Co-assembled Glycyrrhiza nanoparticles embedded supramolecular protein hydrogels to enhance licochalcone A release for acute inflammation management

Abstract

Licochalcone A (LA) garnered remarkable acclaim in acute inflammation therapy, however, poor release capability from the matrix and oral bioavailability restrict its oral delivery. To address this challenge, licorice-derived glycyrrhizic acid (GA) and LA were co-assembled into GA-LA (GLA) binary co-assembled Glycyrrhiza nanoparticles (BCGNs), which were subsequently incorporated into supramolecular hydrogel matrix. GLA BCGNs demonstrated a remarkable capacity to scavenge various reactive oxygen species (ROS) and facilitated the cascade process of O₂^•−-H₂O₂-O₂ in vitro. Subsequently, GLA was dispersed in nano form into ovalbumin (OVA) and rhamnose (Rha) solutions, which were next self-assembled into OVA-Rha-GLA hydrogels. Remarkably, the introduction of Rha induced disordered secondary conformation of OVA, which decreased its mechanical properties and inherent binding energy, thereby shaping the three-dimensional supramolecular spatial structures of OVA-Rha-GLA networks. The assembly mechanisms indicated that the hydrogen bonding predominantly drove the assembly of loose supramolecular networks surrounded by -OH, -CH₂ and C O bonds on the Rha and OVA. Notably, the conformational transformation facilitated faster LA release, confirmed by computational simulation analysis, which was conducive to acute inflammation curation. Therefore, OVA-Rha-GLA exhibited excellent anti-inflammation and ROS-scavenging versatilities, displaying improved oral bioavailability compared to hydrogels lacking BCGNs or Rha in cellular and animal acute inflammation experiments. The results provided novel BCGNs-embedded supramolecular hydrogel systems to improve the drug release and anti-inflammatory bioactivities of LA, which demonstrated great promise in the management of acute inflammation.

Graphical abstract

1. Introduction

Acute inflammation is a rapid defense response of the body to injury or infection, usually characterized by redness, swelling, fever, pain, and dysfunction (Janoff and Schaefer, 1967; Wu et al., 2024), which generally leads to tissue damage, suppurative infections, and chronic inflammation prolongation. It predominantly causes systemic inflammatory response syndrome (SIRS) and Sepsis, threatening approximately 50 million patients' welfare, which remains a major concern in the world. Decreasing the secretion of pro-inflammatory factors such as TNF-α, IL-1β and IL-6, and lowering neutrophil infiltration are effective and critical means for acute inflammation intervention. Flavonoids are a large category of compounds consisting of two benzene rings interconnected by three carbon atoms (Premathilaka et al., 2022a; Wang et al., 2023b). Profiting from their unique structures, flavonoids harbor multifunctional health welfare, including anti-inflammatory, antioxidant, anti-microbial, anti-tumor, and anti-atherosclerotic bioactivities (Li et al., 2023; Pang et al., 2025). Therefore, they garnered remarkable acclaim in inflammation therapy (Xie et al., 2009). Nevertheless, a majority of flavonoids possess sparing solubility in water, low stability and dispersion in food, as well as poor bioavailability within the body (Liu et al., 2021; Zhong et al., 2022), bringing about a formidable challenge in oral administration. Licochalcone A is a typical flavonoid compound derived primarily from Glycyrrhiza inflata Batalin, which possesses prominent anti-inflammatory, antioxidant, and anti-allergy bioactivities (Li et al., 2022b). Specifically, deficiencies also impair its anti-inflammatory and antioxidant efficacy, which results in poor acute inflammation treatments (Liu et al., 2024).

Natural product self-assembled nanoparticles have been ubiquitously applied in drug delivery, bioimaging, and food science, ascribing to their biocompatibility, degradability and low toxicity (Li et al., 2024; Sadeghi et al., 2022; Yao et al., 2025). Notably, it achieves remarkable capability in enhanced drug solubility, targeted drug delivery, and improved drug stability. The natural components formed ordered nanostructures spontaneously through intermolecular forces such as hydrogen bonding, hydrophobic interactions, and electrostatic interactions (Hu et al., 2024c). Glycyrrhizic acid (GA) is an amphiphilic molecule also isolated from the dried roots of Glycyrrhiza that accounts for 3 %–8 % content in the herbal, consisting of a hydrophilic glucuronic acid and a hydrophobic glycyrrhetinic acid (Yu et al., 2021). The unique structure renders GA to assemble into nanoparticles when its concentration is higher than 0.2 mg/mL. Previous literature has highlighted its efficacy in reducing hydrophobicity and toxicity, improving the solubility, bioavailability as well as therapeutic activity of the hydrophobic payload (Selyutina and Polyakov, 2019; Su et al., 2017). Furthermore, edible GA itself owns versatile anti-inflammatory, antioxidant, hepatoprotective, and anticancer activities (Qian et al., 2022). Herein, we aimed to fabricate GA-LA (GLA) that was defined as binary co-assembled Glycyrrhiza nanoparticles (BCGNs), which were expected to boost the solubility, anti-inflammatory and antioxidant effect, as well as bioavailability of LA. Nevertheless, a new challenge emerged that direct BCGNs administration was difficult to achieve an ideal drug release and high stability in the human gastrointestinal tract with gastric acids, enzymes, and bile salts.

Hydrogels have attracted great interest in the delivery vehicles of the nanoparticles, ascribing to their three-dimensional cross-linked networks (Li et al., 2021; Xie et al., 2023). Typically, protein hydrogels are favored attributed to their excellent biocompatibility, biodegradability, nutritional properties, and edibility. They possess a well-defined sequence, more controllable molecular water cross-linking density, and homogeneous network structure when compared to other polymers (Fu et al., 2023; Zhao et al., 2025). Ovalbumin (OVA) is a single globular peptide chain constituting 385 amino acids, with a molecular mass of approximately 45 kDa (Weijers et al., 2002). It harbors the advantages of strong gelation, highly ordered structure, easy availability, and low cost. Previous studies established OVA-flavonoids hydrogels using five flavonoids with different structures and revealed that hydrogen bonding and hydrophobic force were the vital interactions driving their assembly (Zhong et al., 2022). Gan demonstrated that acylated OVA hydrogels exhibited superior drug-controllable abilities, hydrogel hardness, and recovery as well as water holding capacity than pure OVA (Hu et al., 2022). Based on the above, GLA CBGNs were next incorporated in the OVA hydrogels to augment their drug release and stability. However, positively charged OVA hydrogels probably exhibit poor miscibility and compatibility with negatively charged GLA. Furthermore, flavonoid ingredients generally possess poor release entropy in the highly ordered OVA hydrogels, which results in an unsatisfactory drug release for LA in the gastrointestinal tract. More importantly, a rapid LA release is critical for the treatment of acute inflammation. Consequently, it was pivotal to enhance the compatibility between OVA carrier and GLA CBGNs in the water microenvironment and provide a loose supramolecular helical structure for faster drug release.

Supramolecular systems are composed of polymers with a spatial structure consisting of units connected by reversible non-covalent bonds (Song et al., 2024b; Zhang et al., 2025), offering them unique properties (Tang et al., 2022). Rhamnose (Rha) is a natural six-membered ring structure with multiple polyhydroxy groups that have earned acclaim in pharmaceutical and food development attributed to its anti-inflammatory, immunomodulatory, anti-aging, and anti-tumor effects. Some investigations have documented that it effectively stabilized drug nanoparticles in delivery systems (Ren et al., 2023), suggesting that it demonstrated great promise in improving drug compatibility upon introduction into polymer matrices. Additionally, the spatial arrangement of polyhydroxy groups in Rha and pressure-sensitive adhesive (PSA) could form a supramolecular system surrounded by ether or amide bonds in the PSA (Song et al., 2024b). The supramolecular structure augmented binding enthalpy and release entropy, simultaneously increasing drug loading and release. Based on the above, we hypothesized that Rha could form supramolecular helical structures with OVA chains, which improved the stability of the nano-structure of GLA in the hydrogel networks and then increased its drug release. Still, the inherent gelation mechanisms of Rha-based supramolecular hydrogels in the water microenvironment were not systematically elucidated before.

