presence of LPS, confirming the suitability of the chip for cellular experiments.Fig. 1Configuration of the microfluidic chip. a, b Multilayer structure of the soft-hard tissue bionic microfluidic chip. The chip included five layers: a top cover, an upper chamber, a porous membrane, a lower chamber and a bottom cover. c A schematic cross-sectional view of the chip, showing the upper (HGF; red) and lower (HBMSC; blue) cell culture microchannels separated by a porous membrane sandwiched between them. Pores in the membrane allow for signal exchange between the two layers. d Schematic representation of the soft-hard tissue bionic hybrid chip perfusion system. It consists of microfluidic air pumps, microfluidic chips, liquid flow meters, catheters, reservoir bottles, and a computer control program. e, f Inflammatory responses to LPS in microfluidic chips and dishes in HGFs and HBMSCs. The data are expressed as the mean ± SD. ***p < 0.001, **p < 0.01, *p < 0.05 Configuration of the microfluidic chip. a, b Multilayer structure of the soft-hard tissue bionic microfluidic chip. The chip included five layers: a top cover, an upper chamber, a porous membrane, a lower chamber and a bottom cover. c A schematic cross-sectional view of the chip, showing the upper (HGF; red) and lower (HBMSC; blue) cell culture microchannels separated by a porous membrane sandwiched between them. Pores in the membrane allow for signal exchange between the two layers. d Schematic representation of the soft-hard tissue bionic hybrid chip perfusion system. It consists of microfluidic air pumps, microfluidic chips, liquid flow meters, catheters, reservoir bottles, and a computer control program. e, f Inflammatory responses to LPS in microfluidic chips and dishes in HGFs and HBMSCs. The data are expressed as the mean ± SD. ***p < 0.001, **p < 0.01, *p < 0.05 Modulation of the inflammatory responses to LPS under microfluidic chip coculture conditions The addition of 5 μg/mL LPS to HGFs cultured alone in culture dishes resulted in significant increases in the IL6 and IL8 levels. Compared with those in the control group, the levels of the inflammatory factors IL6 and IL8 in HGFs were elevated more than tenfold. In contrast, the IL6 and IL8 levels in HGFs in the microfluidic chips were elevated only approximately threefold after stimulation with LPS (Fig. 1e). When HBMSCs in a culture dish and in the microfluidic chips were stimulated under the same conditions, we observed greater expression of IL6 and IL8 in HBMSCs cultured alone in the dish than in those cocultured in the chip (Fig. 1f). Compared with individual cultivation in traditional culture dishes, chip coculture attenuated the cellular inflammatory responses to LPS. Increases in the IL6 and IL8 expression were notably less pronounced in the chip-based coculture system, reinforcing the benefits and necessity of the bionic microfluidic chip system. Anti-inflammatory effects of myricetin, catechin and quercetin compared to those of TAK-242 To verify whether the above three drugs could reduce the levels of inflammatory factors in HGFs and HBMSCs within this composite chip, IL6 and IL8 mRNA expression was determined using RT‒qPCR. The effects of TAK-242, a specific inhibitor of TLR4, were compared with those of the above three compounds. The results showed that myricetin, catechin and quercetin were all effective at reducing intracellular IL6 and IL8 levels (Fig. 2a, b, d, e). In HBMSCs, catechin reduced the levels of the inflammatory factors to half the original levels (Fig. 2d, e). Among these three drugs, catechin had the strongest anti-inflammatory effect. These three flavonoid compounds demonstrated effects similar to those of TAK-242. It was also confirmed that this hybrid chip could be used as a generalized screening system to explore the effects of drugs.Fig. 2Levels of inflammatory factors and ROS after treatment with myricetin, catechin and quercetin. a IL6 mRNA expression levels in HGFs. b IL8 mRNA expression levels in HGFs. c Immunofluorescence staining of ROS in HGFs on the chip. Nuclear staining (Hoechst) and DCFH-DA staining are shown (scale bar, 100 μm). d IL6 mRNA expression levels in HBMSCs. e IL8 mRNA expression levels in HBMSCs. f Fluorescence staining of ROS in HBMSCs on the chip. Nuclear staining (Hoechst) and DCFH-DA staining are shown (scale bar, 100 μm). g Effects of flavonoid compounds on the osteogenic differentiation potential of HBMSCs. RUNX2, ALP, OSX and OPN mRNA expression in HBMSCs. The data are expressed as the mean ± SD. ***p < 0.001, **p < 0.01, *p < 0.05 Levels of inflammatory factors and ROS after treatment with myricetin, catechin and quercetin. a IL6 mRNA expression levels in HGFs. b IL8 mRNA expression levels in HGFs. c Immunofluorescence staining of ROS in HGFs on the chip. Nuclear staining (Hoechst) and DCFH-DA staining are shown (scale bar, 100 μm). d IL6 mRNA expression levels in HBMSCs. e IL8 mRNA expression levels in HBMSCs. f Fluorescence staining of ROS in HBMSCs on the chip. Nuclear staining (Hoechst) and DCFH-DA staining are shown (scale bar, 100 μm). g Effects of flavonoid compounds on the osteogenic differentiation potential of HBMSCs. RUNX2, ALP, OSX and OPN mRNA expression in HBMSCs. The data are expressed as the mean ± SD. ***p < 0.001, **p < 0.01, *p < 0.05 ROS-scavenging abilities of myricetin, catechin and quercetin According to previous literature reports, 10 μM myricetin, catechin or quercetin maintained good cytocompatibility of HGFs and HBMSCs. Chips with a perfusion solution containing 5 μg/mL LPS were used as the control group. Chips pretreated with 10 μM myricetin, catechin or quercetin were used as the experimental groups. Without drug pretreatment, LPS increased the ROS levels in HGFs and HBMSCs, as indicated by the green fluorescence in both the nucleus and cytoplasm (Fig. 2c, f). A significant reduction in the intracellular ROS levels was observed after pretreatment with myricetin, catechin or quercetin (Figs. S2, S3). Among these three drugs, catechin had the most significant effect, followed by myricetin. Effects of the flavonoid compounds on the osteogenic differentiation potential of HBMSCs To determine how the above three compounds affect the osteogenic differentiation of HBMSCs, we assessed the expression of the osteogenesis-related factors RUNX2, ALP, OSX, and OPN by qPCR. Compared with that in the control group, the expression of osteogenesis-related genes in HBMSCs was significantly upregulated after 3 days of perfusion of osteogenic differentiation medium supplemented with myricetin, catechin, or quercetin. Among the three tested flavonoids, catechin had the most pronounced osteogenesis induction effect (Fig. 2g). Drug concentration screening of fisetin, silybin and icariside II In this study, three flavonoid compounds including fisetin, silybin and icariside II that have not been previously examined for use in periodontal therapy, were selected for drug screening in hybrid chips. The results of live-dead staining showed no significant differences in the survival rates between HGFs and HBMSCs treated with the above drugs at a concentration of 10 μM or 20 μM and those treated with the control media (Fig. 3a–c). The survival rates reached more than 90% (Fig. 3d–f). However, at drug concentrations of 50 μM and 100 μM, the numbers of live cells notably decreased. After treatment with 100 μM fisetin and silybin, the densities of live HGFs and HBMSCs decreased, with concomitant increases in the numbers of dead cells. Only a few green fluorescent live cells were visible under the microscope, and the cells were not attached. When the concentration of icariside II was increased to 100 μM, only a few cells attached to the chip exhibited red fluorescence (Fig. 3c). The likely reason is that most of the dead cells floated and detached from the chip to be carried away by the perfusion system. Based on the above live-dead staining results, 10 μM fisetin, silybin and icariside II were selected for subsequent analyses.Fig. 3Drug screening tests of fisetin, silybin and icariside II. a–c Fluorescence diagram of cell activities after treatment with different concentrations of fisetin, silybin and icariside II. d–f Cell viability analysis in fisetin (d), silybin (e) and icariside II (f) groups. The data are expressed as the mean ± SD. ***p < 0.001, **p < 0.01, *p < 0.05 Drug screening tests of fisetin, silybin and icariside II. a–c Fluorescence diagram of cell activities after treatment with different concentrations of fisetin, silybin and icariside II. d–f Cell viability analysis in fisetin (d), silybin (e) and icariside II (f) groups. The data are expressed as the mean ± SD. ***p < 0.001, **p < 0.01, *p < 0.05 Reduction in the levels of inflammatory factors in coculture chips by fisetin, silybin and icariside II A perfusion medium containing 5 μg/mL LPS was used for the control group. For the experimental groups, 10 μM fisetin, silybin or icariside II was added to LPS. After 1 day of stimulation, the cells within the chip were collected for qPCR analysis. Compared with those in the LPS-treated group, the IL6 and IL8 levels in the groups pretreated with fisetin, silybin, or icariside II were significantly lower in both HGFs (Fig. 4a, b) and HBMSCs (Fig. 4d, e). Interestingly, fisetin demonstrated the greatest ability to inhibit inflammation.Fig. 4Levels of inflammatory factors and ROS after treatment with fisetin, silybin and icariside II. IL6 (a) and IL8 (b) levels in HGFs. c Immunofluorescence staining of ROS in HGFs on the chip (scale bar, 100 μm). IL6 (d) and IL8 (e) levels in HBMSCs. f Fluorescence