Genistein from Thai Fermented Soybean (Thua-nao) Attenuates Osteoarthritis Pathogenesis via Dual Inhibition of NF-κB and MAPK Signaling: Evidence from Ex Vivo, Computational, and In Vitro Models

Abstract

Aglycone isoflavones derived from fermented soybeans demonstrate antioxidant, anticancer, antiobesity, and anti-inflammatory properties, with evidence suggesting protective effects against cartilage deterioration. However, precise mechanisms in chondrocytes remain unclear. This study screened extracts from Thai fermented soybean (Thua-nao) for chondroprotective effects using an ex vivo porcine cartilage explant model. Dichloromethane (DCM) and ethyl acetate (EA) extracts enriched in daidzein and genistein, primary aglycone isoflavones, effectively attenuated the IL-1β/OSM-induced extracellular matrix depletion from explants. Computational analysis predicted genistein to inhibit NF-κB and MAPK signaling effectively, with high gastrointestinal absorption. The in vitro experiments using primary chondrocytes showed that genistein upregulated ACAN while downregulating aggrecanase genes, thereby affecting key chondrogenic markers. Additionally, genistein inhibited IL-1β-mediated MAPK and NF-κB activation and suppressed matrix metalloproteinases (MMP) -1, -3, and -13 expression. Altogether, these findings demonstrate the chondroprotective properties of Thai fermented soybean extract and strengthen the potential of genistein as a promising therapeutic candidate for osteoarthritis.

Introduction

Inflammation plays a pivotal role in the pathophysiology of osteoarthritis (OA), primarily by activating pro-inflammatory cytokines that affect resident joint cells. Interleukin-1 beta (IL-1β), a proinflammatory cytokine secreted by synovial fibroblasts, initiates inflammatory cascades in chondrocytes, resulting in metabolic dysregulation. , The typical metabolic imbalance in OA chondrocytes is characterized by upregulation of matrix metalloproteinases (MMPs) and downregulation of cartilage matrix proteins, including aggrecan and type II collagen. This imbalance leads to degradation of the cartilage matrix and chondrocyte apoptosis, resulting in clinical manifestations such as joint pain, swelling, functional impairment and mobility disorders. Elevated levels of IL-1β have been documented in OA cartilage and synoviocytes, correlating with disease progression. Strategies aimed at downregulating IL-1β expression or mitigating inflammatory signaling pathways have demonstrated the potential to decelerate OA progression.

Current pharmacological interventions for managing OA symptoms predominantly include nonsteroidal anti-inflammatory drugs (NSAIDs) and cartilage-protective agents; however, these treatments are often associated with adverse effects and exhibit limited efficacy. Consequently, there is growing interest in exploring the anti-inflammatory properties of phytochemicals as alternative therapeutic strategies for OA management.

Soybeans, a staple in East Asian diets, are rich in aglycone isoflavones such as daidzein, genistein, and glycitein, as well as their glycoside derivatives. Fermentation significantly increases the aglycone content of soybeans; fermented soybeans (Thua-nao) have been reported to contain higher concentrations of aglycone isoflavones than their nonfermented counterparts. Numerous in vivo and in vitro studies indicate that soybean isoflavones confer substantial health benefits. The anti-inflammatory and antiarthritic properties of fermented soybeans have been demonstrated in mouse macrophage cell lines (RAW264.7) and human chondrocyte cell lines (SW1353). Additionally, oral administration of soybean isoflavones has been shown to reduce cartilage degradation in ovariectomized rat models of osteoarthritis. Daidzein and genistein, the primary aglycone isoflavones in fermented soybeans, exhibit chondroprotective properties, however, their specific anti-inflammatory mechanisms within chondrocytes remain to be fully elucidated. To explore these mechanisms, an ex vivo porcine cartilage explant model was used to evaluate the protective effects of fermented soybean against cartilage destruction. The results demonstrated that the extract enriched in aglycone isoflavone (daidzein and genistein) reduced cartilage deterioration and preserved cartilage matrix molecules. Molecular docking studies were additionally performed to predict the interactions between key aglycone isoflavones in fermented soybean extract (daidzein, genistein, and genistin) and inflammatory-relevant signaling proteins, specifically major IL-1βmediated signaling molecules using AutoDock Vina. Furthermore, oral bioavailability and brain or intestinal permeability of these compounds were predicted using Lipinski’s Rule of Five (Ro5) and the BOILED-Egg model via SwissADME. Through computer-aided drug design (CADD), we identified and optimized potential isoflavones from fermented soybean extracts; notably, genistein demonstrated superior efficacy compared to daidzein and genistin. Subsequent evaluations in human primary chondrocytes revealed that genistein effectively attenuated IL-1β-induced MMP-1, -3, and -13 transcription and secretion by inactivating the MAPK and NF-κB pathways.

This study emphasizes the therapeutic potential of fermented soybean-derived aglycone isoflavones, particularly genistein, as promising candidates for OA management. Their capacity to inhibit key inflammatory signaling pathways provides a foundation for developing innovative phytochemical-based interventions that could minimize the adverse effects commonly associated with conventional treatments.

Materials and Methods

Fermented Soybean Extract Preparation

Thai soybeans (Glycine max (L.) Merrill) were soaked and steamed for 3 h until soft, then packed into a sterile bucket and allowed to spontaneously ferment at ambient temperature for 3 days. Twenty grams of Thai fermented soybeans were added to 95% ethanol (100 mL) and homogenized for 5 min. The mixture was then agitated at 200 rpm for 2 h, followed by centrifugation at 5,000 rpm for 5 min. The crude extract was concentrated by evaporating the ethanol under vacuum using a rotary evaporator (Buchi, Switzerland) at 175 mbar and 40 °C to obtain the ethanolic fraction. The ethanolic fraction was further fractionated using organic solvents of different polarities, including hexane, dichloromethane, ethyl acetate, and methanol. Each extract was then concentrated under vacuum and air-dried to obtain the hexane, dichloromethane, ethyl acetate, and methanol fractions.