In this comprehensive study, OVA-Rha-GLA hydrogels were novelly fabricated to enhance the drug release and anti-inflammatory efficacy of LA for acute inflammation management (Scheme 1). First, GLA CBGNs were facilely fabricated and characterized by X-ray diffraction, Fourier transform infrared spectroscopy (FTIR), Raman spectra, and molecular dynamics simulation (MDS). The various ROS-scavenging effect of GLA at different concentrations was investigated. GLA was next added to the OVA-Rha solution, whose secondary structures and intermolecular interactions were revealed by fluorescence spectroscopy, circular dichroism (CD), and molecular docking. Subsequently, OVA-Rha-GLA solutions were assembled into hydrogels, and their gelation mechanisms were unveiled by scanning electron microscope (SEM), FTIR, and MDS. After that, the rheological properties of different hydrogels were characterized. The in vitro release behaviors of LA from different hydrogels were explored, which were evidenced by the MDS test. Lastly, the enhanced anti-inflammatory, ROS-scavenging effect and oral bioavailability of LA in different hydrogels were studied in the LPS-induced acute inflammatory cellular and rat models. To our best knowledge, a joint application of protein hydrogels, Rha and co-assembled natural products in oral delivery for acute inflammation management has not been documented. In summary, the work proposes an effective system composed of supramolecular networks and co-assembled systems to improve the drug release and anti-inflammatory efficacy of hydrophobic flavonoid ingredients in the inflammation therapy.

Scheme 1.

Schematic of assembly, drug release and enhanced anti-inflammatory effect of OVA-Rha-GLA for acute inflammation management. The introduction of Rha induced disordered secondary conformation of OVA, shaping the three-dimensional supramolecular spatial structures. Moreover, the conformational transformation facilitated faster LA release. Therefore, the bioactive hydrogels exhibited enhanced anti-inflammation, ROS-scavenging, and oral bioavailability in the LPS-induced acute inflammation model.

2. Experimental methods

2.1. Materials, cell lines, and animals

Glycyrrhizic acid (GA, purity>98 %) was purchased from Shanghai Macklin Biochemical Co., Ltd. Licochalcone A (LA, purity>98 %) was supplied by Nanjing Spring & Autumn Biological Engineering Co., Ltd. Ovalbumin (OVA) was from Shanghai Yuanye Biotechnology Co., Ltd. Rhamnose (Rha) was obtained from Beijing Coolaber Technology Co., Ltd. Polyethylene glycol 400 (PEG 400), Trizol, cell counting kit-8 (CCK-8), lipopolysaccharide (E. coli-derived, LPS), 2,2′-azinobis (3-ethylbenzthiazoline-6-sulfonic acid) radical ion (ABTS), 2,2-diphenyl-1-picrylhydrazyl radical (DPPH) and phosphate buffer saline (PBS) were from Beijing Solarbio science & technology co., Ltd. Nitrogen blue tetrazolium chromogenic was obtained from Beyotime Biotechnology Co., Ltd. Fetal bovine serum (FBS) was obtained from Capricorn Scientific (America). PrimeScript RT Master Mix and TB Green Premix Ex Taq were provided by Takara (Japan). High-glucose Dulbecco's modified Eagle medium (DMEM-high), trypsin (0.25 %), and penicillin were from Gibco BRL Co., Ltd., USA. ROS assay kits were supplied by Nanjing Senbeijia Biotechnology Co., Ltd. CellROX-green dye was purchased from Invitrogen. CCK 8 kits were obtained from New Cell and Molecular Biotech. Co., Ltd. IL-1β, INOS, and TNF-α kits were acquired from Enzyme-free Biotechnology Co., Ltd. ARG1 and IL-10 kits were purchased from Biolegend (America). Methanol and acetonitrile for high-performance liquid chromatography (HPLC) were supplied by Merck (Germany). Other reagents are all analytical grade.

RAW264.7 macrophage cells were obtained from the Central Laboratory of Southern Medical University. They were grown in DMEM with 10 % FBS at 37 °C and 5 % CO₂ in an incubator.

Male Sprague-Dawley (SD) rats weighing 180–220 g were purchased from the Laboratory Animal Center of Southern Medical University. All animals were housed at 25 °C, 60 % humidity, and 12 h light/dark cycles and had free access to water and food. All animal experimental procedures Animal Research as per the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines, and received approval from the Ethics Committee at Southern Medical University.

2.2. Preparation of GLA CBGNs

GA-LA (GLA) CBGNs were fabricated using the dry film dispersion method. Briefly, 16.45 mg LA was dissolved in 5 mL of anhydrous ethanol, meanwhile, 100 mg GA was completely dissolved in 5 mL of water/anhydrous ethanol (V/V = 2:3) (The molar ratio of GA to LA was 2.50). Subsequently, the LA solution was dropwise added into GA solution under stirring, followed by the removal of ethanol in the round-bottomed flask using a rotary evaporator at 50 °C. The resulting film was uniform and yellow, which was then dissolved with 10 ml of preheated distilled water and incubated at 50 °C for 2 h. Lastly, the GLA CBGNs were acquired as a clear and transparent yellow liquid.

2.3. Characterization and the MDS behaviors of GLA CBGNs

The diameter distribution, polydispersity index (PDI), and zeta potential of the CBGNs were detected using a Zetasizer Nano ZS (Malvern, Worcs, UK). The morphology of GLA nano-CBGNs was photographed by transmission electron microscope (HITACHI, H-7650) at a voltage of 60 kV employing negative staining methods. Furthermore, the storage stability of GLA, including mean size and PDI within 6 months was exclusively recorded at 4 °C and 25 °C. The successful encapsulation of LA in GA was confirmed by FTIR, X-ray diffraction, and Raman analysis. FTIR technology was conducted on a Nicolet iS50 FTIR spectrometer (American Thermos, New York) from 400 to 4000 cm⁻¹ at a resolution of 2 cm⁻¹. X-ray diffraction was carried out at 5–60°2θ range with a scan rate of 10°/min. Furthermore, different samples were used for Raman analysis using the Renishaw RM2000, England Renishaw. The excitation wavelength was set at 532 nm, and the spectrum data were collected at 400–4000 cm⁻¹ with a spectral resolution of 4 cm⁻¹.

Gromacs software was donated to study the self-assembly behaviors of GLA in aqueous solution. First, a 10x10x10 nm box was constructed, which was randomly filled with GA and LA according to the molar ratio of 2.5:1, and the box was filled with water molecules. It was carried out under constant temperature, constant pressure, and periodic boundary conditions using the GAFF all-atomic force field and TIP3P water model. Next, the steepest descent method was performed to minimize the system energy. A 100 ps NVT (constant particle number, temperature, and pressure) and NPT (constant particle number, temperature, and volume) balance simulation was conducted, followed by a 50 ns MDS. In the MDS process, all the hydrogen bonds involved were constrained by the LINCS algorithm, and the integral step was 2 fs. The PME method was used to calculate the electrostatic interaction, the non-bond L-J interaction truncation distance was 1 nm, the V-rescale method was used to control the temperature at 298.15 K. Lastly, the visualization of the simulation results was completed by using the Gromacs embedded program and Pymol software.

2.4. In vitro CAT-Like and CAT-like activity of GLA

The removal of superoxide anion is the first step in the anti-ROS cascade reaction, classical nitrogen blue tetrazolium chromogenic (NBT) method was performed to assess the sod-like activity of GLA. The reaction system of xanthine and xanthine oxidase generates O₂•⁻, thereby reducing NBT to blue formic acid. SOD removes O₂•⁻, which inhibits the formation of formaldehyde. The final concentrations of GLA were 12.5, 25, 50, 100, and 200 μg/mL, and the reaction solution with NBT/enzyme was incubated with NBT/enzyme for 25 min at room temperature under dark conditions. The sod-like activity levels were then calculated based on the absorbance at 560 nm (Hu et al., 2024a).