staining of ROS in HBMSCs on the chip (scale bar, 100 μm). g Effects of flavonoid compounds on the osteogenic differentiation potential of HBMSCs. RUNX2, ALP, OSX and OPN mRNA expression in HBMSCs. The data are expressed as the mean ± SD. ***p < 0.001, **p < 0.01, *p < 0.05 Levels of inflammatory factors and ROS after treatment with fisetin, silybin and icariside II. IL6 (a) and IL8 (b) levels in HGFs. c Immunofluorescence staining of ROS in HGFs on the chip (scale bar, 100 μm). IL6 (d) and IL8 (e) levels in HBMSCs. f Fluorescence staining of ROS in HBMSCs on the chip (scale bar, 100 μm). g Effects of flavonoid compounds on the osteogenic differentiation potential of HBMSCs. RUNX2, ALP, OSX and OPN mRNA expression in HBMSCs. The data are expressed as the mean ± SD. ***p < 0.001, **p < 0.01, *p < 0.05 ROS-scavenging abilities of fisetin, silybin and icariside II To determine the effects of fisetin, silybin and icariside II on intracellular ROS levels, we used the DCFH-DA probe to assess ROS levels in HGFs and HBMSCs. Similar to the above results, strong green fluorescence was observed in the nuclei and cytoplasm of both HGFs and HBMSCs when the perfusion medium contained 5 μg/mL LPS. After treatment with fisetin, silybin or icariside II, the fluorescence intensity (indicating the intracellular ROS levels) was significantly reduced (Fig. 4c, f). The fluorescence intensity was the lowest in the fisetin group. Effects of the flavonoid compounds on the osteogenic differentiation of HBMSCs To assess the effects of the three drugs on the osteogenic differentiation of HBMSCs, we treated the cells with 10 μM fisetin, silybin or icariside II in osteogenic differentiation medium. PCR results showed that the mRNA levels of RUNX2, ALP, OSX and OPN were significantly elevated in the cells after 3 days of induction (Fig. 4g). The increase induced by fisetin was the most significant. Fisetin modulates the osteogenic differentiation of HBMSCs through the Wnt/β-catenin pathway Compared with that in the LPS group, the fluorescence intensity of β-catenin was significantly greater in HBMSCs cultured in osteogenic differentiation medium supplemented with fisetin (Fig. 5a). After the addition of ICG-001, a specific inhibitor of Wnt, to the osteogenic induction medium, the immunofluorescence of the β-catenin protein significantly decreased, and almost no fluorescence was observed in the nuclei. When both fisetin and ICG-001 were added to the osteogenic medium, the expression of β-catenin significantly increased. Moreover, compared with that in the ICG-001 group, the distribution of β-catenin in the nucleus increased. qPCR further confirmed that Wnt and β-catenin mRNA expression was decreased in the cells after the addition of ICG-001 (Fig. 5b, c). The role of the Wnt/β-catenin pathway in the osteogenic differentiation of HBMSCs was further explored. The expression of the osteogenesis-related factors RUNX2, ALP, and OPN decreased after 3 days of osteogenic induction in the medium supplemented with ICG-001 but rebounded after incubation in the medium supplemented with both fisetin and ICG-001 (Fig. 5d). The levels of the osteogenesis-related factors RUNX2 (Fig. 5e), ALP (Fig. 5f), and OPN (Fig. S4) decreased in the ICG-001 group but increased after fisetin supplementation. The above results suggest that the Wnt/β-catenin pathway plays a key role in flavonoid-induced osteogenic differentiation of HBMSCs.Fig. 5Flavonoids regulate the osteogenic differentiation of HBMSCs through the Wnt-β-catenin pathway. a Changes in β-catenin expression in the presence or absence of ICG-001 and/or fisetin. b, c Changes in the expression of Wnt and β-catenin following the addition of ICG-001. d The expression of Runx2 in the presence or absence of ICG-001 and/or fisetin. e, f The relative expression levels of RUNX2 and ALP in the presence or absence of ICG-001 and/or fisetin. g Schematic representation of how flavonoids regulate osteogenic differentiation in HBMSCs through the Wnt/β-catenin pathway Flavonoids regulate the osteogenic differentiation of HBMSCs through the Wnt-β-catenin pathway. a Changes in β-catenin expression in the presence or absence of ICG-001 and/or fisetin. b, c Changes in the expression of Wnt and β-catenin following the addition of ICG-001. d The expression of Runx2 in the presence or absence of ICG-001 and/or fisetin. e, f The relative expression levels of RUNX2 and ALP in the presence or absence of ICG-001 and/or fisetin. g Schematic representation of how flavonoids regulate osteogenic differentiation in HBMSCs through the Wnt/β-catenin pathway Discussion The previous researches have demonstrated that the interactive crosstalk between HBMSCs and HGFs modulates their cellular behaviours [16]. For instance, Iwata et al. cocultured HBMSCs with HGFs for osteogenic