Rat chondrocytes were isolated and cultured as described in a previous study (Wang et al., 2019; Xu et al., 2019). Sprague Dawley (SD) rat knee cartilage tissues were cut into small pieces and digested with 2 mg/ml collagenase type II (Sigma, United States) for 4–6 h. The undigested tissues were removed, and the isolated cells were seeded in DMEM/F12 (1:1) medium (Gibco, United States) containing 10% fetal bovine serum (FBS) (Gibco, United States). Cells from the second or third passages were used in cell experiments. Animal experiments for this study were approved by the Institutional Review Board (IRB) of Chongqing University.

Hesperetin (HPT) and dexamethasone (DEX) were purchased from Selleck (Selleck Chemicals, United States) and then dissolved in DMSO as a 10 mM stock solution. Rat chondrocytes were treated with HPT at different concentrations (1, 5, 10, 20, 50, and 100 μM); DMSO was used as a control. Meanwhile, exposure to 10 μM HPT was implemented at different time intervals (1, 2, 3, 4, and 5 days). Morphology caused by HPT at different concentrations (1, 5, 10, 20, 50, and 100 μM) was detected by microscopy (Olympus, Japan). In addition, 10 ng/ml of TNF-α alone (the TNF-α-treated group) or TNF-α plus HPT at 10 μM (TNF-α + HPT-treated group) or TNF-α plus DEX at 10 μM (TNF-α + DEX-treated group) was used for treating chondrocytes for 48 h (TNF-α and HPT or TNF-α and DEX were added in the same order). All the experiments were performed in triplicate independently.

Cell Counting Kit 8 (CCK-8) (Beyotime, China) was used to investigate cell viability and cytotoxicity. Chondrocytes were seeded at a density of 1,000 cells/well in a 96-well plate and tested for toxicity for 1–5 days with different concentrations of HPT. DMSO was used as a control. At designated time intervals, the cells were incubated with 10% CCK-8 reagent at 37°C for 2–4 h. After incubation, the corresponding absorbance was measured with a microplate reader at a wavelength of 490 nm. The data were analyzed with GraphPad 8.0. All the experiments were performed in triplicate independently.

Chondrocytes were treated with TNF-α alone or TNF-α plus HPT and then harvested for the qRT-PCR assay. DMSO was used as control. Each sample was harvested using TRIzol reagent, and RNA was isolated. Total RNA was reverse transcribed into cDNA using the RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific). Then, the cDNA was used as template to perform real-time polymerase chain reaction (PCR) on a StepOne Plus thermal cycler (Applied Biosystems). Primers were designed by NCBI primer blast and synthesized using Invitrogen. Data were analyzed according to the 2−ΔΔct method.

Rat chondrocytes were lysed by RIPA lysis buffer to extract the total protein which was separated by SDS-PAGE gel electrophoresis and transferred to PVDF (polyvinylidene fluoride) membrane (0.2 μm; BioRad). After 1 h blocking with 5% skimmed milk, the membrane was incubated with primary antibody ACAN (Abcam, ab36861), COL2A (Santa Cruz, sc-7763), IL-1β (Proteintech, 66,737-1-lg), IL-6 (Zen Bio, 347023), MMP-13 (Zen Bio, 511755), PTGS2 (Zen Bio, 383967), Bcl-2 (Zen Bio, 381702), CDK1 (Abcam,ab18), Cyclin D1 (Zen Bio, 382442), p-AMPKα1 (Zen Bio, 310044), AMPKα (Zen Bio, 380431), and GAPDH (Zen Bio, 384404) antibodies overnight at 4°C. After washing 3 times with 1×TBST, the membrane was incubated with horseradish peroxidase-labeled secondary antibody (goat anti-rabbit/mouse IgG, 1:5,000) at room temperature for 1 h. Finally, target protein was visualized by the ECL Western Detection Kit (Thermo Fisher Scientific, United States) and quantified by ImageJ software.

Flow cytometry was used to determine the effect of HPT on the chondrocyte cycle. After being treated with TNF-α alone or TNF-α plus HPT, the chondrocytes were harvested and washed, followed by fixation using 70% alcohol at 4°C overnight. After washing the cells with PBS, they were treated with ribonuclease RNase A and incubated with propidium iodide (PI) for the cell cycle analysis on a BD FACSVerse^TM flow cytometer (BD Bioscience, San Jose, CA).

RNA sequencing was performed on chondrocytes in different treatments to determine the changes in their mRNA expression profiles. RNA extraction kit (Thermo Fisher Scientific, Waltham, MA) was used to extract total RNA from each group of cells. Send the obtained RNA sample to BGI Co., Ltd., Shenzhen, China, and perform RNA-seq operation on BGISEQ-500. The R language program was used to further analyze the sequencing results to find differential genes in the chip expression profile. Then, use the R program to perform the gene ontology (GO) biological function enrichment analysis and genome encyclopedia (KEGG) signal pathway enrichment analysis. All samples were repeated three times.

