Animal experiments were approved by the Laboratory Animal Ethical and Welfare Committee of Hebei Medical University (Hebei, China; approval no. IACUC-Hebmu-2021007), following The Guideline for the Care and Use of Laboratory Animals (25).

Proteins that interact with OTUB2 were analyzed using the STRING database (https://cn.string-db.org/). Gene Ontology (GO) enrichment analysis of proteins that interacted with OTUB2 was carried out using the DAVID database (https://david.ncifcrf.gov/home.jsp).

A 2nd generation lentiviral vector system was used. For OTUB2 overexpression, OTUB2 cDNA was amplified and subcloned into the lentivirus vector (LV), pLVX–IRES-Puro (Unibio) and referred to as LV-OTUB2. pLVX–IRES-Puro with no cDNA insertion was used as the negative control (NC) for LV-OTUB2 and referred to as LV-NC. For TNF-receptor associated factor 3 (TRAF3) knockdown, TRAF3 short hairpin RNA (shRNA) was synthetized and subcloned into pLVX-shRNA1 (Unibio). shNC, a non-targeting shRNA, was subcloned into pLVX-shRNA1 and used as the NC for LV-shTRAF3. Subsequently, pLVX-OTUB2/pLVX-shTRAF3/NC, pSPAX2 and pMD2.G vectors were mixed at a ratio of 4:3:1 (14.0:10.5:3.5 µg in a 10 cm culture flask) and transfected into 293T cells (iCell Bioscience). Following transfection, recombinant virus-containing supernatants were collected and filtered using a 0.45-µm membrane. In the logarithmic phase of growth (multiplicity of infection, 20) BMSCs were infected with LV-NC, LV-OTUB2, LV-shNC or LV-shTRAF3, as previously described (26). Sequences of shTRAF3 and shNC were as follows: shTRAF3, 5′-ccgCGAAGACAGTGGAGGACAAGTttcaagagaACTTGTCCTCCACTGTCTTCGttttt-3′; and shNC, 5′-ccgTTCTCCGAACGTGTCACGTttcaagagaACGTGACACGTTCGGAGAAttttt-3′. Cells were further subjected to the osteogenic differentiation at 72 h post-transfection.

Male Sprague Dawley rats (age, 12 weeks; weight, 410±10 g) were obtained from Liaoning Changsheng Biotechnology. Rats were housed with food and water ad libitum in a humidity (50±10%) and temperature-controlled (22±1°C) environment for a week of acclimation under a 12/12 h light/dark cycle. The fracture model was established as previously described (4), Briefly, animals were subjected to inhalation of isoflurane for anesthesia, using an induction rate of 5% and a maintenance rate of 2%, as previously described (27). An incision was made on the middle lateral thigh of rats, followed by the blunt dissection of muscle around the femur. The femur was cut and a sterile Kirschner wire was inserted to stabilize the fracture. Subsequently, rats were randomly divided into three groups: i) Control; ii) LV-NC; and iii) LV-OTUB2. Rats in the LV-NC and LV-OTUB2 groups were injected with 1×10⁸ transducing units/ml (50 µl/injection) of LV-NC or LV-OTUB2 at the fracture site and this was repeated once a week for 2, 4 or 8 weeks. Fractures of rats in the control group were left untreated. Rats were sacrificed with 70% vol/min CO₂ at each time point unless they met a prior humane endpoint. All animals were evaluated, weighted and scored daily with set criteria as previously described (28,29). Briefly, any finding listed as a humane endpoint, a score of 3 in two or more categories and a total score >9 led to the animal being euthanised. The categories including general appearance, activity, hydration, respiration, ambulation, surgical complications and weight loss. Following euthanasia, cardiac and respiratory arrest, as well as fixed and dilated pupils were observed in the rats for ~10 min before the animals were confirmed as deceased. A total of 162 rats were used in the present study. Among them, 36 rats [18 rats for western blotting; 18 rats for X-radiography and reverse transcription-quantitative PCR (RT-qPCR); n=6 rats/group] were used at 2 weeks post fracture. A total of 54 rats (18 rats for western blotting; 18 rats for X-radiography and RT-qPCR; 18 rats for pathological staining; n=6 rats/group) were used at 4 weeks post fracture. A total of 72 rats (18 rats for western blotting; 18 rats for X-radiography and RT-qPCR; 18 rats for micro-CT and pathological staining; 18 rats for TNAP activity detection; n=6 rats/group) were used at 8 weeks post fracture.