induction and found matrix mineralization was significantly reduced in HBMSCs after coculture compared to that without coculture in the third week. This effect may be mediated through histone deacetylase regulation [17]. Insulin-like growth factor-1 (IGF-1) secreted by HGFs was shown to inhibit osteogenesis in MSCs through miRNAs and undifferentiated MSC markers [18]. However, current chip models are often limited to simulating the local microenvironments of specific single structures such as the gingiva or periodontal ligament [19, 20]. Some studies have recreated the interactions between soft and hard tissues, conducting indirect co-culture of various cell types. These studies often rely on transwell chambers and are difficult to simulate the fluid microenvironment in vivo [16, 17]. The bionic hybrid chip designed in this study consists of upper and lower cell culture chambers separated by a porous membrane. Driven by perfusion, the bioactive factors secreted by cells in the upper and lower layers mediate an intercellular signalling communication between the two layers through the porous membrane. The results showed that the inflammatory responses in fibroblasts and MSCs in the interactive culture system on the chip were more intense than those in single-type cells in a culture dish, when stimulated with LPS. This finding may result from the inhibition of caspase-3 expression by the paracrine mechanism of MSCs, which attenuates oxidative stress damage in fibroblasts [5]. The potential of plants for treating diseases has been recognized since ancient times. Natural products are ideal drug candidates for a wide range of diseases, owing to their low toxicity, low rate of drug resistance and wide range of pharmacological activities [21, 22]. Flavonoids, for instance, have strong antioxidant properties because they can convert free radicals into stable deprotonated forms. In addition, these molecules exert antitumour effects, control blood sugar, and have hepatoprotective and neuroprotective effects [23, 24]. In the field of periodontal health, studies have been conducted using drugs such as myricetin [25–27], catechin [28–30] and quercetin [31–34] to control ROS levels in periodontal tissues and inhibit the progression of inflammation to promote tissue regeneration. Nevertheless, research on more biocompatible flavonoids that could effectively modulate inflammation and osteogenic differentiation still remains in blank. In this study, a soft-hard tissue bionic hybrid chip was developed for flavonoid drug screening. In addition to the drugs mentioned above, which have been extensively investigated in the past, three flavonoid compounds (fisetin, silybin and icariside II) that have not been used in periodontal therapy were identified by screening in this study. The results of the present study showed that the flavonoid compounds exerted broad regulatory effects on cellular functions by scavenging ROS, inhibiting the expression of inflammatory factors and promoting osteogenesis, with catechin and fisetin showing the most significant effects. In addition to the compounds utilized in this study, the discovery of more flavonoid compounds in the future would be valuable. Drug development is a long and time-consuming process [35]. The effectiveness of the same drugs may vary owing to differences in the genetics, living environments and lifestyles among individuals [36]. Thus, it is essential for precise preclinical to assess the safety of drugs, developing personalized drug regimens, and predicting the safety and efficacy of multiple drug combinations [37]. However, current in vitro studies for drug screening mainly rely on traditional static cell culture in dishes, and most of these studies involve the culture of individual cell types [35]. Traditional in vitro research methods lack a dynamic environment and rarely mimic the in vivo microenvironment of cells. Moreover, information on cell‒cell interactions, such as molecular signal transduction, cannot be obtained. In addition, the removal of cellular metabolites is difficult in traditional culture [8]. Thus, cell biological functions, such as the proliferation, migration, and differentiation, may differ significantly from those under in vivo conditions. In response to these limitations, an S–H hybrid chip that can be used as a novel drug screening system was proposed for the first time in this study. Interactive cultures of fibroblasts and mesenchymal stem cells mimic the bilayer structure of hard and soft periodontal tissues. Microfluidic systems create a dynamic physiological environment. The results of this study showed that the screening system had good biocompatibility and allowed for in situ observation and fluorescence staining with low technical sensitivity. Meanwhile, cells harvested from this