The chondrocytes were seeded on coverslips coated with 0.01% poly L-lysine. After treatment, the cells were fixed in 4% paraformaldehyde for 30 min, followed by rinsing with PBS and permeating with 0.1% Triton X-100 on ice for 5 min. Then, the cells were blocked with a blocking solution at 37°C for 1 h and then incubated with the primary antibody p-AMPKα1 (Zen Bio, 310044) at 4°C overnight and with a fluorescently labeled secondary antibody (concentration 1:200) at 37°C for 1 h, respectively. After counterstaining with DAPI, the samples were imaged and captured under a fluorescence microscope (Olympus, Japan).

The animal experiments were carried out following the protocol approved by the Animal Care and Use Committee of Chongqing University. The OA model was achieved by anterior cruciate ligament transection (ACLT) in SD rats, as described in our previous study (Wang et al., 2020a; Xu et al., 2021). There are 24 rats randomly assigned to four groups: healthy, OA (normal saline group), DEX, and HPT groups. Four weeks after ACLT surgery, 0.5 mg/kg DEX was injected into the knee joint in the DEX group, and 10 mM HPT was injected into the knee joint in the HPT group. Four weeks after drug treatment, all the rats were sacrificed by carbon dioxide anesthesia, then cervical dislocation, and articular cartilage samples were collected for microscopic evaluations. Six samples were collected in each group.

The samples of the OA animal model were collected 8 weeks after ACLT. First, the knee joints of the rats were fixed in 4% paraformaldehyde for 2 days. Subsequently, they were decalcified in 10% EDTA solution for 2 months, dehydrated with a concentration gradient of alcohol, embedded in paraffin, and cut into 10 μm thick paraffin sections. All specimens were stained with hematoxylin and eosin (H&E) and safranin-O fast green. H&E staining was used to observe changes in the surface of articular cartilage, cartilage levels, and the number and morphology of cartilage cells. In addition, the histological immunofluorescent staining for p-AMPKα1 (Zen Bio, 310044) was performed to verify HPT exhibits the effect in vivo on regulating the AMPK pathway. In the microscopic examination, the histological score was scored according to the histological evaluation performed in the previous study (Xu et al., 2019; Wang et al., 2020b). The (SD) score of each group was the content of cartilage matrix captured from the safranin-O-positive staining area by ImageJ software. In short, we selected the ROI of the red area of the cartilage matrix in each group, and then, calculate the intensity value of the algorithm built into the ImageJ software for further statistical analysis. The statistical analysis is based on the mean ± SD.

SPSS 22.0 software package was used to perform statistical processing. One-way analysis of variance was used to determine significant differences between groups. The experimental data were expressed as mean ± SD (mean ± SEM), and p < 0.05 was considered statistically significant.

Figure 1A presents the chemical structure of HPT. CCK-8 assay was used to investigate the cytotoxic effects of HPT on SD rat chondrocytes. The results showed no apparent cytotoxic effects of HPT on chondrocytes at concentrations of 0–20 μM; however, it had cytotoxic effects at concentrations ≥50 μM (Figure 1B). In addition, HPT promoted the proliferation of chondrocytes slightly at 10 μM concentration. Therefore, 10 μM HPT was the concentration selected for further studies. We further investigated the cell proliferation rate after treatment with 10 μM HPT for 5 days. Compared with the control group (DMSO), supplementation with 10 μM HPT increased cell proliferation (Figure 1D). There was no change in the morphology of chondrocytes after treatment with 1–20 μM HPT; however, chondrocytes exposed to 50 and 100 μM HPT showed significant morphological changes, and cell counts decreased in a dose-dependent pattern as well (Figure 1E).

Chondrocytes are generally very sensitive to TNF-α, which reduces the aggrecan and collagen type II synthesis and promotes the development of OA. To investigate the protective function of HPT in TNF-α-treated chondrocytes, we examined the effect of HPT on aggrecan and collagen type II using the Western blotting analysis and RT-PCR. Moreover, we then examined the impact of HPT on TNF-α-induced production of IL-1β, IL-6, MMP13, and PTGS2 in chondrocytes. The cells were treated with 10 μM HPT and 10 ng/ml of TNF-α for 48 h. Compared with the control group, TNF-α significantly increased mRNA levels of IL-1β, IL-6, MMP-13, and PTGS2 and inhibited ACAN and COL2A gene expression (*p < 0.05; TNF-α vs. control), while HPT recovered these gene levels (#p < 0.05; TNF-α + HPT vs. TNF-α) (Figures 2A–F). Like gene expression results, HPT restored TNF-α-induced downregulation of COL2A and ACAN and reversed TNF-α-induced IL-1β, IL-6, MMP-13, and PTGS2 protein production (Figures 2G,H). Moreover, DEX also restored TNF-α-induced downregulation of COL2A and reversed TNF-α-induced IL-1β, IL-6, and MMP-13 protein production (Figures 2I,J). Finally, 20 and 30 μM HPT also restored TNF-α-induced downregulation of COL2A and reversed TNF-α-induced IL-1β and MMP-13 protein production (Figures 2K,L). These findings indicated that HPT could activate cartilaginous ECM synthesis and suppress the associated TNF-α-induced expression of inflammatory genes.