The CAT-like activity of GLA was evaluated by determining the amount of oxygen produced by the decomposition of H₂O₂ using a dissolved oxygen meter. GLA suspensions (12.5, 25, 50, 100, and 200 μg/mL) were added to deionized water to achieve a final volume of 5 mL. Subsequently, H₂O₂ (1 mL, 2.0 M) was added and sealed with an anti-corrosion film. Then, a dissolved oxygen meter was used to measure dissolved oxygen in 300 s. Moreover, the capacity of GLA CBGNs to scavenge ABTS and DPPH was also performed.

2.5. Facile fabrication of OVA-Rha-GLA protein hydrogels

100 mg OVA and 50 mg (or 25 mg) Rha were completely dissolved in 5 mL distilled water using the ultrasound method, followed by the dropwise addition of GLA CBGNs solution under stirring. The resulting solution was incubated at 80 °C for 2 h for the assembly of OVA-Rha-GLA hydrogels. Meanwhile, OVA, OVA-Rha, OVA-LA, and OVA-GLA hydrogel systems were constructed.

2.6. Characterization of OVA-Rha-GLA solutions

2.6.1. Three-dimensional fluorescence spectroscopy

The interaction among OVA, Rha, GA, and LA was analyzed by spectrofluorometer (Fluoromax-4, HORIBA Scientific). The OVA concentration of all hydrogels was kept at 1.0 mg/mL. OVA, OVA-Rha, OVA-Rha-GLA, and OVA-Rha-GA were used for analysis. The excitation wavelength was fixed at 300 nm, while the emission wavelength was set at 300 nm to 600 nm. To describe the three-dimensional fluorescence spectrum, the ranges of excitation wavelength and emission wavelength were set as 250–600 nm and 300–600 nm, respectively. In the test, 10 nm acted as the excitation and emission slit widths.

2.6.2. Circular dichroism (CD)

The CD spectrum of OVA at a concentration of 0.5 mg/mL in different samples was recorded using a circular dichroism spectrometer (Chirascan qCD, Applied Photophysics). The wavelength range was 190 nm to 260 nm, and the bandwidth was fixed at 1 nm (Liu et al., 2021).

2.7. Characterization of OVA-Rha-GLA protein hydrogels

2.7.1. SEM, FTIR, and X-ray diffraction

The surface pore diameter and morphology of lyophilized hydrogels were observed using electron microscopy (SEM, FEI Quanta 400 FEG, American FEI) at 5.0 KV. FTIR was done to demonstrate the characteristic peak displacements and interaction details among different components of different hydrogels as previously described. Peak fit 4.12 (San Jose, CA) software was then donated to analyze the percentage and distribution of protein secondary structure within the range of 1500–1600 cm⁻¹. X-ray diffraction was performed to confirm the existence form of LA in different hydrogels at the 5–60°2θ range.

2.7.2. Rheological behaviors

The rheological characteristics of OVA-Rha-GLA and other hydrogels were investigated by Modular Compact Rheometer (MCR 302e). The plate-plate geometry was 20 mm with a gap of 1 mm. Each hydrogel sample was subjected to frequency sweeps and stress amplitude sweeps tests in triplicate. Frequency sweeps test was conducted under a continuous stress range of 0.1 to 100 Hz with a constant strain of 1 % at 25 °C. Stress amplitude sweeps were done in the range from 0.1 to 100 Pa with a fixed frequency of 1 rad s⁻¹ at room temperature (Wang et al., 2024c).

2.7.3. Molecular docking

Molecular docking was conducted to shed light on the interaction mode and strength between OVA and GA, LA, or Rha using Discovery Studio software. The structures of GA, LA, and Rha were downloaded from PubChem and optimized in the Clean Geometry and Full Minimization modules under the CHARMm force field using Discovery Studio. Meanwhile, the Human Protein Data Bank was used to download the structure of OVA (1OVA). The poly-configuration of OVA was removed, and incomplete amino acid residues were added and hydrogenated. GA, LA, and Rha were taken as ligand molecules, and OVA as receptor molecules for simulation in the receptor-ligand interaction module and input protein molecule module. After harboring the optimal 2D and 3D conformations of the receptor ligands, the binding energy and matching scores were calculated.

2.7.4. Molecular dynamics simulation

The 3D structures of GA, LA, Rha, and water molecules were generated via Openbabel 3.0.0 software. OVA-GLA and OVA-rhamnose-GLA systems were constructed using Packmol 18.169 in a box size of 60*60*90, among which the OVA was placed in the center of the box. OVA-GLA hydrogels consisted of 1 molecule of OVA, 40 molecules of GA, 8 molecules of LA, and 1000 molecules of water. 250 molecules of Rha were added to OVA-GLA to form OVA-Rha-GLA hydrogels. The small molecule GAFF force field was next generated using Sobtop 1.0, while the Amber99sb-ILDN force field was used for the protein. Subsequently, the system was minimized using Gromacs 2022.5. The number of steps was set to 5000, while emtol was set to 1000, nstlist set to 1, cutoff-scheme set to Verlet, coulombtype set to PME, and rcoulomb set to 1.2. Next, the systems were subjected to NVT equilibration for 100 ps, during which dt was set to 2 fs, thermostat was set to V- rescale, system temperature was set to 298 K, coulombtype set to PME, and rcoulomb set to 1.2. Lastly, 100 ns of NPT equilibration was performed, with the cutoff scheme set to Verlet, coulombtype set to PME, tcoupl set to V-rescale, and voltage controller set to Berendsen. The kinetic results were analyzed using the MSD module of Gromacs to calculate the diffusion coefficient. Additionally, the energy module was conducted to calculate the binding energy (including Coul-SR and LJ-SR).

2.8. In vitro Antibacterial Assay

The antibacterial capabilities of GLA, OVA-GLA, and OVA-Rha-GLA hydrogels were explored by the optical density values (OD₆₀₀) and agar plate method. Escherichia coli (E. coli) was regarded as the representative bacteria. The activity of OVA-Rha-GLA in inhibiting biofilm formation was also investigated according to our previous literature (Wang et al., 2024c).

2.9. In vitro release test

The in vitro release experiments were performed via Franz diffusion cells (Tianjin Jingtuo Instrument Technology Co., Ltd., China). The semi-permeable membrane was fixed between donor chambers and receiving chambers, in which phosphate buffer saline (PBS) at pH 7.2/ PEG 400 (V/V = 70/30) was donated as receiving medium. The receiving medium was immersed in the water bath at 32 °C and 350 rpm stirring. 0.3 g of hydrogels were evenly placed on the donor chambers, respectively, and each sample was performed in quadruplicate. 1 mL of receiving medium was drawn out at predetermined intervals (0.5, 1, 2, 3, 4, 6, 8, 10, 12, 24, 36, and 48 h) and replenished with 1 mL volume of preheated receiving medium. The concentration of LA at different intervals was determined by high-performance liquid chromatography (HPLC, Agilent 1260, USA) equipped with a C18 column (5 μm, 4.6 × 250 mm) and a DAD detector. The mobile phase consisted of acetonitrile, methanol, and 0.1 % phosphoric acid aqueous solution (20:45:35, V/V/V) with a flow rate of 1 mL/min. The detection wavelength was set at 300 nm.

2.10. Anti-inflammatory and anti-ROS assays, migration of RAW264.7 Cells

First, the cytotoxicity of different hydrogels to RAW264.7 macrophage cells was assessed by CCK8 methods at the concentration ranges of 6.25–100 μg/mL for LA. The gene expression of major inflammatory factors (IL-6, IL-1β, COX-2, TNF-α, and MCP-1) was studied using q-PCR. In short, RAW264.7 macrophages were treated with different hydrogel solutions and LPS (E. coli-derived) for 24 h, and cells were collected for reverse transcription polymerase chain reaction (RT-PCR) analysis(Li et al., 2022a). Total cellular mRNA was extracted with Trizol lysate, and the concentration of mRNA was determined using a spectrophotometer (Thermo Scientific, NanoDrop 2000, America). The RNA was transcribed into cDNA using PrimeScript RT Master Mix (Takara, RR036A, Japan), and then subjected to DNA amplification by TB Green Premix Ex Taq (Takara, RR420A, Japan). PCR was carried out using the RealTime PCR detection system (BIOER, FQD-96 A, China). The primers (Sangon Biotech, China) for PCR are listed in Table S1. Furthermore, the ROS content in the cell supernatant was measured using the ELISA method according to the manufacturer's instructions. For ROS staining, CellROX-green dye was added to a complete medium for 30 min at 37 °C.

