Six human samples of the tibial plateaus were obtained from the OA patients who had undergone knee replacement surgery with approval from the Ethics Committee of Nanfang Hospital, Southern Medical University. The lateral portion of the tibial plateau that did not show significant wear was used as a control. Three levels of the sample sections were processed for each patient in histological examination and OARSI scoring as described previously. The number of positively stained cells in the subchondral bone was counted in 3 different areas for each patient in each group.
Wild-type (WT; Nox4+/+) and Nox4-deficient (C57BL/6N-Nox4em1cyagen; Nox4-/-) mice of C57BL/6 background were purchased from Cyagen Biosciences (Guangzhou, China). The transgenes were genotyped using the following primer pairs: 5’-ATTGGAGGGACAAGTTCTGATAG-3’, 5’-GAGGAGTCTTGTGGAAGAGTATG-3’, and 5’-CAAGTTCATGTTTCTCTTCTCTCTG-3’.
C57BL/6J (wild-type) mice were purchased from the Experimental Animal Center of Southern Medical University (Guangzhou, China). Unless otherwise stated, all mice were 12 weeks old and maintained on a sterile diet and autoclaved water in specific pathogen-free housing facilities at Southern Medical University (Guangzhou, China). Male mice were used for all in vitro experiments. All experiments were approved by the Southern Medical University Animal Care and Use Committee.
Mice knee OA model induced by anterior cruciate ligament transection (ACLT)
ACLT was performed on 12-week-old mice to induce abnormal mechanical loading-related OA in the knee joint, according to a previous study62. A sham operation was only opening the joint capsule and then suturing the incision. Mice were housed under standard conditions with 12-hour light and 12-hour dark cycles and fed and watered ad libitum in a temperature-controlled room. To examine the effect of GKT137831 on early OA, we made comparisons between the sham operation mice, ACLT mice and mice undergoing gavage with GKT137831 (60 mg/kg/d) after ACLT for two weeks.
Tissue sampling and preparation
All mice were euthanized by 1.2% tribromoethanol before sampling at 2 and 4 weeks post-ACLT (n=8). The mice knee joints freshly dissected were fixed in 4% paraformaldehyde at 4°C for 24 hours and then decalcified in 0.5 M EDTA (pH 7.4) for 7 days. Tissue was embedded in paraffin (3.5 µm) and OCT (18 µm) and sectioned continuously.
Knee specimens were exanimated by a high-resolution μ-CT scanner (SkyScan 1172, Bruker) with a voltage of 55 kV, a current of 145 mA, and a resolution of 12 μm per pixel. The longitudinal images of the medial tibial plateau were used for 3D reconstruction and histomorphometry analysis, with the entire medial compartment of the subchondral bone as region of interest. The 3D structural parameters included: relative bone volume or bone volume fraction (BV/TV), ratio of bone surface area to bone volume (BS/BV), bone mineral density (BMD), trabecular thickness (Tb.Th), and trabecular bone pattern factor (Tb.Pf).
Histochemistry, immunohistochemistry, and histomorphometric analysis
Serial sections were dewaxed in xylene, hydrated with graded ethanol, and stained with Safranin O/Fast Green and H&E as previously reported62. TRAP staining (Wako 294-67001, Japan) was performed according to the manufacturer's instructions, followed by restaining with methyl green. Next, immunohistochemical staining was performed following standard protocols. After dewaxing and hydration, 0.01 M sodium citrate buffer (ph 6.0) was used for antigen retrieval, and 3% hydrogen peroxide to reduce endogenous peroxidase activity. Tissue sections were permeabilized with 0.1% Triton X-100 before they were blocked in 10% goat serum to reduce nonspecific staining and then incubated with anti-Nox4 (Abcam, ab133303, 1/500) at 4°C overnight. A horseradish peroxidase-streptavidin detection system (Dako) was used to restain the sections with hematoxylin (Dako). Micrographs of sections were captured by the ortho-fluorescence microscope (Olympus DP71). We quantified articular cartilage degeneration in the tibial plateau joint using the OARSI scoring system as described previously55. The distance from the tideline to the subchondral bone plate (SBP) was measured as the thickness of calcified cartilage, and the distance from the tideline to the articular cartilage surface as the thickness of hyaline cartilage. Quantitative histomorphometric analysis was performed blinded using the software Image J (ImageJ 1.53). The number of positively stained cells was counted in the entire subchondral bone region of the tibia in each specimen and in five sequential sections per mouse in each group.