Investigation of Isoflavone in Fermented Soybean Extracts via HPLC

The isoflavone content of the fermented soybean extracts was examined using high-performance liquid chromatography. Briefly, 1 g of powder was extracted with 5 mL of methanol by shaking at 60 rpm in a 37 °C water bath for 12 h. The extracts were recovered by centrifugation at 12,000 × g for 15 min at 4 °C. The sample solution was filtered through a 0.45 μm membrane and then analyzed by HPLC. HPLC analyses were carried out on an Agilent Technologies 1100 series system equipped with an autosampler and a diode-array detector at 254 nm. Separation was performed on the Zorbax SB C18 reversed-phase column (15 cm × 4.6 mm i.d., 5 μm) from Agilent Technologies, USA using a gradient solvent system that started with 90% solvent A (MeOH:water, 10:90, v/v) containing 0.1% formic acid and 10% solvent B (MeOH with 0.1% formic acid), progressing to 100% solvent A for 25 min. The flow rate was established at 1.0 mL/min and the column temperature was maintained at 40 °C. Quantitative data for soybean isoflavones (daidzin, daidzein, genistin, and genistein) were obtained by comparison with respective authentic standards.

Investigation of Ex Vivo Chondroprotective Effects

The porcine articular cartilage explant culture was used as the screening model for the chondroprotective effects of fermented soybean extracts, as described in a previous study. Cartilage from the porcine metacarpophalangeal joint was sectioned into 25 cm³ discs and incubated in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 1% penicillin/streptomycin (100 U/mL penicillin, 100 μg/mL streptomycin) and 5% fetal calf serum (FCS) for 60 min at 37 °C in a 5% CO₂ environment for sterilization. The cartilage pieces (approximately 2.1 mm in diameter, 7.0 mm in length, and 1.0 mm in thickness) were then aligned and cultured in 24-well plates (30–35 mg per well) in DMEM for 24 h at 37 °C with 5% CO₂. Cartilage degradation was induced with IL-1β and OSM (each at 10 ng/mL), and the cartilage was cotreated with fermented soybean extract at concentrations of 12.5–50 μg/mL. Culture media were collected and replaced on days 0, 7, 14, 21, 28, and 35 to quantify s-GAG and hydroxyproline release. On day 35, cartilage discs were digested with 10 units of papain, and the amounts of hydroxyproline and uronic acid remaining were measured. All experiments were performed in triplicate using tissue obtained from the same donor animal. Each experiment was independently triplicated, including all measurements.

Measurement of Sulfated Glycosaminoglycan (s-GAG) Levels

The s-GAG released into the conditioned culture medium was measured using the dimethylmethylene blue (DMMB) assay. Two hundred microliters of DMMB were added to 50 μL of either standard chondroitin sulfate-C (0–40 μg/mL) or culture medium, resulting in the formation of the DMMB-s-GAG complex. The mixture was measured at 520 nm using a microplate reader (MULTISKAN EX, Thermo Fisher Scientific, USA), and the amount of s-GAG was determined from the standard curve.

Measurement of Uronic Acid Remaining in Explants

Remaining uronic acid content in the papain-digested cartilage disc was quantified using the m-hydroxydiphenyl in a colorimetric assay. Three hundred microliters of concentrated sulfuric acid-borate reagent were added to either standard glucuronic acid lactone or the samples, and the mixture was incubated for 15 min at 100 °C. After cooling on ice, 12 μL of carbazole solution was added, and the mixture was incubated for 15 min at 100 °C. The mixture was measured at 540 nm using a microplate reader (MULTISKAN EX, Thermo Fisher Scientific, USA), and the uronic acid (UA) level was determined from the standard curve.

Measurement of Hydroxyproline Levels

The collagen content in the conditioned media or papain-digested cartilage explants was determined using the hydroxyproline assay. Samples were initially hydrolyzed with 6 N HCl for 24 hours at 100 °C, freeze-dried, and reconstituted with distilled water. The hydroxyproline in each sample was oxidized to a pyrrole using chloramine T at pH 6. The intermediate developed a pink color when combined with 4-dimethylaminobenzaldehyde. The samples were then mixed with diluent solution (67% propan-1-ol), oxidant solution (50 mM chloramine T), and color reagent (7.5% 4-(dimethylamino)­benzaldehyde in propan-1-ol). The reaction was performed at 70 °C for 10–20 min. Absorbance was read at 540 nm using a microplate reader (MULTISKAN EX, Thermo Fisher Scientific, USA). Hydroxyproline levels in each sample were determined from the standard curve.

Histological Analysis

Cartilage samples were fixed in 4% paraformaldehyde overnight, embedded in wax, and then cut into 5 μm-thick sections perpendicular to the articular cartilage surface. Sections were stained with hematoxylin-eosin (H&E) and safranin-O to observe the cell/tissue morphology and s-GAG localization, respectively.