2.11. Bioavailability and in vivo acute anti-inflammatory study

Twenty-five male Sprague-Dawley (SD) rats weighing 180–220 g were purchased from the Laboratory Animal Center of Southern Medical University. After 3 days of acclimatization, the rats were randomly divided into the Control group, LPS group, OVA-GLA group, OVA-Rha-LA group, and OVA-Rha-GLA group (n = 5). The rats were intraperitoneally injected with LPS (E. coli-derived) at a dose of 2.0 mg/kg in addition to the control group. After 12 h, the rats were fasted for 12 h before administration but were allowed to drink freely. The hydrogels were administered by oral gavage at a dose equivalent to 25 mg LA/kg body weight after another 12 h. After that, the blood of the rat retro-ocular venous (0.3 mL) was collected at predetermined intervals (0.25, 0.5, 1.0, 1.5, 2.0, 3.0, 4.0, 6.0, 8.0, 10, and 24 h) under anesthesia. The rats were anaesthetized with a dose of pentobarbital sodium (200 mg/kg) and euthanized by cervical dislocation after another 24 h. The blood was collected for Elias assays of IL-1β, INOS, TNF-α, ARG1, and IL-10 according to the manufacturer's instructions. The timeline of LPS injection and dosing was given in Fig. 9a. Subsequently, the plasma samples were centrifuged and analyzed by LC-MS/MS (Table S2 and Table S3), and the pharmacokinetic parameters were calculated using DAS 2.0 software.

Fig. 9.

Mean plasma concentration-time profiles and anti-inflammatory effects of OVA-Rha-GLA and other hydrogels in LPS-induced acute inflammation rat model. (a) The timeline of LPS injection and dosing of the animal experiments; (b) Mean plasma concentration-time profiles within 24 h; (c) Determination of TNF-α, (d) IL-1β, (e) INOS, (f) ARG1, and (g) IL-10 levels in the plasma after treated with different hydrogels. (Bar graphs represent mean ± SD, n = 5, ***p < 0.001, ****p < 0.0001 versus Control; ^##p < 0.01, ^###p < 0.001, ^####p < 0.0001 versus LPS; ^&p < 0.01, ^&&&p < 0.001 versus OVA-LA; ^$p < 0.01, versus OVA-GLA).

2.12. Statistical analysis

The statistical data were processed via GraphPad Prism version 8 or Origin 2021. All experiments were repeated three times or over three times, and the data produced were expressed as mean ± standard deviation (SD). For the comparison of differences between multiple groups, ordinary one-way analysis of variance (ANOVA) or two-way ANOVA with Dunnett's multiple comparisons was used. P < 0.05 was regarded as statistically significant.

3. Results and discussion

3.1. The characterization and self-assembly properties of GLA CBGNs

To improve the solubility and dispersion of LA in the supramolecular hydrogels, licorice-derived GA and LA were co-assembled into natural GLA BCGNs. The solubility of LA was significantly boosted ascribing to the solubilization and surfactant-like effect of GA. GLA CBGNs were synthesized via the self-assembly of GA and LA in aqueous solution, and the schematic diagram of GLA CBGNs is depicted in Fig. 1a. The average particle diameters of GLA nanoparticles were 27.25 ± 1.06 nm in a uniform distribution, with a polydispersity index (PDI) value of 0.28 ± 0.06 and a potential of −25.14 ± 0.98 MV (Fig. 1b). TEM images illustrated that GLA possessed a mean particle dimension of approximately 25 nm with an unbroken spherical structure, which was aligned with dynamic light scattering measurements (Fig. 1c). Previous literature had elucidated the critical micelle concentration of amphipathic GA in the aqueous solution (Su et al., 2017; Wang et al., 2025; Zeng et al., 2022). In FTIR analysis, the characteristic peaks of LA completely disappeared after the assembly with GA, indicating their molecular form dispersed in the CBGNs (Fig. 1d). Moreover, no significant shift or peak area changes of the -OH and C O characteristic peaks in GA were observed, suggesting that π-π conjugation and weak hydrogen bond interaction were the primary intermolecular strength between GA and LA, which encouraged the self-assembly of GLA CBGNs. Consistent with FTIR analysis, X-ray diffraction and Raman spectrum also observed the disappearance of LA absorption peaks upon the co-assembly of GA and LA, further underscoring their assembly efficiency (Fig. 1e and f). Moreover, the particle size and PDI of GLA CBGNs remained stable throughout 180 days of storage at 4 °C and 25 °C (Fig. S1 and S2).

Fig. 1.

Preparation and characterization of GLA CBGNs. (a) Schematic diagram of the self-assembly of GLA CBGNs; (b) The particle size, zeta potential, PDI, and size distribution of GLA CBGNs; (c) TEM pictures of GLA CBGNs (Bar = 50 nm); (d) FTIR spectrum, (e) X-ray diffraction and (f) Raman curves of GA, LA, physical mixture and GLA CBGNs; (g) Structural changes of the GLA CBGNs over 50 ns in the water environment during simulations (Green represents GA, cyan represents LA molecules). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Subsequently, MDS was conducted to shed light on the process of self-assembly of LA with GA. At 20 ns, the GLA system began to self-assemble into several nanoclusters in the simulated box system filled with water molecules. With the progress of simulation, GA and LA accumulated continuously under the interaction, forming a stable nanofiber cluster structure. After 50 ns, the GLA system self-assembled into sole fibrous nanoclusters (Fig. 1g). As nanofiber self-assembly continued, the solvent-accessible surface area (SASA) reduced, suggesting the formation of a tighter nanofiber cluster. The SASA value reduced significantly initially, then stabilized after 40 ns, indicating GLA could self-assemble into tighter nano-fiber clusters, reducing exposure to the solvent. Moreover, it was found that GA and LA primarily formed π-π stacking structures with weak hydrogen bonding interaction, which was similar to the FTIR results.

3.2. The ROS-scavenging efficacy of GLA

In an acute inflammation, a large number of ROS are generated, including O₂^•−, •OH, and H₂O₂ (Picouet et al., 2016; Sato et al., 2010). The inflammatory response triggers ROS production, which further amplifies inflammation and leads to tissue damage (Butenko et al., n.d.). Therefore, enhancing antioxidant defense is a critical strategy for treating inflammation-related diseases, especially acute inflammations. Idea pharmaceutical products or function foods should possess excellent content of superoxide dismutase (SOD)-like activity, which convert toxic O₂^•− to H₂O₂, rendering it potential antioxidant capacity. Furthermore, excessive inflammatory activation and ROS production are mediated by H₂O₂, which is a typical inflammatory molecule (Hu et al., 2024c). Therefore, it is vital to disproportionate the toxic O₂^•− cascade into harmless oxygen and water. The SOD-like activity of GLA was investigated by monitoring the capacity of GLA to remove O₂^•− from the xanthine/xanthine oxidase mixture. GLA exhibited remarkable SOD-like activity, which enhanced with increasing concentration, and the O₂^•− inhibition rate reached 44 % at 200 μg/mL (Fig. 2a and d). GLA then acted on H₂O₂, converting it into safe H₂O and O₂ in high efficiency, which generated 5 mg/L O₂ at 200 μg/mL (Fig. 2b and e). Interestingly, GLA scavenging O₂^•− and H₂O₂ raised within 300 s, which possessed great capacity to facilitate the whole cascade process of O₂^•−- H₂O₂- O₂ (Fig. 2c and f)_.