The qRT-PCR and Western blotting were used to investigate the gene and protein expression of Bcl-2, CDK1, and Cyclin D1. The results showed that TNF-α significantly decreased Bcl-2, CDK1, and Cyclin D1 mRNA levels (*p < 0.05; TNF-α vs. control), while HPT recovered these gene levels (#p < 0.05; TNF-α + HPT vs. TNF-α) (Figures 3A–C). Similar to the transcription level, the decreased Bcl-2, CDK1, and Cyclin D1 levels induced by TNF-α were restored to the control group level in the group with TNF-α + HPT (*p < 0.05, TNF-α vs. control; #p < 0.05, TNF-α + HPT vs. TNF-α) (Figures 3D,E). We further investigated the cell cycle to determine whether HPT protects chondrocytes against the inhibitory effect of TNF-α on the proliferation of chondrocytes by flow cytometry. The results showed significant decreases in the S phase following TNF-α treatment (*p < 0.05; TNF-α vs. control), while the percentages of chondrocytes in the S phase were restored to the control group level in the group with TNF-α + HPT (#p < 0.05; TNF-α + HPT vs. TNF-α) (Figures 3F,G).

The RNA-seq experiments were performed in cells affected by TNF-α induction with or without HPT treatment to further investigate the protective and anti-inflammatory mechanism of HPT in TNF-α-induced chondrocytes. The results demonstrated that compared with the TNF-α group, there were many differentially expressed genes (DEGs) in chondrocytes treated with TNF-α and 10 μM HPT, of which 105 genes were upregulated and 444 genes were downregulated (Figure 4A). Furthermore, GO functional annotations were made on the identified DEGs, which were highly involved in cell processing, biological regulation, metabolic processes, and biological process regulation (Figure 4B). Furthermore, the KEGG signaling pathway enrichment analysis indicated that these DEGs were highly enriched in functions related to the longevity-regulating pathway, regulation of lipolysis in adipocytes, insulin signaling pathway, AMPK signaling pathway, and glucagon signaling pathway (Figure 4C).

Based on the results of highly enriched signaling pathways, we selected AMPK signaling pathways for further analyses. Western blot analysis showed that TNF-α induction suppressed the phosphorylated-/total-AMPKα1 (p/t-AMPKα1) ratios; however, the phosphorylation of AMPKα1 increased after HPT supplementation (Figures 5A,B) (*p < 0.05). Furthermore, immunofluorescence staining with p-AMPKα1 was performed, which further supported the Western blot results (Figure 5C), indicating that HPT plays a key role in the activation of AMPK phosphorylation in chondrocytes. To determine the role of the AMPK signaling pathway in HPT-induced alleviation of inflammation, we examined the effect of HPT on TNF-α-induced inflammation with or without compound C, a selective AMPK inhibitor. Inhibition of AMPK activation by compound C abolished the HPT-induced decrease in IL-1β and MMP-13 levels (Figures 5D–F) (^a p < 0.05). In addition, it also diminished the upregulation of p-AMPKα1 and COL2A expression by HPT treatment (Figures 5D,G,H) (^a p < 0.05). These data suggested that HPT relieves TNF-α-induced inflammation of chondrocytes through the AMPK signaling pathway.

ACLT was used to establish the rat OA model. Healthy rats served as a control group. HPT was locally injected into the joint cavity 4 weeks after ACLT surgery. As shown in Figure 6, the articular cartilage had a regular shape in the healthy group. The articular cartilage morphology in the OA group was uneven, with reduced cartilage thickness compared with the healthy group. The articular cartilage in the HPT group was more regular, with significantly higher cartilage thickness compared with the OA group. The cartilage matrix of the OA group was significantly thinner than that of the healthy group (*p < 0.05). HPT treatment significantly changed this situation (#p < 0.05). In addition, compared with the healthy group, the histological score of the OA group was significantly different (*p < 0.05); however, DEX and HPT treatment significantly improved cartilage degeneration, and the histological score was significantly different from that of the OA group (#p < 0.05), indicating that HPT treatment successfully inhibited cartilage degeneration. In addition, the expression of p-AMPKα1 was measured in vivo by immunofluorescent staining (Figure 6). The result showed that TNF-α induction suppressed the phosphorylation of AMPKα1; however, the phosphorylation level of AMPKα1 was recovered after DEX and HPT supplementation.