Diagnoses and degree classifications of patients were strictly applied according to the national criteria of KBD in China [WS/T 207–2010]. All subjects with alterations, such as defects and sclerosis on the bone end of phalanges combined with compression changes of the calcaneus and talus and enlarged/deformed limb joints (e.g., hand, elbow, knee, and ankle) manifested on X-ray, were diagnosed with KBD. Subjects were excluded if they were suffering or had previously suffered from any other osteoarticular diseases (such as rheumatoid arthritis, gout, or skeletal fluorosis) or any other type of macrosomia, osteochondrodysplasia disorder, or chronic disease (such as hypertension, diabetes, or coronary heart disease) or had received any treatment in the past six months. Clinical information was collected from patient records. Articular cartilage samples from KBD and OA patients were collected from individuals who underwent arthroplasty of the knee. Healthy controls were obtained from patients who suffered trauma or amputation due to an accident.

Donors signed a written informed consent form. Subjects were of Chinese Han lineage. For the verification of TSG-6 by qRT-PCR and immunohistochemistry, specimens of adult articular cartilage were collected from five KBD patients who had arthroplasty in the knee and five healthy subjects who had an amputation due to an accident in Xi’an Red Cross Hospital (Supplementary Table S1). For silencing and overexpression of TSG-6, articular cartilage samples were obtained from five KBD patients and three controls for the same reason described above (Supplementary Table S2).

All articular cartilage samples, including all of the cartilage zones (including calcified) and subchondral bone, were harvested from the lateral tibial plateau and obtained within 1 h after operation. Chondrocytes were isolated as follows: articular cartilage specimens were washed twice with sterile phosphate-buffered saline (PBS) with antibiotics (penicillin and streptomycin), cut into pieces (1 mm³), and subjected to enzymatic digestion with 0.25% trypsin at 37 °C in an atmosphere of 5% CO₂ for up to 30 min. Cell suspensions were centrifuged at 1,000×g for 5 min, the supernatant was completely aspirated, and the cells were digested in basal media supplemented with 0.2% type II collagenase at 37 °C using an Eppendorf Thermomixer for 12–16 h.

Cartilage tissues were fixed with 4% (w/v) paraformaldehyde for 24 h immediately after acquisition and decalcified in 10% (w/v) ethylenediaminetetraacetic acid disodium salt (EDTA-Na2) for 2–3 weeks. Samples were dehydrated in an alcohol series, cleared in xylene, and embedded in paraffin wax. Paraffin sections were cut into 5 µm sections, mounted on slides, and stored at room temperature until ready for staining. The paraffin-embedded sections were baked at 65 °C for 1 h, deparaffinized with xylene and then rehydrated in decreasing concentrations of ethanol. Endogenous peroxidase activity was blocked by 0.3% (w/v) hydrogen peroxide for 10 min at room temperature, and the sections were then washed with 1×PBS. Then, sections were incubated in a 10 mol/L urea solution diluted with water at 37 °C for 20 min and washed with 1×PBS. Sections were incubated in 0.1% trypsin/CaCl₂ at 37 °C and digested for another 20 min for antigen retrieval. After blocking with 5% (w/v) goat serum for 20 min at room temperature, anti-TNFAIP6 (1:100 dilution, PA5-76008, Invitrogen) antibodies and IgG as a negative control were applied onto the sections, and the samples were further incubated overnight at 4 °C. After washing with 1×PBS, sections were incubated using a human serum amyloid P (SAP) kit (solution B contains a goat anti-rabbit secondary antibody; Zhongshan, Jinqiao, Guangzhou, China) according to the manufacturer’s instructions. The substrate 3,3′-diaminobenzidine was added to stain the sections, and hematoxylin counterstaining was performed. Finally, sections were dehydrated and mounted under alcohol-cleaned coverslips. All IHC staining was assessed under light microscopy by two pathologists who were blinded to the origin of the samples. Articular cartilage was divided into three cell morphologies, namely, the superficial, middle, and deep zones, according to light microscopy observation. Chondrocytes in the superficial zone were relatively small and flat and were oriented with the long axis parallel to the surface; cells in the middle zone were larger and more rounded and were randomly distributed in the matrix with fibers running in oblique directions; and cells in the deep zone were larger in size and were arranged in a columnar manner perpendicular to the surface. An assessment of staining throughout each cartilage zone included systematic counting of positive and negative cells starting from the cartilage surface and progressing down through all layers of cartilage. Five randomly chosen fields in each zone were counted at 50× magnification. The percentage of positive cells was calculated using the number of positively stained cells divided by the total number of cells (positively and negatively stained cells) in the chosen fields of view. Percentages of positive cells in different zones were calculated for each case and for the different groups.