Total RNA was extracted from the callus tissue of the femur of rats using TRIpure lysis solution (BioTeke Corporation) according to the manufacturer's protocols. Subsequently, total RNA was reverse transcribed into cDNA using the BeyoRT II kit (Beyotime Institute of Biotechnology) according to the manufacturer's instructions. Briefly, total RNA, oligo (dT), reaction buffer, RNase inhibitor, dNTP Mix and BeyoRT II M-MLV reverse transcriptase were used to form a reaction system. According to the manufacturers' instructions, the system was incubated at 42°C for 50 min, followed by incubation at 80°C for 10 min. Subsequently, cDNA was amplified using 2X Taq PCR MasterMix and SYBR Green PCR Master Mix (Beijing Solarbio Science & Technology Co., Ltd.) in accordance with manufacturer's instructions. The reaction system included 1 µl cDNA, 1 µl primer, 0.3 µl SYBR Green, 10 µl 2X Taq PCR MasterMix and 7.7 µl ddH₂O. The thermocycling conditions used were as follows: 94°C for 5 min, followed by 40 cycles of 94°C for 15 sec, 60°C for 25 sec and 72°C for 30 sec. The relative expression of targeted genes was quantified using six independent measurements and the 2^−ΔΔCq method (30). β-actin was used as the internal reference gene. Primer sequences are listed in Table I. The experiments were repeated thrice.

Total protein was extracted from BMSCs or the callus tissue in the femur of rats using Radioimmunoprecipitation assay buffer (Beyotime Institute of Biotechnology) mixed with phenylmethanesulfonyl fluoride (100:1; Beyotime Institute of Biotechnology). Samples were centrifuged at 10,000 × g for 5 min at 4°C and supernatants were collected. Total protein was quantified using a bicinchoninic acid protein assay kit (Beyotime Institute of Biotechnology) according to the manufacturer's protocol. A total of 20 µg of protein was separated by SDS-PAGE s with a 5% stacking gel and a 10% running gel. Separated proteins were transferred onto PVDF membranes and blocked with 5% skimmed milk diluted in tris-buffered saline tween-20 for 1 h at room temperature. Following blocking, membranes were incubated with the following primary antibodies at 4°C overnight: Polyclonal rabbit anti-OTUB2 (1:500; Sabbiotech), polyclonal rabbit anti-TNAP (1:1,000; ABclonal Biotech, Co., Ltd.), polyclonal rabbit anti-RUNX2 (1:500; Proteintech Group, Inc.), polyclonal rabbit anti-osteoprotegerin (OPG; 1:1,000; Affinity Biosciences, Ltd.), polyclonal rabbit anti-receptor activator of nuclear factor-κB ligand (RANKL; 1:500; Affinity Biosciences, Ltd.), polyclonal rabbit anti-TRAF3 (1:1,000; ABclonal Biotech, Co., Ltd.), polyclonal rabbit anti-Flag (ABclonal Biotech, Co., Ltd.), polyclonal rabbit anti-ubiquitin (Proteintech Group, Inc.) and monoclonal mouse anti-β-actin (1:1,000; Santa Cruz Biotechnology, Inc.). Following primary incubation, membranes were incubated with the following secondary antibodies at 37°C for 45 min: Goat anti-rabbit horseradish peroxidase-conjugated IgG (1:5,000; Beyotime Institute of Biotechnology) and goat anti-mouse horseradish peroxidase-conjugated IgG (1:5,000; Beyotime Institute of Biotechnology). Protein bands were visualized using ECL reagent (Beyotime Institute of Biotechnology) on the WD-9413B gel imaging system (Liuyi Biotechnology). Six independent measurements were obtained using callus tissue samples and three independent measurements were obtained using BMSC samples. The band densities were quantified with the Gel-Pro-Analyzer software (version 4.0; Media Cybernetics, Inc.).