Fig. 2.

ROS-integrated catalysis and scavenging capabilities of GLA. In vitro kinetic curves of (a) SOD-like activity, (b) CAT activity, and (c) integrated cascade catalysis for GLA at 200 μg/mL; In vitro (d) SOD-like activity, (e) CAT activity, and (f) integrated cascade catalysis of GLA at various concentrations; (g) ABTS, (h) DPPH, and (i) OH radicals scavenging rate of GLA and LA solution at different concentrations. (n = 3, Data are presented as mean values ± SD).

Although O₂^•− occupies the majority of ROS, several other free radicals also attack biomolecules. Therefore, we evaluated the capacity of GLA to scavenge other radicals such as ABTS and DPPH. The results showcased that GLA harbored remarkable radical scavenging capacity with concentration-dependent kinetics. After incubation, the inhibition rate ascended with the increase of GLA concentration, verifying its free radical-scavenging ability (Fig. 2h). At the concentration of 200 μg/mL, GLA was capable of scavenging 57.3 % of the free radicals (0.4 mM DPPH) within 500 s (Fig. 2i). In addition, GLA effectively scavenged -OH in a concentration-dependent manner through the reaction with 3,3′,5,5′-tetramethylbenzidine (TMB) (Fig. 2j). In contrast, free LA exhibited a significantly lower capacity to expel ABTS, DPPH, and -OH in comparison with GLA. Overall, GLA systems exhibited great performance in scavenging different types of free radicals, which were promising complexes for radical scavengers in the management of acute inflammations.

3.3. Three-dimensional fluorescence spectra and CD spectra of OVA-Rha-GLA hydrogel solutions

OVA solution was difficult to assemble into protein hydrogels in their natural state. The components ratio that likely impacted the hydrogels' crosslinking and rigidity was first investigated. We observed that both OVA and OVA-Rha colloidal solutions were white, transparent, and clear. Upon the dropwise addition of GLA into the OVA solution, the mixtures became turbid owing to the attraction of the negatively charged GLA and the positively charged OVA (Fig. 3a). However, OVA-Rha-GLA solution became more transparent in comparison to OVA-GLA, benefiting from the charge neutralization and clarifying effect of Rha. Remarkably, a higher Rha amount encouraged the formation of a clearer OVA-Rha-GLA solution (Fig. 3a). TEM images illustrated that GLA still exhibited a spherical structure in OVA-Rha-GLA mixture with a mean particle dimension of approximately 30 nm (Fig. 3b). This suggested that GLA was dispersed in nano form into OVA-Rha-GLA solution, which was beneficial to maintain the molecular form and distribution of LA in the OVA protein skeleton. Nevertheless, the amount of Rha posed a potential impact on the gelation time, rheological properties, and rigidity of the hydrogels. Therefore, the mass ratio of Rha to OVA was confirmed as 1:2 in the subsequent analysis.

Fig. 3.

Preparation and characterization of OVA-Rha-GLA and other hydrogel solutions. (a) Schematic diagram of the fabrication of OVA-Rha-GLA solution and OVA-Rha-GLA hydrogels; (b) TEM pictures of GLA CBGNs after incorporation into OVA-Rha solution (Bar = 50 nm); (c) Three-dimensional fluorescence spectra of OVA, OVA-Rha, OVA-GLA, and OVA-Rha-GLA solution; (d) Circular dichroism spectra of different hydrogel solutions (The red arrows represent amino acid residues of OVA). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Three-dimensional fluorescence spectra could reveal the changes in protein conformations. The three-dimensional fluorescence spectra of OVA in different solutions are illustrated in Fig. 3c. The addition of Rha significantly changed the three-dimensional fluorescence spectra of OVA, appearing a new fluorescence peak of an amino acid residue (Fig. 3c). These emphasized that a strong intermolecular interaction force was formed between OVA and Rha, which altered the secondary structure of OVA, probably leading to the formation of supramolecular structures. It has been reported that Rha encouraged the self-assembly of supramolecular structures with other polymers through hydrogen bonding interaction and π-π conjugation (Song et al., 2024a; Song et al., 2024b; Zhao et al., 2014). Nevertheless, after the incorporation of GLA CBGNs, the secondary structure of OVA did not exhibit significant changes in comparison with OVA. Conversely, a joint application of GLA and Rha resulted in the appearance of fluorescence peaks of multiple amino acid residues in OVA, which indicated that GA complexes and Rha further altered the secondary structure of OVA, achieving a more stable supramolecular structure. To further understand the alternations of secondary structures, the CD spectroscopy of different solutions was recorded. As illustrated in Fig. 3d, the CD spectrum of OVA had a broad negative peak at 221 nm. GLA induced slight changes to the width and displacement of the OVA characteristic peak. In contrast, in OVA-Rha and OVA-Rha-GLA solutions, the peak moved to 224 nm and became weaker and narrower, verifying the contributory effect of Rha in changing the secondary conformations of OVA.

3.4. The assembly behaviors and mechanisms of supramolecular spatial protein hydrogels

Next, we unveiled the assembly mechanisms of different hydrogel networks. All hydrogel solutions were immediately heated at 80 °C to record the gelation time of the hydrogels. Interestingly, the gelation time of OVA-Rha (41.3 min) was significantly longer than pure OVA hydrogels (32.5 min). Similarly, OVA-GLA (34.0 min) required significantly less time for the assembly of hydrogels compared to OVA-Rha-GLA hydrogels (44.7 min) (Fig. 4a). These suggested that the altered secondary structures induced by Rha disturbed the normal assembly of OVA, which reshaped the gelation of the hydrogels, thus significantly prolonging the gelation time. All hydrogel solutions were then incubated at 80 °C for 120 min to construct supramolecular steric networks.

Fig. 4.

Physical properties, microstructures, and assembly mechanisms of different hydrogels. (a) The gelation time of different hydrogels. (b) SEM pictures of the lyophilized hydrogels (scale bar = 50 μm); (d) Pore size of OVA, OVA-Rha, OVA-GLA, and OVA-Rha-GLA hydrogels (n = 50). (Bar graphs represent mean ± SD, **p < 0.01, ****p < 0.0001 versus OVA; ###p < 0.001, ####p < 0.0001 versus OVA-GLA; +++p < 0.01, versus OVA-Rha); (d) X-ray diffraction properties of LA, GLA, OVA-GLA, and OVA-Rha-GLA; (e) FTIR spectra of different hydrogels; (f) Secondary structure fractions analysis of the FTIR spectra ranging from 1500 to 1600 cm⁻¹ of OVA; (g) The 2D schematic interaction between OVA and Rha, LA, GA, displaying the lowest CDOCKER interaction energy.

The surface microstructure of the hydrogels was visualized by SEM. OVA hydrogels displayed a dense and compact structure with no apparent pores (Fig. 4b), suggesting sufficient assembly and gelation. The introduction of Rha facilitated the appearance of uniform pore size on the surface, shaping a loose supermolecule spatial structure for OVA-Rha. The surface of OVA-GLA hydrogel exhibited no significant difference compared to pure OVA. Similar with OVA-Rha, Rha also significantly enlarged the pore diameters and shaped the three-dimensional spatial structure for OVA-Rha-GLA hydrogels (Fig. 4b, c, and Fig. S3). These verified that Rha could weaken the dense crosslinking of OVA hydrogels but reinforced their spatial properties, which was conducive to drug release, underscoring its capability in the therapy of acute inflammation. We next measured the mean pore of OVA-Rha-GLA and OVA-Rha using ImageJ software and found that OVA-Rha-GLA demonstrated a significantly larger pore diameter and porosity than OVA-Rha (Fig. 4b, c, and Fig. S3). X-ray diffraction was conducted to confirm the drug's existence form of the hydrogels. It was observed that the LA characteristic diffraction peaks all disappeared in GLA, OVA-GLA, and OVA-Rha-GLA hydrogels (Fig. 4d), demonstrating that LA was molecularly dispersed into the hydrogels. This laid a foundation for the formation of intermolecular forces with other components.