Molecular Docking and Pharmacokinetic Analysis

To gain insight into the efficacy of isoflavones from fermented soybean extract (daidzein, genistein, and genistin) in impairing the NF-κB and MAPK signaling pathways, the crystal structures of nine principal signaling molecules were established as receptors for molecular docking analysis. The receptors comprised IκBα (AlphaFold DB: AF-B0LXP6-F1, IκBβ (PDB: 4KIK, NF-κB (PDB: 3GUT, Ubc13 (PDB: 4ONM ), UbcH7 (PDB: 4Q5E ), Linear ubiquitin assembly complex or LUBAC (PDB: 6KC6 ), ERK (PDB: 2GPH, MEK1/2 (PDB: 7JUT, JNK1 (PDB: 1UKH, JNK2 (PDB: 3E7O, MAPK14 or p38α (PDB: 1WBV, and MAPK11 or p38β (AlphaFold DB: AF-Q15759-F1. Molecular docking analysis was conducted via AutoDock Vina (version 1.2.5) alongside a genetic algorithm (GA) for the initial screening. , The protonation states of the signaling proteins were evaluated via PROPKA, incorporated in the PDB 2PQR system at pH 7, whereas the pK _a values of the isoflavones were assessed using ChemAxon. Additionally, the drug-likeness characteristics of the isoflavones were evaluated based on Lipinski’s Rule of Five (Ro5) criteria. The brain-or-intestinal estimated permeation (BOILED-Egg) model was employed through SwissADME to concurrently assess the gastrointestinal absorption and cerebral accessibility of the compounds. The interaction profiles and critical interacting residues of signaling proteins were visualized utilizing UCSF Chimera software and Accelrys Discovery Studio 3.0 (Accelrys Inc.).

Investigation of In Vitro Anti-Inflammatory Effect

The anti-inflammatory properties of genistein were examined in primary human articular chondrocytes (HAC) isolated from the amputated joints obtained from the patients at Maharaj Nakorn Chiang Mai Hospital, Chiang Mai University, Chiang Mai, Thailand, with informed consent and ethical committee approval (Ethics approval no. ORT-2563-07112). All methods were carried out in accordance with relevant guidelines and regulations.

Briefly, HACs were treated with 1 ng/mL IL-1β and varying concentrations of genistein (10, 20, 40 μM, Extrasynthese, France) for 24 h. Cell lysates were collected to assess chondrogenic and proteinase gene expression via qRT-PCR. Conditioned culture media were obtained for analysis of MMP-1, -3, and -13 protein expression using ELISA.

In experiments investigating cellular signaling pathways, HACs were pretreated with varying concentrations of genistein for 2 h, followed by exposure to 1 ng/mL IL-1β for an additional 10 min for the MAPK pathway and 15 min for the NF-κB pathway. Cell lysates were collected to investigate phosphorylation of signaling proteins via Western blotting. Additionally, the MAPK and NF-κB signaling inhibitors, specifically U0126 (ERK inhibitor), SP600125 (JNK inhibitor), SB203580 (p38 inhibitor), and BAY11-7082 (NF-κB inhibitor), were utilized for comprehensive analysis. HACs were pretreated with either 40 μM genistein or 10 μM signaling inhibitors (BAY11-7082, U0126, SB203580, SP600125) for 2 h, followed by 1 ng/mL IL-1β and incubated for an additional 24 h prior to the determination of respective phosphorylated signaling proteins.

Quantitative Reverse Transcription-Polymerase Chain Reaction (qRT-PCR)

An RNA extraction kit (Cytiva, USA) was utilized to isolate total RNA from cell lysate samples. First-strand cDNA was synthesized from 500 ng of total RNA using a reverse transcription kit (Bioline, Meridian Bioscience, USA), followed by cDNA amplification on an ABI 7000 real-time fluorescence quantitative PCR instrument (Applied Biosystems; Thermo Fisher Scientific, USA) with HOT FIREPol EvaGreen qPCR Mix Plus (ROX) (Solis BioDyne, Estonia). The quantitative PCR was performed according to the following thermal cycling protocol: initial activation at 95 °C for 12 min, followed by 40 cycles of 95 °C for 15 s, 60 °C for 20 s, and 72 °C for 20 s. The sequences of the oligonucleotides used as specific primers to amplify genes are shown in Table . All primers were obtained from Macrogen, Inc. (South Korea). GAPDH (the housekeeping gene) was used as an endogenous control. Data were calculated using the 2^–ΔΔCT relative quantification method.

1. Primer Sequences Used for RT-qPCR Were as Follows.

Name	Sequence (5′-3′): Forward (F); Reverse (R)
ACAN	F: ACTTCCGCTGGTCAGATGGA
	R: TCTCGTGCCAGATCATCACC
COL2A1	F: GGTGGCTTCCATTTCAGCTATG
	R: TTGCAGTGGTAGGTGATG TTCTG
SOX9	F: CATGAGCGAGGGCACTCC
	R: C TCGCTTCAGG TCAGCCTTG
MMP-1	F: CTGGCCACAACTGCCAAATG
	R: GCTGTCCCTGAACAGCCCAGTACTTA
MMP-3	F: TTTTGGCCATCTCTTCC TC
	R: TGTGGATGCCTCTTGGGTAT
MMP-13	F: TCCCAGGAA TTGGTGATAAAGTAGA
	R: CTGGCATGACGCGAACAATA
ADAMTS4	F: GCAACGTCAAGGCTCCTCTT
	R: CTCCACAAATCTACTCAGTG AAGCA
ADAMTS5	F: AGGAGCACTACGATGCAGCTATC
	R: CCCAGGGTGTCACATGAATG
GAPDH	F: AGGGCTGCTTTTAACTCTCGT
	R: CCCCACTTGATTTTGGAGGGA