According to the previous gene expression profile in cartilage from KBD (Wang et al., 2009) and OA microarray data GSE57218 from the GEO database in NCBI, TSG-6 (TNF-stimulated gene 6, also known as TNFAIP6) was identified as one of the common differentially expressed genes of importance in ECM reconstruction and by maintaining cellular homeostasis. Differential expression of TSG-6 at the mRNA level in KBD chondrocytes was verified respectively by quantitative real-time polymerase chain reaction (qRT-PCR). TaKaRa Company (Japan) synthesized the TSG-6 primer (forward-GGAGTGAAAGATGGGCATGCC, reverse-CTCATTTGGGAAGCCTGGA). Experiments were conducted according to the manufacturer’s instructions.

To clarify the effect of TSG-6 abnormality in the ECM degradation of KBD chondrocytes, we silenced this gene in KBD chondrocytes by RNA interference technology and then separately overexpressed TNF-α induction. A pretest was applied to determine the optimum multiplicity of infection (MOI) and transfection time of the lentivirus. We designed five culture conditions combined with three preset MOI (Supplementary Table S3) and selected the group with an infection efficiency above 80% along with good cell growth status. The lentivirus system (GENECHEM, Shanghai) consisted of the objective vector GV248, pHelper 1.0, and pHelper 2.0. We designed three sections of TSG-6 for construction of the lentivirus vector including TNFAIP6-RNAi(54322–1), TNFAIP6-RNAi(54323–1), and TNFAIP6-RNAi(54324–11) (Supplementary Table S4). Transfected 293T cells were introduced into the lentivirus system for 48–72 h and then the supernatant was collected. Next, viruses were concentrated and purified by 0.45 μm filtration and ultracentrifugation. Finally, the quality of lentiviruses was determined involving physical form, sterility, and a titer test (via fluorescence).

Primarily, chondrocytes were cultured in 96-well plates (5×10⁴ cells/well) for 36 h. Then, 200 μl TNF-α (0, 5, 15, 25, 50, 75, 100 ng/ml, 5 repeats for each concentration) were added to further culture chondrocytes with 5% CO₂ for 24 h, 48 h, and 72 h (three repeats for each time point) at 37 °C. TNF-α is not supposed to remarkably affect the proliferation of chondrocytes. We attempted cell culture with serum-free medium as described in previous studies (Zhang et al., 2012; Yoshida et al., 2018). However, a downward trend of cell proliferation was observed when the TNF-α concentration added up to 100 ng/ml by MTT. This might have been due to the serum-free culture condition. Therefore, we added 2% serum into the medium, but the proliferation still presented a downward trend at 100 ng/ml of TNF-α. Hence, 100 ng/ml was set as the upper limited concentration. Both trends emerged at 48 h. No significant difference in the affection of cell proliferation was observed between the previously mentioned concentrations, which suggests that there is no need to try so many sub-dosages. Finally, we selected four concentrations of TNF-α (0, 5, 25, 100 ng/ml) with serum-free medium and one optimistic time point (48 h) for TSG-6 induction.

qRT-PCR was used to detect the expression changes of extracellular matrix MMP1, MMP3, MMP13, COL2A1, and proteoglycan (Abcam, United Kingdom, 1:500 dilution) before and after treatments. Information about the primers (TaKaRa, Bejing) of each gene is listed in Supplementary Table S5. Proliferation and viability of chondrocytes was observed by MTT before and after the gradient concentrations of TNF-α treatment.