Fractured femurs were imaged using the CSM-2R X-ray apparatus (Softex) at 2-, 4- and 8-weeks post-fracture. Radiographic fracture healing was scored from Grade 1–6, as previously described (3). Briefly, fracture healing was scored as follows: i) Grade 1, no calcification; ii) Grade 2, patchy calcification; iii) Grade 3, calcification takes on the appearance of a callus; iv) Grade 4, callus bridging across the fracture gap; v) Grade 5, continuity of bone trabeculae; and vi) Grade 6, remodeling to normal bone.

Fractured femurs were scanned using the Quantum GX micro-CT imaging system (PerkinElmer, Inc.). Briefly, fractured femurs were scanned at 90 kV and 88 µA using a 50-µm voxel. Using the fracture line as the center, 50 slices were measured between the distal and proximal edges. The callus was segmented from the background as a region of interest. Subsequently, bone volume fraction (BV/TV) and bone mineral density (BMD) were auto-obtained using micro-CT analysis and 3D images were captured (31,32).

Tissues were fixed in 10% formalin solution at 4°C for 48 h. The fixed tissues were rinsed in running water for 4 h at room temperature. Samples were dehydrated in gradient ethanol (50–100 %), embedded in paraffin and then cut into 5 µm sections using a microtome (cat. no. RM 2235; Leica Microsystems, Inc.). Paraffin-embedded sections of fracture callus were deparaffinized in xylene and rehydrated in an ethanol gradient with distilled water. Formation of callus was examined using H&E staining (33). Cartilaginous ossification was analyzed using safranine O-fast green staining. Briefly, sections were stained with safranine O for 5 min at room temperature, followed by rinsing with ethanol gradient. Sections were subsequently stained with fast green for 1 min at room temperature and imaged using a BX53 microscope in bright-field mode (Olympus Corporation).

Isolation and osteogenic differentiation of BMSCs were performed as previously described (24). Briefly, femurs and tibiae were collected from rats and muscles surrounding the bone were removed. The marrow cavity was flushed with DMEM (Wuhan Servicebio Technology Co., Ltd.) for cell harvesting and cells were centrifuged at 300 × g for 7 min at room temperature. Subsequently, cells were cultured in DMEM supplemented with 10% fetal bovine serum (Sijiqing Biological Engineering Materials Co., Ltd.) at 37°C in an incubator with 5% CO₂. BMSCs were cultured until the third passage was reached and used in subsequent experiments after ~2 weeks. BMSC phenotype was confirmed via flow cytometry with surface antigens.

Osteogenic differentiation was performed at 72 h post-transfection, as previously described (34). Briefly, cell medium was removed and replaced with medium for inducing osteogenic differentiation [100 nM dexamethasone (MilliporeSigma), 50 µM vitamin C (Sinopharm Chemical Reagent Co., Ltd.) and 10 mM β-sodium glycerophosphate (MilliporeSigma)]. Differentiation medium was changed every 2 days. At 3 days post-induction, total protein was extracted for western blotting analysis (35). At 7 days post-induction, TNAP activity was detected and at 14 days post-induction, Alizarin Red S staining was performed to assess calcium deposition, as previously described (35,36).

Following centrifugation at 300 × g for 7 min at 4°C, BMSCs were rinsed and suspended with 100 µl phosphate buffer saline. In total, 1×10⁶ cells were incubated with 0.25 µg fluorescein isothiocyanate (FITC)-conjugated anti-CD45 (MultiSciences Biotech), 1 µg FITC-conjugated anti-CD29 (BioLegend, Inc.), 0.06 µg FITC-conjugated anti-CD90 (BioLegend, Inc.) and 0.25 µg PE-Cyanine 7-conjugated anti-CD31 (Thermo Scientific, Inc.) at 4°C in the dark. Subsequently, cells were analyzed via NovoCyte flow cytometry (ACEA Biosciences), using three independent measurements. Data was analyzed using NovoExpress software (version 1.4.1; Agilent Technologies, Inc.).