ELISA for MMP-1, -3 and -13 Quantification

Matrix metalloproteinases (MMP-1, -3, and -13) were measured in culture media using commercial ELISA kits (R&D Systems, USA) in accordance with the manufacturer’s guidelines. For each MMP assay, 100 μL of Assay Diluent (RD1-52 for MMP-1 and -3; RD1-34 for MMP-13) was added to precoated microplate wells. Subsequently, 100 μL of sample or standard was added for MMP-1 and -3 assays (standard range: 0–10 ng/mL), whereas 50 μL was added for MMP-13 (standard range: 0–5,000 pg/mL). Plates were incubated for 2 h at room temperature on an orbital microplate shaker set at 500 rpm. After four washes with Wash Buffer, 200 μL of the respective MMP conjugate was added to each well and incubated as before. After another four washes, 200 μL of substrate solution was added and incubated for 30 min at room temperature, protected from light. The enzyme reaction was stopped by adding 50 μL of stop solution to each well. Absorbance was measured at 450 nm using a microplate reader (MULTISKAN EX, Thermo Fisher Scientific, USA). MMP concentrations were determined by fitting a four-parameter logistic curve to the standard values.

Western Blotting

The cell lysate was centrifuged at 20,000 × g for 20 min. The supernatant was analyzed to quantify total protein using the Bradford method. A total protein sample of 50 μg was subjected to 10% SDS-PAGE and subsequently transferred to a nitrocellulose membrane. After blocking with 5% nonfat milk in PBS containing 0.2% Tween-20, the membrane was incubated with antibodies against p-JNK, Total-JNK, p-ERK, Total-ERK, p-P38, Total-P38, p-P65, Total-P65, p-IKK, and Total-IKK (Cell Signaling Technology, USA), followed by suitable horseradish peroxidase (HRP)-conjugated secondary antibodies at a dilution of 1:20,000. The protein bands visualized by enhanced chemiluminescence were captured with a ChemiDoc Imaging System (Bio-Rad, USA) and densities of the bands were quantified using ImageJ software with normalization to the loading control, β-actin.

Statistical Analysis

All data are presented as mean ± SD (standard deviation). For statistical analysis, we used the independent t test and one-way ANOVA. Statistical significance was expressed as p values of *, # < 0.05; **, ## < 0.01; ***, ### < 0.001 and ****, #### < 0.0001. All statistical analyses were performed using Prism GraphPad software.

Results

Analysis of Isoflavones in Thai Fermented Soybean Extracts

Ethanolic extraction of Thai fermented soybeans yielded 7.52%. Fractionation yielded hexane (0.60%), dichloromethane (0.11%), ethyl acetate (0.11%), and methanol (0.29%). Subsequently, HPLC was used to quantify the aglycone isoflavones in the fermented soybean extract, based on previous studies reporting elevated levels of these compounds in fermented soybean products. The HPLC profiles and isoflavone contents of the fermented soybean extract are presented in Figure and Table , respectively. Our analysis revealed that the dichloromethane (DCM) and ethyl acetate (EA) fractions contained high levels of daidzein and genistein. Notably, a marked difference in the relative abundance of these two isoflavones was observed in the EA fraction, where genistein content was 2.27 times higher than that of daidzein.

1.

HPLC profile of daidzein or genistein. The HPLC profiles and isoflavones contents of fermented soybean extract were quantified in different fractions including Standard (A), Methanol (B), Ethanol (C), Hexane (D), Dichloromethane (E) and Ethyl Acetate (F).

2. Isoflavone Contents of Fermented Soybean Extract in Different Fractions.

Fraction	Ethanol	Hexane	Dichloromethane	Ethyl acetate
Genistin (μg/mg extract)	10.06 ± 1.52	3.11 ± 0.52	13.64 ± 1.21	157.46 ± 25.79
Daidzein (μg/mg extract)	76.94 ± 11.03	62.93 ± 7.15	489.95 ± 67.32	323.57 ± 49.22
Genistein (μg/mg extract)	90.22 ± 13.58	119.06 ± 12.97	416.79 ± 57.73	735.88 ± 114.94

Effect of Fermented Soybean Extracts on Cartilage Degradation

The chondroprotective effects of fermented soybean extracts were assessed using IL-1β/OSM-induced cartilage degradation model. After 35 days of treatment, IL-1β/OSM significantly increased s-GAG release from cartilage explants by 1.7-fold compared with untreated explants. Co-treatment of IL-1β/OSM with DCM or EA fractions reduced s-GAG release, with high doses decreasing it by 1.8- and 1.6-fold, respectively, compared with IL-1β/OSM treatment alone (Figure B). Additionally, IL-1β/OSM significantly increased the accumulation of hydroxyproline, a modified amino acid found in collagen, in the media by 2.2-fold (Figure C). Consistently, cotreatment with high doses of DCM and EA extracts reduced IL-1β/OSM-induced hydroxyproline release by 3.7- and 1.7-fold, respectively, compared with IL-1β/OSM treatment alone (Figure C). The levels of UA and hydroxyproline in papain-digested cartilage explants, reflecting remaining glycosaminoglycan and collagen, were significantly decreased (4.2- and 98.8-fold, respectively) in IL-1β/OSM-treated explants (Figure D,E), whereas cotreatment with DCM extract attenuated the reduction of both markers by 1.3- and 3.5-fold, respectively. Similarly, cotreatment with EA extract also mitigated these effects by 1.2- and 3.2-fold, respectively. Safranin-O staining confirmed s-GAG content in the cartilage matrix, showing lower staining intensity in IL-1β/OSM-treated explants (Figure F,G). Co-treatment with DCM and EA extracts increased safranin-O intensity compared with IL-1β/OSM treatment alone. In contrast, neither methanol nor hexane extracts showed protective effects against IL-1β/OSM-induced s-GAG loss from the cartilage matrix. All results from cartilage explant models demonstrated that a high dose (50 mM) of each extract had no effect on cartilage degradation.