Subsequently, FTIR was done to shed light on the intermolecular details among different components in the hydrogels (Wang et al., 2024c) (Fig. 4e). The introduction of Rha caused an apparent red shift of the -OH group (3281.21 cm⁻¹) and C O group (1626.41 cm⁻¹) of OVA protein. A significant -CH₂ stretching vibration peak of OVA (1047.93 cm⁻¹) was also observed upon the introduction of Rha. In general, a red-shift movement reflected lower energy and more active functional groups in the OVA. Remarkably, the -OH groups on the OVA protein became broader after the incorporation of Rha, which was indicative of the strong hydrogen bond interaction between OVA and Rha, exerting a pivotal effect on the clarification of the OVA-Rha-OVA solution. GLA also induced the vibration of the –OH group and C O group, but the contributory effect was less significant than Rha (Fig. 4e). Conversely, GLA demonstrated a similar effect to facilitate the stretching vibration of the -CH₂ group to that of Rha. The combined addition of GLA and Rha exerted a more significant disturbing effect on the vibration of the -CH₂ group than pure OVA and Rha, without an observable effect on the vibration of the -OH group. These results revealed that Rha and GLA synergistically encouraged the assembly of supermolecule spatial structures predominantly surrounded by the -OH, C O, and -CH₂ bonds on the OVA.

To further understand the detailed alterations of secondary structures in different OVA hydrogel systems, the FTIR spectra ranging from 1500 to 1600 cm⁻¹ of OVA were determined (Fig. 4f). Secondary structure fractions analysis exhibited that OVA possessed 30.8 % α-helix, 28.7 % β-sheet, 23.6 % β-turn, and 17.9 % random coil in its secondary structure composition. Upon the addition of Rha, the proportions of β-sheet and random coil were raised to 31.2 % and 19.5 %, whereas the contents of α-helix and β-turn were reduced to 26.6 % and 26.7 %, respectively. It underscored that the 3D conformation of OVA was transited from an ordered condition to a disordered status. Conversely, pure GLA did not induce the apparent changes of OVA peaks, with no observable secondary structure changes. The combined application of GLA and Rha possessed the strongest capacity to alter the secondary structure of OVA with apparent conformational transformation. As a result, it led to a composition of 25.4 % α-helix, 33.3 % β-sheet, 20.5 % β-turn, and 20.7 % random coil, which facilitated the formation of supermolecule networks with apparent pores. The findings were in accord with the results of three-dimensional fluorescence spectra and CD spectra.

3.5. Molecular docking

Afterward, we gained insights into the miscibility and interaction performance between OVA and Rha, GA, or LA using molecular docking, respectively. The positional arrangement with the lowest energy fraction was defined as the most stable molecular conformation (Zhong et al., 2022), which represented the binding interactions of OVA with other compounds. Notably, the interaction energy of OVA-Rha (−30.48 Kcal/mol) was lower than that of OVA-GA (−38.24 Kcal/mol) and OVA-LA (−45.91 Kcal/mol), indicating a better compatibility (Fig. 4g). The results emphasized that Rha played a pivotal role in stabilizing OVA hydrogel skeletons, which further increased the clarity of OVA-Rha-GLA ternary systems. The 2D schematic interactions and 3D docking conformation results between OVA and other compounds are displayed in Fig. 4g and Fig. S4. The salt bridges and π-π conjugation were mainly involved in OVA-LA interaction, while the -OH and C O bonds were not involved in the hydrogen bond interaction with OVA residues, thereby resulting in their poor miscibility. In contrast, the strong hydrogen bond interactions predominantly dominated the OVA-Rha interaction. Rha predominantly used -OH and C O bonds as hands to specifically contact AGR117, LEU156, etc., of the keratin 8K8H for binding. These suggested -OH and C O bonds of Rha also played a vital role in the assembly of supermolecule spatial structures.

Altogether, the introduction of Rha resulted in secondary conformational transformations of OVA, decreasing the binding energy of the hydrogel systems, which shaped supermolecule steric OVA-Rha-GLA structures predominantly surrounded by -OH and C O bond on the Rha and the -OH, C O, and -CH₂ bonds on the OVA.

3.6. Rheological performance and mechanical properties of different hydrogels

To testify the above results, the rheological behaviors and mechanical properties of different hydrogels were determined. In the frequency sweeps test, the elasticity modulus (G') was higher than the loss modulus (G") for all hydrogels, indicating their solid-like behaviors (Fig. 5a). The G' and G" values all exhibited an increasing tendency in the range of linear frequency sweep values. As expected, the G' and G" values of OVA-Rha-GLA and OVA-Rha were significantly lower than OVA-GLA and OVA (Fig. 5a), indicating a weaker mechanical property, which corresponded to the results of SEM observations. OVA-GLA and OVA demonstrated close G' and G" values, however, OVA-Rha-GLA possessed higher G' and G" values than OVA-Rha, suggesting that GLA also played an essential role in the hydrogels' crosslinking. In the strain sweep analysis, the critical point of four kinds of hydrogels was ranked as OVA-Rha-GLA ≈ OVA-GLA > OVA-Rha > OVA (Fig. 5b). A higher critical point is indicative of a more flexible hydrogels with higher critical tensile elongation (Wang et al., 2023a; Wang et al., 2024c). These suggested that the weak crosslink density and supermolecule spatial structure rendered OVA-Rha-GLA networks softer and flexible, contributing to a higher critical tensile elongation. In addition, all hydrogels demonstrated a pronounced strain-dependent viscoelastic response. G' was independent of strain and was consistently larger than G" in the linear viscoelastic region (0.1–10 %), exhibiting the properties of a viscoelastic material. Subsequently, we measured the mechanical properties, including hardness and stiffness of different hydrogels. Consistent with the rheological results, the incorporation of Rha resulted in a significant reduction of hardness (Fig. 5c) and stiffness (Fig. 5d) values for OVA-Rha-GLA and OVA-Rha. Furthermore, pure OVA and OVA-GLA displayed similar hardness and stiffness. The water holding capacity (WHC) reflected the mechanical properties and stability of hydrogels, which were next investigated. Intriguingly, Rha-based hydrogels also harbored a significantly lower WHC than other hydrogels (Fig. 5e), indicating their lower mechanical property.

Fig. 5.

Rheological performance and mechanical properties, as well as antibacterial properties of different hydrogels. (a) Frequency sweep measurement of OVA, OVA-Rha, OVA-GLA, and OVA-Rha-GLA hydrogels (Kept at 1 % strain); (b) Strain sweep determination of the hydrogels (maintained at 1 rad s⁻¹); (c) The water holding capacity of different hydrogels; (d) Hardness and springiness of different hydrogels. (Bar graphs represent mean ± SD, n = 3, *p < 0.05, **p < 0.01 versus OVA; ^#p < 0.05, ^##p < 0.01 versus OVA-GLA). (d) Photographs of E. coli bacteria clones on the surface of different hydrogels; The corresponding (f) OD600 growth, (g) bacterial survival rate of E.coli after treated with different hydrogels; (h) The E.coli biofilm clearance rate of different hydrogels. (n = 3, ****p < 0.01 versus control).

3.7. In vitro release behaviors of different hydrogels

In the treatment of acute inflammation, the rapid release of the bioactive components from the matrix encourages the rapid reduction of pro-inflammatory factors and the corresponding cascade of reactions, which helps to curb the deterioration of the disease. Therefore, the regulation of drug release by hydrogel networks is a critical step in the management of acute inflammation (Yao et al., 2017). An in-depth understanding of drug diffusion behaviors in the hydrogel matrix could predict drug absorption and bioavailability in the body (Palem et al., 2021). At present, most literature on drug release characteristics in edible hydrogels has ignored the role of water. The water in the hydrogels is classified into non-freezable water, intermediate water, and free water according to the interaction between water molecules and functional groups on the polymer chain. When water molecules are in different states, they possess different flow properties and crystallization capacity. Polymer chains in hydrogels are prone to trapping nearby water molecules to form non-freezable water via strong interactions such as hydrogen bonding, which possess low mobility and cannot crystallize even at very low temperatures. Water molecules far from the polymer chain are defined as free water, which has high mobility and does not interact directly with the hydrogel chains (Song et al., 2024b; Zhou et al., 2019). We observed that the proportion of free water in OVA-Rha-GLA networks was lower than that in OVA-GLA (Fig. 6a), as evidenced by the results of molecular dynamics simulation (Fig. 7a).