SA-β-Gal staining and immunofluorescence staining of bone tissue sections
Senescent cells were detected using the Senescence β-Gal Staining Kit (Cell Signaling Technology, #9860) according to the manufacturer's instructions and photographed with an ortho-fluorescent microscope (Olympus, DP71). For immunofluorescence staining, after dewaxing hydration, antigen retrieval, permeabilizing, and non-specific site closure, the paraffin sections were stained with anti-RANK (Abcam, ab13918, 1:200), anti-p16 (Proteintech, 10883-1-AP, 1:200), anti-p21 (Cell Signaling Technology, #2947, 1:200), anti-Nox4 (Abcam, ab133303, 1:200), anti-P-p38 (Cell Signaling Technology, #4511, 1:1000), anti-P-ERK1/2 (Cell Signaling Technology, #4370, 1:1000), and anti-P-JNK (Cell Signaling Technology, #4668, 1:1000) at 4 °C overnight. Then the sections were incubated with fluorescent secondary antibodies (Abbkine, A23210 & A23420, 1:200) at 37 °C for 1 hour. Cell nuclei were re-stained with DAPI (Solarbio, S2110) and observed under a confocal microscope (Zeiss, LSM 980). Analyzed were the data of 8 mice for each treatment group and 5 different regions of the tibial subchondral bone for each sample.
Cell culture, viral infections and osteoclast differentiation
As previously described21, bone marrow cells (BMMs) were isolated from the femur and tibia of 4-week-old male mice, screened and induced into osteoclast precursors (OCPs). Briefly, whole bone marrow cells were isolated by flushing the bone marrow cavities of the femur and tibia before cultured in α-MEM (Gibco) containing 10% serum (Gibco), and 1% penicillin-streptomycin (HyClone) for 24 hours. Floating cells collected were cultured with 70 ng/mL M-CSF (PeproTech, 315-02-10) for 48 hours to induce OCPs formation. OCPs were transfected using a lentivirus overexpressing Nox4 (GeneCopoeia, LPP-Mm06833-Lv201-400) or negative control (GeneCopoeia, LPP-NEG-Lv201-400). Successful gene overexpression was confirmed by immunoblot analysis. To induce the formation of mature osteoclasts, OCPs were cultured with 70 ng/mL M-CSF and 100 ng/mL RANKL (PeproTech, 315-11C-10) for 6 days. We used the TRAP staining kit (Sigma, 387A-1KT) to stain the mature osteoclasts differentiated from OCPs according to the manufacturer's instructions. TRAP+ osteoclasts with 5 or more nuclei were counted. To obtain the supernatants containing SASP factors, we collected the medium of transfected-OCPs treated with or without 2 mM NAC for 24 hours. H2O2, NAC and RANKL were added into SASP-containing medium to culture untreated OCPs for durations indicated in the figure legends.
RANKL and NAC treatment started the third day after complete transfection in transfected OCPs while RANKL was stimulated 2 days after formation of knockout OCPs. The above cell incubations were performed in a humidified incubator at 37 °C and 5% CO2, if not additionally indicated.
Cell proliferation assay
BMMs seeded in 96-well plates at 30% confluency were incubated with M-CSF at 37 °C for 2 days to induce OCPs formation. After respective treatments, cell proliferation was determined using an EdU imaging kit (UElandy, C6016) according to the manufacturer's instructions. The absorbance at 555 nm were observed under an inverted fluorescence microscope (Olympus, IX73). EdU assays were performed in triplicate each time and repeated three times.
Measurement of ROS
Tissue ROS levels were measured using the OxiSelect in vitro ROS/RNS detection kit (Cell Biolabs, STA-347). Frozen tissue sections were thawed at room temperature for 15 minutes, incubated with 20 μmol/L dichlorofluorescein diacetate at 37 °C for 20 minutes, washed and mounted, and observed under an upright fluorescence microscope (Olympus, BX63) (excitation wavelength, 485 nm; emission wavelength, 535 nm). Starved OCPs were incubated with dihydroethidium (KeyGEN, KGAF019, 1:1000) at 37 °C for 15 minutes, washed, and stimulated with M-CSF and α-MEM with or without RANKL for 15 minutes. Fluorescence images (excitation wavelength, 518 nm; emission wavelength, 605 nm) were acquired in an inverted fluorescence microscope (Olympus, IX73).
Determination of bone resorption
After suspension BMMs seeded on bone slices (IDS, DT-1BON1000-96) were induced into OCPs after 2 days of M-CSF stimulation, the OCPs were induced to differentiate into osteoclasts with 70 ng/ml M-CSF and 100 ng/ml RANKL for 10 days. Bone slices were washed with 6% sodium hypochlorite and PBS. The absorption pit area on the air-dried disk was observed under a microscope and measured with ImageJ (ImageJ 1.53).
The cells were inocubated on the cell slides for culture. Fresh cells were fixed with 4% paraformaldehyde and permeabilized with 0.1% Triton X-100. After washing, non-specific binding sites were blocked with 10% goat serum for 1 hour and overnight at 4 °C in the dark with anti-H3K9me3 (Cell Signaling Technology, # 13969) or anti-Ki67 (Abcam, ab15580). Fluorescent secondary antibodies (Abbkine, A23210 & A23420) functioned together for 1 hour at 37 ° C. Then, the cell slides were removed and sealed with DAPI. Phalloidin (AAT Bioquest, AAT-23119) was additionally added to mark the F-actin belts during the secondary antibody incubation phase. A confocal microscope (Zeiss, LSM 980) was used for the above observation and software Image J (ImageJ 1.53) for calculation.