2.

Effects of fermented soybean extracts on cartilage degradation. Schematic representation of the experimental design. Porcine cartilage explants were cultured in the presence of IL-1β and OSM to induce cartilage degradation, followed by treatment with different solvent extracts (methanol, hexane, dichloromethane, and ethyl acetate) at concentrations of 12.5, 25, or 50 μg/mL for 35 days (A) The GAGs (B) and hydroxyproline (C) were quantified in the culture media. The porcine cartilage explants were digested and the amount of uronic acid (D) and hydroxyproline (E) were detected. The porcine cartilage explants were cut and processed, then stained with H&E or Safranin O, observed under a microscope (F), and the Safranin O images were converted to intensity (G). p values of **, ## < 0.01, and ***, ### < 0.001. * compared with CTRL group, # compared with IL-1β treated group.

Molecular Docking-Based Insights into the Potential Modulation of IL-1β-Mediated Signaling Pathways by Fermented Soybean Isoflavones

It is well established that IL-1β activates activator protein-1 (AP-1) and NF-κB in chondrocytes through the MAPK family and IKK-mediated phosphorylation of IκB, respectively. Activation of these signaling pathways contributes to cartilage degradation in OA. Several matrix-degrading enzymes, including aggrecanases (ADAMTS-4 and -5), MMP-1, -3, and -13, are upregulated via IL-1β signaling pathways. To explore the potential mechanisms for inhibitory effects of isoflavones on IL-1β-induced cartilage degradation, molecular docking analyses were performed using the major isoflavones present in the DCM and EA fractions of fermented soybean extracts (daidzein, genistein, and genistin) against MAPK family and NF-κB-related proteins.

To refine virtual screening and examine the potential interactions of isoflavones with signaling molecules, a molecular docking study was conducted using AutoDock Vina. Daidzein, genistein, and genistin showed binding free energies (ΔG _bind) ranging from −4.7 to −9.8 kcal/mol, comparable to or more negative than those of their respective reference inhibitors (Table ). This suggests that these isoflavones bind to signaling molecules with similar or greater affinity than the reference systems. Regarding the binding interactions between isoflavones and MAPK/NF-κB signaling molecules, molecular docking results indicated that hydrophobic interactions were the primary contributors to binding.

Among the MAPK signaling proteins, majority of interactions were hydrophobic. Key interacting residues included L118, I141, and M143 (MEK1/2); I32, V40, and L168 (JNK2); L74 (p38α); and K53 and L167 (p38β), as depicted in Figure A and Supplementary Figures S1. In addition, these isoflavones formed H-bonds with T23 and C95 (IκBα) and with E97 and C99 (IκBβ) in the NF-κB system, and with K97 and V127 (MEK1/2) and with R127 and K160 (JNK1) in the MAPK system, as illustrated in Figure A.

3.

Three-dimensional orientation of key isoflavones: daidzein (blue), genistein (yellow), and genistin (pink), with the active site residues of MAPK (A), NF-κB (B) signaling molecules and ubiquitination enzymes/complex that activate IKK complex (C). The revealed residues interact with the compounds through hydrogen bonds. Brain-to-intestinal estimated permeation (BOILED-Egg) plots for all isoflavones (D). Molecules predicted to permeate the blood-brain barrier (BBB) are in the yellow zone. The white area represents absorbable gastrointestinal (HIA) molecules. The blue and red dots denote predicted permeable glycoprotein substrates (PGP⁺) and nonsubstrates (PGP^–).

Isoflavones interacted with several key residues in NF-κB signaling proteins, including L21, V29, and I164 of IκBα, as well as L21, V29, A42, and V152 of IκBβ. Hydrogen bonds (H-bonds) were formed between the potent compounds and residues T23 and C98 of IκBα. Similarly, H-bonds were observed between isoflavones and residues E97 and C99 of IκBβ. Isoflavones also contacted the catalytic cysteines of Ubc13, UbcH7, and LUBAC, which are specific targets of BAY 11-7082, the known inhibitor of NF-κB signaling pathway. The compounds interacted with R85 and C87 of Ubc13 through multiple forces, including hydrophobic, electrostatic, and H-bond contacts, corresponding to the interactions with K67 and C86 of UbcH7. Furthermore, in LUBAC, residues C885, L992, and R935 interacted through hydrophobic, H-bond, and hydrophobic interactions, respectively, as shown in Figure B,C, and Supplementary Figure S1.

Subsequently, drug-likeness analysis was performed using the SwissADME platform to assess the physicochemical properties and bioavailability of the molecules. The results indicated that only daidzein and genistein complied with Lipinski’s rule of five (Ro5) criteria. These criteria include (i) a molecular weight (MW) not exceeding 500 Da, (ii) a maximum of 5 hydrogen bond donors (HBD), no more than 10 hydrogen bond acceptors (HBA), and (iii) a log P value of 5 or lower. Moreover, gastrointestinal absorption of these compounds was estimated using the BOILED-Egg method. As illustrated in Figure D, genistein falls within the white zone, suggesting superior gastrointestinal absorption relative to daidzein. Based on computational analyses, genistein was identified as a promising candidate for drug screening and subsequently selected for further in vitro investigation.