To explore whether TSG-6 is the intermediate element in the chondrocyte ECM degradation process caused by T-2 toxin, gradients (0, 12, 50, 100, 150, 200, 250 nM, and 200 μl) of HT-2 toxin were added to intervene with the C28/I2 chondrocytes. Conventional culture conditions and repeat counts were the same as the overexpression section.

The morphological changes of chondrocytes were observed under a light microscope. The MTT test showed a dose-dependent inhibition effect of HT-2 on cell viability at 0, 12, 50, and 100 nM. However, the inhibitory effect did not continue to significantly increase with the increase in intervention concentration from 100, 150, and 200–250 nM. Therefore, we selected 0, 12, 50, and 100 nM to utilize in the following study.

Total RNA was extracted in subsequent experiments. The expression of TSG-6, MMPs, COL2A1, and proteoglycans in the extracellular matrix were detected using qRT-PCR before and after intervention according to previously described methods (Wang et al., 2017).

A statistical analysis was performed using the Statistical Package for the Social Sciences for Windows version 18.0 (SPSS, Inc.). Individual samples were divided into triplicate for the study. The differences in means were determined by one-way analysis of variance (ANOVA) for multiple comparisons. Student’s t-test was applied to determine the difference between two groups. The normality and homogeneity of variance of the data were tested before any further analyses. A non-parametric test was performed if the conditions for normality and homogeneity properties were not fulfilled, and p values less than 0.05 were considered significant.

We used immunohistochemistry to detect TSG-6 protein expression in KBD cartilage tissue (Figure 1A). Results showed that TSG-6 was upregulated in the superficial, middle, and deep zones of KBD cartilage compared to cartilage tissue from normal controls. We validated the gene expression of TSG-6 in patients with KBD using qRT-PCR and found that the results were consistent with those of the differentially expressed TSG-6 gene in KBD chondrocytes screened by previous microarray analysis (Figure 1B).

The lentiviral transfection pretest determined that the optimal transfection conditions were MOI = 100, ENi.S + P(E)+virus for culture, and 96 h for the duration. TNFAIP6-RNAi (54322–1) had the best infection efficiency and showed a 72% interference rate (Table 1; Figure 1C). After TSG-6 silencing, the expression levels of MMP1, MMP3, MMP13, and proteoglycan were significantly decreased, while the expression of COL2A1 was significantly increased in KBD chondrocytes (p < 0.05, Figure 1D).

Generally, with increasing TNF-α concentration and prolonged time, cell proliferation was faintly affected regardless of whether the chondrocytes were cultured with or without serum. However, cell proliferation showed a decreasing trend at 48 h and a significant difference at a dosage 100 ng/ml without serum compared to the blank group (p < 0.05) (Figure 2A). And cell proliferation showed no significant difference when cultured with serum (Figure 2B).

Before intervention, the expression levels of MMP1, MMP3, and MMP13 in the KBD group notably increased, while the expression levels of COL2A1 and proteoglycan significantly decreased compared with those in the healthy control group. TNF-α intervention results demonstrated that the expression of TSG-6 increased in KBD and healthy chondrocytes in a dose-dependent manner. Increased MMPs and decreased COL2A1 and proteoglycan were more significant in TNF-α treated KBD chondrocytes than in the blank KBD group (p < 0.05, Figure 2C). Moreover, healthy chondrocytes treated with TNF-α also had significantly increased MMPs and decreased COL2A1 and proteoglycan compared with the blank control group (p < 0.05, Figure 2D).

C28/I2 chondrocytes were treated with gradient concentrations of HT-2 toxin to clarify whether TSG-6 mediates environmental exposure and chondrocyte ECM degradation. By using MTT and light microscopy, we found that the survival rate of chondrocytes was correspondingly reduced in a dose-dependent manner (Figure 3A). Compared with the blank control group, the mRNA levels of TSG-6, MMP1, MMP3, and MMP13 were increased in the HT-2 toxin intervention group (0, 12, 50, and 100 nM), while the expression of COL2A1 was significantly decreased (p < 0.05, Figure 3B).