Cells were fixed with 4% paraformaldehyde at room temperature for 15 min and stained with Alizarin Red for 30 min at room temperature. Images were captured using a microscope (IX53l; Olympus Corporation). Calcium deposits were quantified following the addition of 10% cetylpyridinium chloride (Shanghai Macklin Biochemical Co., Ltd.) for 15 min at room temperature. Absorbance was measured at a wavelength of 570 nm (37).

The levels of TNAP were assessed using the TNAP Detection Kit (Wanleibio Co., Ltd.) according to the manufacturer's instructions. Briefly, ultrasonic cell disruption (300 W; ultrasonication for 3 sec with an interval of 30 sec repeated 5 times) was carried out in an ice bath and cells were subsequently centrifuged at 421 × g for 10 min at 4°C. In total, three independent supernatant samples were used for the detection of TNAP levels.

Callus tissues were homogenized with normal saline (1:9; g/v). The tissue homogenate was centrifuged at 421 × g at room temperature for 10 min. The resulting supernatant was obtained for the detection of TNAP levels. A total of six independent supernatant samples were used for the detection of TNAP levels.

Wild-type fragments of OTUB2 cDNA, wild-type TRAF3 cDNA and OTUB2 fragments with a point mutation at C51S were synthetized by General Biotech (Anhui) Co., Ltd. Subsequently, wild-type OTUB2 fragments and OTUB2 fragments with a point mutation were inserted into pcDNA3.1(+)_myc-His A vectors (General Biotech (Anhui) Co., Ltd.). Wild-type fragments of TRAF3 were inserted into p3×FLAG-CMV-10 vectors (General Biotech (Anhui) Co., Ltd.). All constructs were verified using enzyme digestion and sequencing. Subsequently, Lipofectamine^® 3000 was used to transiently transfect OTUB2-Myc/OTUB2-Myc (C51S) and TRAF3-Flag plasmids into 293T cells. One day prior to transfection, cells were cultured in 6 well-plates until 80% confluence was reached. In addition, cells were starved for 2 h prior to transfection and subsequently incubated with OptiMEM containing 5 µl Lipofectamine^® 3000 (Invitrogen; Thermo Fisher Scientific, Inc.), 5 µl P3000 reagent and 2 µg plasmids. Following 48 h of transfection, 293T cells were lysed with radioimmunoprecipitation assay buffer mixed with phenylmethanesulfonyl fluoride (100:1), followed by mixing in liquid nitrogen. Cells were centrifuged at 10,000 × g for 5 min at 4°C and the supernatant was collected and subjected to immunoprecipitation using a co-immunoprecipitation kit (cat. no. 26149; Thermo Fisher Scientific, Inc.). Briefly, 20 µl of resin slurry was used to pre-clear the lysates, which were further incubated with 1 µg of immobilized antibodies (Flag-tag or Myc-tag). Following elution and centrifugation (1,000 × g for 5 min at 4°C) to obtain the precipitates, proteins were used for subsequent western blot analysis.

OTUB2 and TRAF3 sequences were obtained from the National Center for Biotechnology Information (https://www.ncbi.nlm.nih.gov/) and crystal structures were obtained via homology modeling on the SWISS-MODEL database (https://swissmodel.expasy.org/). Interactions between proteins were forecasted using the GRAMM database (https://gramm.compbio.ku.edu/request) and a graphic representation of the protein-protein interaction was established using PyMOL software (version 2.0; Schrodinger). Grey was indicative of the OTUB2 protein, purple was indicative of the TRAF3 protein and stick structures of red and blue were indicative of binding. Notably, the yellow dotted line was indicative of a hydrogen bond.