Real-time quantitative PCR and microarray analysis
To determine the expression levels of SASP factors in GFP-transfected OCPs, Nox4-transfected OCPs and Nox4-transfected OCPs with NAC, total RNA was extracted using the EZ-press RNA purification kit (EZBioscience, B0004D) according to the manufacturer's instructions, for reverse transcription using Color Reverse Transcription Kit (EZBscience, A0010CGQ), and qPCR using 2 × Color SYBR Green qPCR Master Mix (EZBscience, A0012-R2) in ABI QuantS tudio5 (Applied Biosystems, QuantStudio5). The relative expression of each target gene was calculated using the 2-ΔΔCt method. Details of the primers are listed below.
IL-1α: F-5'-CGAAGACTACAGTTCTGCCATT-3', and R-5'-GACGTTTCAGAGGTTCTCAGAG-3'; IL-1β: F-5'-GCAACTGTTCCTGAACTCAACT-3', and R-5'-ATCTTTTGGGGTCCGTCAACT-3'; IL-6: F-5'-TAGTCCTTCCTACCCCAATTTCC-3', and R-5'-TTGGTCCTTAGCCACTCCTTC-3'; MMP3: F-5'- ACATGGAGACTTTGTCCCTTTTG', and R-5'-TTGGCTGAGTGGTAGAGTCCC-3'; TNF-α: F-5'-CCCTCACACTCAGATCATCTTCT-3', and R-5'- GCTACGACGTGGGCTACAG-3'.
Western blot analysis
Total protein was extracted using a total protein extraction kit (KeyGEN, KGP2100) and diluted with loading buffer (Pythonbio, AAPR39). Proteins (15 μg) were separated using SDS-Page and electrotransferred to a polyvinylidene difluoride (PVDF) membrane. After blocking, each membrane was incubated with a primary antibody at 4 ° C overnight and then with a secondary antibody corresponding to the host species of the primary antibody for 1 hour at room temperature. A luminescent substrate (Millipore, WBKLS0500) was used for visualization. Antibodies used were as follows: anti-p21 (Cell Signaling Technology, # 2947, 1:1000), anti-p16 (Proteintech, 10883-1-AP, 1:600), anti-Nox4 (Huabio, ET1607-4, 1:600), anti-NFATc1 (Santacruz, sc-7294, 1:200), anti-TRAF6 (Santacruz, sc-8409, 1:200), anti-c-Fos (Cell Signaling Technology, #2250, 1:1000), anti-CatK (Santacruz, sc-48353, 1:200), anti-β-actin (Abmart, m20011, 1:2000), anti-P-p38 (Cell Signaling Technology, #4511, 1:1000), anti-p38 (Cell Signaling Technology, #8690, 1:1000), anti-p-JNK (Cell Signaling Technology, #4668, 1:1000), anti-JNK (Proteintech, 66210-1-Ig, 1:600), anti-p-ERK1/2 (Cell Signaling Technology, #4370, 1:1000), anti-ERK1/2 (Cell Signaling Technology, #4695, 1:1000).
The microarray expression profiling dataset GSE51588, based on the GPL13497 platform, was downloaded from the Gene Expression Omnibus (GEO, https://www.ncbi.nlm.nih.gov/geo/). The sequencing included 50 samples the data of which had been normalized, 10 from human non-OA knees and 40 from human OA knees. To better characterize OA, we screened the samples according to the following criteria: (1) the medial tibial plateau; (2) male: female = 1:1; (3) normal: OA patients = 1:1; (4) eligible patients randomized. Consequently, 4 samples from human non-OA knees and 4 ones from human OA knees were obtained as follows: GSM1248763, GSM1248764, GSM1248765, GSM1248766, GSM1248791, GSM1248798, GSM1248800, and GSM1248804.
Differentially expressed genes (DEG)
After the data of the above 8 samples log2 transformed, DEGs in the medial subchondral bone of the tibial plateau in OA knees compared to non-OA knees were selected using the limma package. P < 0.05 and | log2FC | > 1 were used as thresholds to screen differential genes.
Functional enrichment analysis of DEGs
We used Sangerbox Tools (http://vip.sangerbox.com/) for functional enrichment analysis of DEGs. We displayed the functional enrichment genes from oxidative stress and cellular senescence, and screened out the most relevant differential genes.
All images resulted from at least three independent experiments with similar results. All data were expressed as mean ± SD using GraphPad Prism 9.0.0 (GraphPad Software, San Diego, CA, USA). One-way ANOVA followed by Tukey’s t tests, two-way ANOVA followed by Tukey’s t tests or two-tailed paired Student’s t tests were used to compare the means among groups. Significance levels were set at P < 0.05 and indicated by "*", P < 0.01 by "**", and P < 0.001 by "***".