Fig. 6.

The intermolecular interactions among different components controlling LA release in the hydrogels. (a) The mode and strength of intermolecular interactions in different hydrogels; (b) The release curves of different hydrogels during 48 h. (Bar graphs represent mean ± SD, n = 4).

Fig. 7.

MDS analysis of LA diffusion from different hydrogel systems. (a) The dynamic snapshots of LA diffusion into the OVA-GLA and OVA-Rha-GLA hydrogels network at 0, 10, 20, 40 and 100 ns (Green represents OVA chains, while purple represents LA, blue green represents GA, Yellow represents Rha and red represents water molecules); (b) MSD analysis of LA from different hydrogels within 90 ns NPT analysis; (c) The binding energy of different systems; (d) The interaction energy between GA, LA, Rha and OVA respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Herein, the in vitro drug release test unveiled the drug diffusion properties and intermolecular interaction behaviors among different components of the hydrogel systems. It was reported that intermolecular force was a pivotal factor that affect the drug release (Wang et al., 2022; Wang et al., 2024c). The interaction model and strength among different components in the hydrogels are illustrated in Fig. 6a. In OVA-LA systems, LA mainly formed hydrophobic-hydrophobic interactions and weak water bridge hydrogen bonding with OVA polymers. Weak hydrogen bonding between GLA and OVA, and moderate water bridge hydrogen bonding were the interaction forces that predominantly drove the self-assembly of OVA-GLA. In contrast, strong water bridge hydrogen bonding and hydrogen bonding among Rha, OVA, and water were the primary interaction strengths that existed in OVA-Rha-GLA hydrogels. As described in molecular docking and FTIR analysis, -OH and C O bonds on the Rha and the -OH, C O, and -CH₂ bonds on the OVA shaped the supermolecule steric OVA-Rha-GLA structures. Therefore, the OVA-Rha-GLA network exhibited a larger volume and spatial pore than OVA-GLA and OVA-LA hydrogels (Fig. 6a).

The dynamic concentration kinetics of LA from different hydrogels were detected within 48 h (Fig. 6b). It was apparent that LA release behaviors from OVA-Rha-GLA demonstrated a first-order release equation curve, whereas it exhibited zero-order release equation properties in OVA-1/2Rha-GLA, OVA-GLA, and OVA-LA systems. The LA release percent reached 23.80 % within 24 h, which achieved only 39.6 % after 48 h in OVA-LA hydrogel. Intriguingly, LA displayed a lower release capacity in OVA-GLA than that in the OVA-LA (Fig. 6b). As demonstrated before, LA predominantly formed hydrophobic-hydrophobic interaction (Fig. 4) and weak water-bridge hydrogen bonding (Fig. 6a) with OVA chains, which controlled LA sustained release. However, when LA was co-assembled with GA, GA possessed a higher compatibility and binding energy with OVA chains than LA (Fig. 4), thereby leading to a lower LA release percent in OVA-GLA within 48 h.

Compared with OVA-GLA, the introduction of a small amount of Rha displayed no remarkable enhancement to the release of LA. Interestingly, the augment of Rha amount significantly raised the release percent of LA, facilitating 45 % and 61.28 % LA release from OVA-Rha-GLA hydrogels within 24 h and 48 h respectively (Fig. 6b). On the one hand, OVA-Rha-GLA possessed weaker viscoelasticity and mechanical rigidity than OVA-GLA owing to its supramolecular spatial structure induced by Rha (Fig. 5). On the other hand, OVA exhibited strong binding energy (Fig. 4), hydrogen bonding and water-bridge hydrogen bonding (Fig. 6a) with Rha and water molecules. The introduction of Rha could occupy the interaction sites between OVA chains and GLA CBGNs, resulting in the free form of GLA complexes in the OVA-Rha-GLA networks. Based on the above, a higher drug release capacity was successfully achieved in OVA-Rha-GLA hydrogels than other protein hydrogels. The rapid release of LA was expected to provide effective concentration and enhance bioavailability of LA in vivo, thereby improving its anti-inflammatory and anti-ROS effects in the acute inflammation treatments.

3.8. Molecular dynamics simulation

To further unravel the interionic interaction mechanisms and drug diffusion properties of different hydrogels at the molecular level, we construct amorphous cell systems comprising OVA, Rha, GA, LA, and water. As illustrated above, GA and LA could self-assemble into fibrous clusters in the aqueous solution after a 50 ns simulation (Fig. 1g). OVA-GLA and OVA-Rha-GLA were selected for analysis. The 3D conformations of different systems at 0, 10, 20, 40, and 100 ns are displayed in Fig. 7a. In both OVA-GLA and OVA-Rha-GLA systems, it was seen that the free water diffused faster as the increasing of time. Remarkably, the proportion of free water in the OVA-Rha-GLA systems was significantly lower than that in the OVA-GLA, verifying that Rha formed a strong water bridge bonding with OVA and non-freezable water. As a result, the OVA protein chains were more ordered, compact, and stable in the OVA-Rha-GLA networks than in OVA-GLA (Fig. 7a). Next, the binding energies among different compounds in the simulation process were calculated. Interestingly, OVA-Rha-GLA possessed a binding energy of 2.87*10⁷ Kcal/mol, which was lower than OVA-GLA (3.66*10⁷ Kcal/mol) (Fig. 7c). A higher binding energy generally manifested as stronger intermolecular force, mechanical strength and inherent entangle of the hydrogels (Hu et al., 2024b; Wang et al., 2023b). These reinforced the viewpoints that the conformational transformation caused by Rha loosened the hydrogel networks, reducing the binding energy of the hydrogel systems, which was in accord with the results of rheological test (Fig. 5a) and FTIR analysis (Fig. 4e and f). Subsequently, the MSD of LA in different systems was calculated using the MSD module of Gromacs. As expected, the diffusion rate of LA in OVA-Rha-GLA was higher than that in OVA-GLA with 90 ns (Fig. 7b), further elucidating that Rha induced the hydrogel networks to a laxer state, which was conducive to LA release. The results were aligned with the results of in vitro drug release (Fig. 6).

Next, we analyzed the interaction strength between different constituents in the hydrogels. Interestingly, both GA and LA exhibited a higher interaction force with OVA in OVA-GLA than in OVA-Rha-GLA (Fig. 7d). Moreover, Rha possessed a higher binding energy with OVA proteins than that of GA or LA in the Rha-based supermolecule structures, which corresponded to the results of molecular docking (Fig. 4g). These revealed that Rha occupied the interaction sites between OVA and GLA or LA, contributing to a free form of GLA or LA in the water environment. Consequently, GA possessed higher interaction force with LA after the introduction of Rha (Fig. 7d). Based on the above, we hypothesized that LA in OVA-Rha-GLA was more easily absorbed by the organisms and possessed higher anti-inflammatory bioactivities and oral bioavailability in the acute inflammation rat models.

3.9. The antibacterial capacity of hydrogels

LPS, a major component of the cell wall of Gram-negative bacteria, triggers a strong inflammatory cascade by activating the immune system's toll-like receptor 4, eventually resulting in the development of acute inflammation (Wang et al., 2024a). Herein, the antibacterial effects of different hydrogels were observed by using LPS-derived E.coli, further to evaluate their anti-inflammatory bioactivities. Compared with the control, OVA-Rha-GLA and OVA-GLA displayed significantly decreased OD₆₀₀ values after incubation with E. coli for 24 h. However, the inhibitory effect was significantly reduced in the GLA group (Fig. 5e-g). Moreover, OVA-Rha-GLA was superior to clearing the biofilm of E.coli than others at different concentrations (Fig. 5h). These results indicated that OVA-Rha-GLA hydrogels were efficient in treating the acute inflammation caused by E.coli.