Data are presented as the mean ± standard deviation, or as box and whisker plots, with the ‘box’ depicting the median, 1st quartile and 3rd quartile and the ‘whisker’ depicting the standard deviation. Data were analyzed using GraphPad Prism software (version, 8.0; GraphPad; Dotmatics). In vitro experiments were independently repeated three times and in vivo experiments were independently repeated six times (6 rats/group). Differences between multiple groups were analyzed using one-way ANOVA followed by Tukey's post hoc test and comparisons between two groups were analyzed using an unpaired Student's t-test. Effect size was calculated to determine the power of the analysis, due to the small sample size included in the present study, as previously described (38). The sample size of animals was chosen according to the results of power analysis. The priori/post-hoc power analysis was carried out using G*power 3.1.9.7 software (Franz Faul). It was found appropriate to complete the study with at least 6 samples for each group (α err probe was 0.05; effect size was 0.9; power was 0.8). Previous studies reported that Student's t-tests are suitable when sample size is small and effect size is large (39,40). In addition, bone fracture healing scores were analyzed using a Kruskal-Wallis test followed by Dunn's post hoc analysis. P<0.05 was considered to indicate a statistically significant difference.

To investigate the potential effects of OTUB2 on bone fracture healing, a fracture model was established through the injection of lentivirus targeting OTUB2 around the fracture sites of rats (Fig. 1A). At 2-, 4- and 8-weeks post-fracture, mRNA and protein expression levels of OTUB2 were elevated in the fracture callus of rats in the LV-OTUB2 group (Fig. 1B and C). The effects of OTUB2 overexpression were subsequently investigated using X-radiography and micro-CT analysis. Results of the present study revealed that the speed of fracture healing was increased and fracture healing remodeling was improved in the LV-OTUB2 group at 8 weeks post-fracture, when compared with the LV-NC group (Fig. 1D and E). In addition, the elevated callus area was observed at the fracture sites (Fig. 1D). Results of the present study also revealed that specimens exhibited severe loss and erosion of the femur following bone fracture; however, this was alleviated following OTUB2 overexpression. Representative micro-CT 3D images of fractured femur are displayed in Fig. 1F. In addition, quantification of the micro-CT 3D analysis demonstrated that BMD and BV/TV were increased following OTUB2 overexpression, indicative of accelerated bone fracture healing and bone formation (Fig. 1G). These results highlighted that OTUB2 may promote bone fracture healing.

The potential effects of OTUB2 on bone callus formation and cartilaginous ossification were further investigated in the present study. Results of the H&E staining analysis demonstrated that OTUB2 overexpression was associated with improved recovery at the fracture sites, with the formation of woven bone at 4- and 8-weeks post-fracture, compared with the LV-NC group (Fig. 2A). In addition, OTUB2 overexpression promoted soft callus formation at 4- and 8-weeks post-fracture. OTUB2 overexpression was associated with the domination of woven bone at the fracture sites at 8 weeks post-fracture and this was close to union (Fig. 2A). Results of the safranine O-fast green staining analysis revealed that the cartilage area of rats in the OTUB2 overexpression group was reduced at 4- and 8-weeks post-fracture, compared with the LV-NC group (Fig. 2B). At week 8, a few calcified cartilages and a limited number of uncalcified cartilages were observed in the fracture calluses of rats in the OTUB2 overexpression group (Fig. 2B), with ossification of the cartilage callus. Collectively, these data suggested that OTUB2 may potentiate fracture callus formation and cartilaginous ossification.

Bone formation and mineralization, cartilage maturation and bone mass play vital roles in the bone remodeling of fracture. Thus, genes and proteins associated with bone fracture were investigated at 2 weeks post-fracture. Results of the present study revealed that OTUB2 overexpression was associated with increased TNAP activity (Fig. 2C). In addition, OTUB2 overexpression was associated with increased TNAP and RUNX2 mRNA expression levels and an increase in the ratio between OPG and RANKL expression (Fig. 2D-F). Notably, results of the western blot analysis were comparable with those obtained using RT-qPCR (Fig. 2G and H).