Effect of Genistein on Chondrocyte Gene Expression

OA pathological conditions are mediated by interleukin-1beta (IL-1β), a principal inflammatory cytokine present in OA joint tissues. This cytokine activates genes pertinent to osteoarthritis, including those associated with inflammation and proteolysis, while simultaneously downregulating chondrogenic genes. Thus, the effects of genistein on IL-1β induced human articular chondrocyte (HAC) were evaluated.

The results indicated that 24 h-treatment with IL-1β (1 ng/mL) significantly altered the markers of cartilage homeogenesis, downregulating the expression of key chondrogenic and cartilage matrix genes, specifically aggrecan core protein gene (ACAN), SRY-Box Transcription Factor 9 (SOX9), and collagen type II (COL2A1). ACAN expression significantly increased in the genistein (20, 40 μM) cotreated group compared with the IL-1β alone-treated group, whereas COL2A1 and SOX9 expressions remained unchanged (Figure A). The upregulation of aggrecanases (ADAMTS-4 and -5) following IL-1β induction has been previously documented. Our results showed that genistein significantly reduced IL-1β-mediated aggrecanases expression reflecting the potential chondroprotective effects (Figure B).

4.

Effects of Genistein on gene expression of IL-1β induced chondrocytes. HAC was cotreated with IL-1β (1 ng/mL) and daidzein or genistein (0–40 μM) for 24 h, then the mRNA levels of the chondrogenic gene expression (ACAN, COL2A1, SOX9) (A) and catabolic gene expression (ADAMTS-4, -5,MMP-1, -3,-13) (B, C) were explored. The levels of MMP-1, MMP-3 and MMP-13 in the culture medium were measured via ELISA (D). p values of *, # < 0.05; **, ## < 0.01, and ***, ### < 0.001. * compared with CTRL group, # compared with IL-1β treated group.

Effect of Genistein on Matrix Metalloproteinases

In addition to aggrecanases, we also examined the regulation of matrix metalloproteinases by genistein in IL-1β-treated HAC. MMP-1 and MMP-13 are recognized as potent enzymes that degrade collagen type II, particularly during the progression of osteoarthritis. Specific substrates of MMP-3 include proteoglycans and collagen types IX and XI, which associate intricately with type II collagen to form collagen fibrils within the cartilage extracellular matrix. , MMP-3 also activates other cartilage MMPs and amplifies ECM breakdown. Following 24 h of treatment, IL-1β (1 ng/mL) markedly enhanced the expression of MMP-1, -3, and -13 at both mRNA and protein levels (Figure C,D). Co-treatment with genistein resulted in a significant reduction of MMP-1, -3, and -13 at both the gene and protein secretion, implicating ameliorating ability of genistein on cartilage matrix degradation. High doses of genistein alone did not influence the expression of MMP-1, -3, or -13, as anticipated.

Effect of Genistein on IL-1β-Mediated Signaling Pathways

IL-1β-induced MMP expression signals through the MAPK and NF-κB pathways. The molecular mechanisms by which genistein modulates the IL-1β signaling were investigated in chondrocytes. The findings of the in vitro investigation of genistein’s effect on IL-1β-mediated phosphorylation of the MAPK and NF-κB signaling molecules aligns with the molecular docking results, as illustrated in Figure and Table . There is no evidence that genistein inhibits MAPK phosphorylation in human chondrocytes; however, genistein (50–100 μM) has been shown to inactivate ERK and JNK in human synoviocytes (MH7A cells) and p38 in rat intervertebral disc cells. In this study, genistein (40 μM) significantly reduced phosphorylation of JNK and p38, as shown in Figure A. Inhibition of ERK activation was also observed with genistein and IL-1β cotreatment, but no significant difference was observed compared with IL-1β alone. These in vitro findings related to the MAPK pathway are consistent with the molecular docking results, which showed moderate to strong binding to the signaling molecules (binding affinity between −5.6 to −8.1 kcal/mol). Moreover, an inhibition of the NF-κB signaling pathway by genistein was observed (Figure B), which corresponds to the previous studies. , The binding affinity of genistein to NF-κB signaling molecules was around −8.2 to −8.5 kcal/mol.

3. Binding Affinity (kcal/mol) of the Majority of Isoflavones and NF-κB/MAPK Signaling Molecules.

Signaling pathways	Enzyme/complex/signaling molecules	ΔG bind (kcal/mol)
		BAY11-7082 (inhibitor of IKK complex activation)	Daidzein	Genistein	Genistin
Ubiquitination enzymes and complex that activate IKK complex	Cysteine residues on Ubc13	–3.9	–6.1	–6.4	–7.4
	Cysteine residues on UbcH7	–3.0	–5.7	–5.5	–5.2
	Linear ubiquitin assembly complex (LUBAC)	–4.5	–4.7	–4.9	–4.9
NF-κB	IκBα	–6.3	–7.8	–8.2	–8.4
	IκBβ	–6.8	–8.4	–8.4	–9.6
	NF-κB	–6.5	–8.0	–8.5	–9.8
ERK/MAPK		U0126 (inhibitor of MAP1/2)	Daidzein	Genistein	Genistin
	ERK	–6.0	–7.1	–7.1	–6.8
	MEK1/2	–8.0	–7.8	–8.0	–9.2
JNK/MAPK		SP600125 (inhibitor of JNK)	Daidzein	Genistein	Genistin
	JNK1	–6.2	–5.4	–5.6	–5.9
	JNK2	–8.9	–8.4	–8.0	–9.3
p38/MAPK		SB203580 (inhibitor of MAPK)	Daidzein	Genistein	Genistin
	MAPK14 (p38α)	–8.9	–7.6	–7.9	–8.6
	MAPK11 (p38β)	–9.2	–8.1	–8.1	–8.1

5.