system could be used for molecular biology assays, such as RT‒qPCR, demonstrating the practical utility of the system. The effects of myricetin, catechin, and quercetin on human gingival fibroblasts (HGFs) and human bone marrow stem cells (HBMSCs) were comparable to or better than those of TAK-242 in controlling inflammation. These findings are consistent with the results of previous studies. The scientific validity of the microfluidic chip system was further confirmed. This system is highly practical and can bionically simulate the in vivo environment. Thus, our system can be used as an ideal model for preclinical studies of periodontal therapeutic drugs. These regulatory effects of the above-described flavonoid compounds may occur through the Wnt/β-catenin pathway. Under the influence of inflammatory cytokines, nuclear factor κB in host cells is activated, thereby inhibiting the production of Wnt/β-catenin and Runx2 [38]. The classic Wnt/β-catenin pathway is important for regulating the osteogenic differentiation of stem cells [39]. In the Wnt/β-catenin pathway, β-catenin is a key protein located in the cytoplasm. The stability of this protein is influenced by the destruction complex (DC), which consists of Axin, adenomatous polyposis coli (APC), casein kinase 1 (CK-1), and glycogen synthase kinase 3 (GSK-3). β-Catenin degradation is inhibited when exogenous stimuli cause a decrease in the level of the APC-Axin-GSK3β complex [40]. β-Catenin accumulates in the cytoplasm and can translocate to the nucleus to induce the expression of osteogenic factors, ultimately promoting the osteogenic differentiation of MSCs [41]. In the present study, fisetin stimulation significantly increased the intracellular level of β-catenin compared with that in the LPS group. After perfusion with a culture medium containing ICG-001, the total level of β-catenin in HBMSCs significantly decreased. Moreover, the immunofluorescence results showed a decrease in β-catenin levels in the cytoplasm and a near absence of β-catenin in the nucleus. In contrast, when the cells were cultured in a medium supplemented with fisetin and ICG-001, the β-catenin level in the cytoplasm increased, and β-catenin was partially translocated to the nucleus. Therefore, the Wnt/β-catenin signalling pathway may be important for the regulation of fisetin-induced osteogenic differentiation of HBMSCs. This process may also be a key regulatory mechanism by which flavonoid compounds act on MSCs. The specific mechanisms by which flavonoids enhance the osteogenic differentiation of MSCs require further in-depth exploration in the future. In conclusion, this study presents the first application of a soft-hard tissue bionic hybrid chip for screening flavonoids and other types of drugs. The flavonoid compounds tested effectively reduced the levels of the inflammatory factors IL6 and IL8 in HGFs and HBMSCs. These treatments decreased the intracellular ROS levels. In addition, the flavonoids strongly induced the osteogenic differentiation of HBMSCs. Among the three molecules that have not yet been applied in periodontal therapy, fisetin had the strongest effect, showing the greatest potential for periodontal tissue regeneration. This effect may perform function through the Wnt/β-catenin pathway. Despite the encouraging results, research on flavonoid compounds still faces several challenges. First, the molecular mechanisms by which flavonoids affect periodontal tissue regeneration are not fully understood. And further research is needed to clarify the specific mechanisms of the Wnt/β-catenin pathway in regulating osteogenic differentiation. Additionally, future studies should include in vivo research to verify and expand upon the results of this study. On the other hand, It is crucial to explore different formulations and delivery systems to enhance the stability and bioavailability of these compounds. Looking ahead, since current findings are inspiring, a substantial amount of work is required to fully exploit the therapeutic potential of flavonoids in periodontal regeneration. Conclusions Flavonoids effectively reduce inflammation and intracellular reactive oxygen species in HGF and HBMSC, while simultaneously promoting osteogenic differentiation in HBMSC. The innovative soft-hard tissue hybrid chip facilitates signal crosstalk between soft and hard tissues and integrates a dynamic culture system to biomimetically replicate real tissue structures. This feature makes the chip an effective tool for drug screening, particularly for the selection and evaluation of flavonoid compounds. These capabilities underscore the potential of flavonoids as therapeutic medicine in the treatment of periodontal diseases and other related conditions. Supplementary Information Supplementary Material 1. Supplementary Material 1.