Results of the present study revealed that OTUB2 exhibited protective effects on bone fracture healing in vivo. Thus, it was hypothesized that OTUB2 may play a similar role in vitro. In the present study, a model of osteogenic differentiation was established using BMSCs (Fig. 3A). BMSCs were isolated from rats and the phenotype was identified using flow cytometry with surface antigens. Results of the present study exhibited positivity for antigens CD29 and CD90 and negativity for antigens CD45 and CD31, indicative of a high cell purity (Fig. 3B). Notably, OTUB2 expression was increased in BMSCs infected with LV-OTUB2 (Fig. 3C). Results of the Alizarin Red S staining analysis revealed that OTUB2 potentiated the mineralization of BMSCs, with an increased number of calcified nodules that were bright red in color (Fig. 3D and E). In addition, OTUB2 overexpression was associated with increased TNAP expression levels (Fig. 3F), increased RUNX2 and OPG expression levels and reduced RANKL expression levels (Fig. 3G and H). Collectively, these results suggested that OTUB2 may facilitate the osteogenic differentiation and mineralization of BMSCs.

The present study aimed to determine the specific mechanisms underlying the OTUB2-induced increases in osteogenesis and mineralization. Notably, deubiquitinase OTUB2 interacts with substrate proteins to remove covalently-attached ubiquitin, thereby controlling substrate abundance. Thus, it was hypothesized that the protective role of OTUB2 may be mediated by the stability of downstream substrates and proteins that interacted with OTUB2 were analyzed using the STRING database. As displayed in Fig. 4A, results of the present study revealed 10 proteins that interacted with OTUB2. To further investigate the specific functions of these proteins, GO enrichment analysis was performed (Fig. 4B). Results of the present study revealed that the OTUB2-interacting proteins were enriched in ‘protein deubiquitination’, ‘ubiquitin-dependent protein catabolic process’ and ‘protein modification by small protein conjugated or removal’. These enriched terms provided direction for further investigation of the mechanisms downstream of OTUB2 and TRAF3, a protein enriched in ‘regulation of protein polyubiquitination’ was selected for subsequent analyses (Fig. 4C). As a member of the TRAF family, TRAF3 plays a role in promoting bone formation and remodeling (41). Thus, it was hypothesized that the protective effects of OTUB2 in fracture healing and the osteogenic differentiation of BMSCs may be associated with TRAF3. In the present study, co-immunoprecipitation analysis was used to verify the physical interaction between OTUB2 and TRAF3 in 293T cells (Fig. 4D and E). Results of the protein molecular docking analysis revealed a hydrogen bond between OTUB2 and the TRAF3 protein, indicative of binding (Fig. 4F). In addition, results of the present study revealed that OTUB2 reduced the ubiquitination of TRAF3 (Fig. 4G). Subsequently, HEK293 cells expressing inactive enzyme mutant OTUB2 C51S were generated and results of the present study revealed that the OTUB2 mutation enhanced the ubiquitination of TRAF3 (Fig. 4H). These results highlighted that OTUB2 may repress TRAF3 ubiquitination and this is dependent on the deubiquitinase activity.

As displayed in Fig. 2G and 3G, TRAF3 protein expression levels were increased following OTUB2 overexpression in vivo and in vitro. As a downstream factor of OTUB2, the role of TRAF3 attracted our interest. It was found that TNAP activity was reduced following TRAF3 knockdown (Fig. 5A) and the mineralization of BMSCs was markedly repressed (Fig. 5B and C). Results of the present study also revealed that TRAF3 knockdown reduced the expression levels of RUNX2 and OPG and promoted RANKL expression (Fig. 5D and E). Collectively, these results demonstrated that TRAF3 knockdown may reduce the osteogenic differentiation of BMSCs.

In the present study, TRAF3 expression was reduced following transfection with LV-shTRAF3 (Fig. 6A). Results of the present study revealed that TNAP expression levels were reduced and the mineralization of BMSCs with OTUB2 overexpression was repressed following TRAF3 downregulation (Fig. 6B-D). Collectively, these results demonstrated that OTUB2 may promote the osteogenic differentiation and mineralization of BMSCs through upregulation of TRAF3 (Fig. 6E).