Effects of genistein on MAPK and NF-κB pathway in IL-1β induced chondrocytes. HACs were pretreated with varying concentrations of genistein for 2 h, followed by 1 ng/mL IL-1β for an additional 10 and 15 min for the MAPK (A) and NF-κB (B) pathways, respectively. Cell lysate was harvested to study the phosphorylation of signaling proteins by Western blotting. p values of *, # < 0.05; **, ## < 0.01, and ***, ### < 0.001. * compared with CTRL group, # compared with IL-1β treated group.

To determine whether genistein inhibits IL-1β-induced MMP expression via MAPK and NF-κB pathways, chondrocytes were pretreated with genistein or specific pathway inhibitors (SP600125, SB203580, or BAY11-7082) prior to the assessment of signaling pathways activation and MMP production. Genistein (40 μM) and inhibitors (10 μM) significantly decreased phosphorylation of JNK, p38, and p65 (NF-κB), with a concomitant reduction in ERK phosphorylation (Figure ).

6.

Comparative of effects of genistein and specific signaling inhibitors on MAPK and NFκB pathway in IL-1β induced chondrocytes. HACs were pretreated with either genistein (40 μM) or respective signaling inhibitors (10 μM) for 2 h, followed by 1 ng/mL IL-1β for an additional 10 and 15 min for the MAPK (A) and NF-κB (B) pathways, respectively. Cell lysates were harvested to examine the phosphorylation of signaling proteins using Western blotting. p values of *, # < 0.05; **, ## < 0.01; ***, ### < 0.001, and ****, #### < 0.0001. * compares with the CTRL group, # compares with IL-1β treated group.

In line with the effects of genistein, inhibitors of MAPK and NF-κB pathways also reduced IL-1β-induced MMP-1, -3, and -13 expression (Figure ), suggesting that genistein may act through these pathways.

7.

MAPK and NF-κB pathway involvement in genistein-mediated suppression of IL-1β-induced MMP expression. The MAPK and NF-κB signaling inhibitors were used to investigate the direct effects of genistein on IL-1-induced MMP expression. HAC was pretreated with either 40 μM genistein or 10 μM signaling inhibitors along with 1 ng/mL IL-1β for 24 h. The cell lysate and culture medium were collected to measure the mRNA expression of MMP-1, MMP-3, and MMP-13 (A) and protein expression (B) using qRT-PCR and ELISA, respectively. p values of # < 0.05; ## < 0.01; ### < 0.001, and *** < 0.001 are considered significant. * indicates comparison with the CTRL group, and # indicates comparison with the IL-1β-treated group.

Discussion

This study revealed that the primary isoflavones in Thai fermented soybean (Thua-nao) extract are daidzein and genistein, produced through Bacillus spp. fermentation. These aglycone isoflavones also present in other traditional fermented soybeans from different countries: tempeh, Doenjang and chungkookjang. Fermentation can increase aglycone content by hydrolyzing glucoside precursors (e.g., daidzin to daidzein), potentially improving bioavailability compared to nonfermented soybean rich in glucosides. Isoflavones from soy products, including fermented forms, have shown potential health benefits in preclinical studies, such as antioxidant, anti-inflammatory, and metabolic effects, however, human clinical outcomes remain inconsistent and require further validation.

A study by Okabe et al. revealed that dietary aglycone isoflavones from fermented soybeans, rich in daidzein and genistein, inhibited bone resorption and bone loss in ovariectomized rats. On the other hand, soybean with high levels of isoflavone glucosides (daidzin, glycitin, and genistin) was also shown to restrain cartilage degeneration in rats, suggesting that both forms may confer benefits for bone health. Aglycone forms exhibit superior intestinal absorption and bioactivity in some models, however, their advantages are context-dependent, with limited direct comparisons in osteoarthritis (OA) models and variable translation to humans due to factors such as gut microbiota influencing equol production from daidzein.

The ex vivo porcine cartilage explants exhibited chondroprotection by DCM and EA extracts of fermented soybean, which contained high levels of daidzein and genistein. Moreover, these extracts attenuated inflammation-induced degradation of the cartilage matrix by reducing the loss of glycosaminoglycans and collagen from IL-1β/OSM-induced cartilage explants. Analysis of the culture medium corresponded to the levels of uronic acid and hydroxyproline remaining in the cartilage. The histological study additionally confirmed that DCM and EA extracts protected against the loss of sulfated glycosaminoglycan content in the cartilage matrix. Previous studies utilizing computational approaches have shown that genistein interacts with the key receptor-binding domain of the spike protein, potentially inhibiting viral spread, and have also suggested its ability to suppress the progression of castration-resistant prostate cancer. For analysis of the potential candidates in DCM and EA extracts, the computer-aided drug design (CADD) was further utilized to screen isoflavones with potential to inhibit IL-1β mediated signaling pathways in OA. The molecular docking analysis demonstrated that the three isoflavonesdaidzein, genistein, and genistinexhibited strong binding affinities with the majority of IL-1β mediated signaling molecules in comparison to reference compounds. Notably, genistein not only demonstrated a potential inhibitory effect against NF-κB/MAPK signaling-related molecules but also was predicted to possess favorable physicochemical characteristics, aligning with Lipinski’s criteria and demonstrating high passive gastrointestinal absorption (white ellipse plot on BOILED EGG graph (Figure D). Previous studies reported that genistein can be detected in circulation after oral administration and absorption, with an oral bioavailability of 3–7%, sufficient to achieve detectable levels in synovial fluid. , Oral administration of genistein (40 mg/kg) was shown to reduce IL-1β levels in synovial fluid in a rat OA model and suppress TNF and VEGF in synovial fluid of OA mice, supporting its effective oral bioavailability in agreement with our computational findings. Moreover, genistein demonstrated greater efficacy over daidzein in suppressing OA pathology, with broader anti-inflammatory and chondroprotective effects both in vitro and in vivo, whereas daidzein showed a limited efficacy in OA model.