3.10. Anti-inflammatory and ROS-scavenging effects at the cellular level

The LPS-induced macrophage cell (RAW264.7) model was donated to assess the anti-inflammatory activity of different hydrogels. The cytocompatibility of different hydrogels were first investigated using CCK8 test. The results showcased that OVA displayed remarkable safety to the RAW264.7 cells, while OVA-LA were toxic to the cells when the LA concentration was exceeded 50 μg/mL. However, the treatment of Rha and GLA enhanced the biocompatibility, exhibiting a high cell viability lower than 50 μg/mL of LA (Fig. S5).

Decreasing the secretion of pro-inflammatory factors such as TNF-α, IL-1β as well as IL-6 is effective strategy for acute inflammation intervention. The expression of key inflammatory factors (IL-6, TNF-α, IL-1β, MCP-1, COX-2) was evaluated to confirm the anti-inflammatory activity of OVA-Rha-GLA and other hydrogels at the cellular level. LPS significantly raised these key inflammatory factors, and OVA-Rha-GLA demonstrated the highest capability to reduce the expression of IL-6, TNF-α, IL-1β, MCP-1, and COX-2 (Fig. 8a-8e). In contrast, OVA and OVA-Rha exhibited an inability to lower the levels of these inflammatory factors at the current concentration. OVA-LA could suppress part of the key inflammatory factors (Fig. 8a, c, and d), but was significantly less effective than OVA-Rha-GLA. Excessive inflammatory activation leads to ROS accumulation (Butenko et al.), therefore, it was urgent to dispel the production of ROS. Consistent with the content of the anti-inflammatory test, OVA-Rha-GLA also displayed a higher capacity to decrease the production of ROS than other hydrogel systems (Fig. 8f). To visualize the impact of different hydrogels on ROS production, we treated cells with different hydrogels and measured ROS levels after LPS stimulation. ROS production was also significantly reduced in OVA-Rha-GLA compared with LPS and OVA-LA (Fig. 8g), in accord with the ELISA results. Altogether, OVA-Rha-GLA displayed excellent anti-inflammatory and ROS-scavenging capabilities in LPS-induced Inflammatory cell model.

Fig. 8.

The anti-inflammatory, ROS-scavenging effects and cell migration promotion of OVA-Rha-GLA and other hydrogels. The expression levels of inflammatory factors of (a) IL-6, (b) MCP-1, (c) IL-1β, (d) TNF-α, and (e) COX-2 upon treatment of different hydrogels; (f) The production of ROS after treating different hydrogels; (g) The RAW264.7 cells were stimulated with 1 μg/ml LPS, followed by staining with Cell ROX-green to quantificationally detect ROS. Scale bar = 20 μm; (Bar graphs represent mean ± SD, n = 4, ****p < 0.0001 versus OVA; ^#p < 0.05, ^##p < 0.01, ^###p < 0.001, ^####p < 0.0001 versus LPS; ^&p < 0.05, ^&&p < 0.01 when compared with OVA-LA;) (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.11. Enhanced bioavailability and anti-inflammatory effects in vivo

An LPS-induced inflammatory rat model was established to evaluate the bioavailability and anti-inflammatory effects of LA in different hydrogel formulations. The poor water solubility of flavonoids poses a tremendous impact on cellular absorption and effective concentration, which is the key factor for their low bioavailability (Premathilaka et al., 2022b; Wang et al., 2024b). The mean plasma concentration-time curves of LPS-induced rats after oral administration of OVA-LA, OVA-GLA, and OVA-Rha-GLA are showcased in Fig. 9a. Moreover, the pharmacokinetic parameters of different hydrogels are provided in Table S4. LA harbored low C_max (1.314 ± 0.423 ng/mL) and AUC_0-t (18.417 ± 6.521 ng∙h/mL) values in OVA-LA hydrogels. After co-assembly with GA, the C_max (2.04 ± 0.849 ng/mL) and AUC_0-t (29.038 ± 12.743 ng∙h/mL) of LA in OVA-GLA were enhanced, elucidating that improving the drug solubility contributed to a higher bioavailability. Intriguingly, 326 % of the C_max and 287 % AUC_0-t of OVA-Rha-GLA were achieved compared with OVA-LA, suggesting that CBGNs-based supramolecular hydrogels exhibited a remarkable improvement in the oral bioavailability of LA. Furthermore, the T_max of OVA-Rha-GLA was shorter than OVA-GLA and OVA-LA, was indicative of a faster function (Fig. 9a). Thus, OVA-Rha-GLA demonstrated improved pharmacokinetic behaviors and bioavailability compared with the OVA-LA and OVA-GLA, which was beneficial from the enhanced water solubility and drug release rate of LA.

Next, we determined the concentrations of various pro-inflammatory factors (IL-1β, TNF-α, and INOS) and anti-inflammatory factors (ARG1 and IL-10) in the plasma using ELISA method after treated with different hydrogels (Fig. 9c–e). It was found that LPS injection induced the up-regulation of IL-1β, TNF-α, and INOS and the down-regulation of ARG1 and IL-10, indicating the successful establishment of an acute inflammation model. In terms of pro-inflammatory factors, OVA-Rha-GLA treatment could significantly lower the expression of IL-1β, TNF-α, and INOS in the plasma when compared to the LPS and OVA-LA groups. Moreover, OVA-Rha-GLA was superior to decreasing the levels of IL-1β and INOS than OVA-GLA (Fig. 9d and e). In terms of anti-inflammatory factors, OVA-Rha-GLA was capable in promoting the levels of ARG1 and IL-10 compared with the LPS group. OVA-GLA exhibited a certain capacity to enhance the expression of ARG1, but its capacities were lower than OVA-Rha-GLA (Fig. 9f and g). Collectively, supramolecular hydrogels and CBGNs together enhanced the anti-inflammatory effects and bioavailability of LA in the acute inflammation model.

4. Conclusion

In this paper, we proposed novel CBGN-embedded supramolecular protein hydrogels to enhance LA release for acute inflammation management. Licorice-derived GLA CBGNs harbored great capacity to scavenge various ROS, including O₂^•−, •OH, and H₂O₂. It was next incorporated into the OVA-Rha binary solutions and assembled into OVA-Rha-GLA supramolecular hydrogels. The hydrogels were formed via hydrogen bonding and hydrophobic-hydrophobic interaction mediated by -OH and C O bonds on the Rha and the -OH, C O, and -CH₂ bonds on the OVA. Notably, the introduction of Rha changed the secondary conformations of OVA, which shaped the three-dimensional supramolecular spatial structures of OVA-Rha-GLA hydrogels. A looser supramolecular network encouraged a faster LA release, which was beneficial for acute inflammation treatments. As a result, LA possessed higher oral bioavailability in OVA-Rha-GLA than other hydrogels that lacked BCGNs or Rha in acute inflammatory models. In LPS-induced inflammatory cellular and rat models, OVA-Rha-GLA possessed excellent capabilities of alleviating anti-inflammation and dispelling ROS. Overall, a joint application of supramolecular hydrogels and co-assembled nano-systems is an effective and promising system to improve the drug release, anti-inflammatory effect, and oral bioavailability of flavonoid ingredients, which exhibited remarkable promise in the therapy of acute inflammation.

CRediT authorship contribution statement

Zhuxian Wang: Writing – original draft, Software, Methodology, Investigation, Funding acquisition, Formal analysis, Conceptualization. Jun Liu: Validation, Methodology, Investigation. Yufan Wu: Visualization, Software, Methodology. Yamei Li: Software, Investigation. Hongxia Zhu: Supervision, Resources. Qiang Liu: Writing – review & editing, Validation, Supervision. Bin Yang: Writing – review & editing, Supervision, Project administration, Funding acquisition.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could influence the work reported in this paper.

Data availability

Data is provided within the manuscript or supplementary information files, and data will be made available on request.