Based on the literature review and CADD results, the molecular mechanisms of genistein in primary human articular chondrocytes were explored, and IL-1β was selected as the inflammatory inducer in the in vitro study due to its pivotal role in OA pathogenesis. A previous study demonstrated the inhibitory activity of genistein via downregulation of NO and COX2 in LPS and IL-1β induced chondrocytes, respectively. Our study demonstrated that genistein attenuates the IL-1β-mediated upregulation of chondrocyte catabolic genes and suppresses matrix metalloproteinases-1, -3, and -13, thereby preserving collagen type II and sulfated proteoglycans in the porcine cartilage explant model. The results of this study corroborate those of Liu FC and colleagues, who showed that genistein reduced MMP-1, -3, and -13 protein expression in IL-1-induced human chondrocytes through Nrf2-mediated inhibition of NF-κB signaling, consistent with findings from rat models. By tracking IL-1 signaling pathways in human chondrocytes, this study proposes that genistein transcriptionally regulates MMPs by inactivating MAPK and NF-κB, thereby down-regulating MMP-1, -3, and -13 expression at both the mRNA and protein levels in human chondrocyte lysates and conditioned medium, respectively. Genistein has previously demonstrated multitarget actions, including cytokine suppression, oxidative stress depletion, ECM preservation, and regulation of chondrocyte apoptosis. Inhibition of the NF-κB and MAPK signaling pathways in OA chondrocytes has been shown to confer significant chondroprotective and anti-inflammatory effects. Numerous studies have demonstrated that blocking NF-κB activity markedly reduces the expression of catabolic enzymes, particularly matrix metalloproteinases (MMPs) such as MMP-1, MMP-3, and MMP-13, as well as aggrecanases like ADAMTS-4 and ADAMTS-5; both of which are central mediators of cartilage extracellular matrix degradation. Additionally, NF-κB inhibition suppresses the production of pro-inflammatory cytokines (e.g., IL-6, TNF-α), inducible nitric oxide synthase (iNOS), and cyclooxygenase-2 (COX-2), thereby attenuating the inflammatory milieu that drives OA progression. Importantly, NF-κB suppression also helps restore the synthesis of key cartilage matrix components, such as type II collagen and aggrecan, and can reduce chondrocyte apoptosis, contributing to cartilage repair and maintenance.

Similarly, inhibition of the MAPK pathways downregulates both catabolic and inflammatory genes, further protecting cartilage integrity by limiting the breakdown of collagen and proteoglycans. MAPK pathway inhibition not only reduces abnormal chondrocyte activity but also enhances anabolic processes, thereby encouraging matrix regeneration. When both the NF-κB and MAPK pathways are targeted simultaneously, studies report synergistic chondroprotective effects, including robust suppression of matrix-degrading enzymes and inflammatory mediators, and enhanced cartilage matrix synthesis. , Collectively, these findings underscore the therapeutic potential of dual NF-κB and MAPK inhibition to slow OA progression and preserve cartilage homeostasis.

In conclusion, this study provides evidence that genistein in Thai fermented soybeans (Thua Nao) exhibits chondroprotective effects by preserving cartilage matrix molecules, thereby mitigating cartilage degradation. A combination of the in vitro signaling pathway study in primary human chondrocytes and the molecular docking model, comparing genistein with a specific IL-1β inhibitor, implies that genistein effectively suppresses MMP production via both the MAPK and NF-κB signaling pathways. Collectively, these findings highlight the dual-pathway inhibitory action of genistein and underscore its potential as a therapeutic agent for managing the pathogenesis of osteoarthritis. Despite the encouraging findings, this study is subject to certain limitations. A key limitation is that CADD predictions are in silico and lack direct evidence of cellular target engagement; biophysical assays (e.g., kinase inhibition, binding kinetics) are needed for validation. Besides, fermentation benefits are promising but not unique to aglycones, and human trials are essential to confirm their efficacy, dosing, and safety.

Data is provided within the manuscript or Supporting Information files. Further inquiries can be directed to the corresponding author.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jafc.5c11447.

• Two-dimensional diagrams of ligand-protein interactions for the three main isoflavones (daidzein, genistein, and genistin) with NF-κB and MAPK signaling molecules are shown in Supplement Figure S1 (PDF)

B.W., A.C., T.C., and P.P. contributed to the ideas and concepts of the study, analysis and interpretation of data, drafting, and revising of the manuscript. B.W., P.S., A.C., and A.W. contributed to the collection, assembly, and interpretation of data. H.C. and P.P. contributed to computational analysis. T.P., P.K., J.R., and P.P. contributed to the critical revision of the final manuscript. P.S. created all images in the TOC figure. P.P. obtained funding and provided final approval of the article. All authors read and approved the manuscript. B.W. and A.C. contributed equally to this work as first authors.

This research was funded by Chiang Mai University (FF66/023).

Use of artificial intelligence tools: During the preparation of this manuscript, the authors used Grammarly to assist with language editing. The authors reviewed and revised all AI-generated content, taking full responsibility for the accuracy and integrity of the work.

The authors declare no competing financial interest.