All experiments were performed in accordance with the standard guidelines approved by the Sichuan University Animal Care and Use Committee (No. 2019155A). A combination of physical, chemical and enzymatic treatments was used to prepare the sDCB scaffold according to our published protocols with some modifications (Yang et al., 2018a; Qing et al., 2019). First, all soft tissues, including blood and marrow, were removed from fresh bovine bones (New Hope Group, China). Then, cortical bone cylinders (approximately 40 mm in length and 30 mm in diameter) were made using an automatic cutting machine (Q-80Z, Sanfeng Instrument Technology, Changzhou, China) under cool water condition. Sandpapers and files removed all cancellous bone in the cortical bone cylinders. Next, bone cylinders were manufactured into bone slices (10 mm × 5 mm × 0.5 mm) using an automated diamond wire saw (STX-202A, Kejing Auto-Instrument, Liaoning, China) along the long axis. Each bone slice was manually divided into two regions: undemineralized bone section within the bone slice (sDCB-D) and complete demineralized bone section within the bone slice (sDCB-D).

The sDCB-D region of sDCB was immersed in 0.6 N hydrochloric acid (HCl, Chron Chemicals, Chengdu, China) for 140 min at 25°C for complete demineralization according to previous studies (Lewandrowski et al., 1996; Summitt and Reisinger, 2003). After demineralization, samples were rinsed overnight in phosphate-buffered saline (PBS, HyClone, Logan, UT, United States). Next, the samples were degreased with methanol (Chron Chemicals) and ethanol (Chron Chemicals) (v/v, 1:1) for 12 h. After being rinsed in PBS, samples were decellularized with 0.5% Triton X-100 (Amresco, OH, United States), DNase (150 IU/ml, Roche, Mannheim, Germany), and RNase (100 μg/ml, Roche, Mannheim, Germany) for 24 h. Next, the sDCB-D region of sDCB was soaked in 3% hydrogen peroxide for 12 h to remove the inherent osteoinductivity. Finally, the sDCB was rinsed, lyophilized, and sterilized. The surface morphology and distribution of calcium (Ca) and phosphorus (P) in sDCB were characterized by field scanning electron microscopy (SEM, JSM-7500F, JEOL, Japan) with energy dispersive X-ray (EDX) detector. In addition, the composition within the two different regions (sDCB-B and sDCB-D) of sDCB was determined using Fourier transform infrared spectroscopy (FT-IR, Nicolet 6,700, Thermo Scientific, Waltham, MA, United States).

TDSCs and BMSCs from Sprague−Dawley (SD) rats were isolated and cultured according to our published protocol (Ning et al., 2015). In brief, TDSCs were isolated from Achilles tendons and flexor tendons of rats, and BMSCs were isolated from the bone marrow of femur and tibia. Then, TDSCs and BMSCs were cultured in a culture medium (low-glucose Dulbecco’s modified Eagle’s medium (L-DMEM), 20% fetal bovine serum (FBS) and 1% penicillin/streptomycin (Gibco, Amarrilo, TX, United States). The culture medium was changed every 3 days, and all experiments were performed with TDSCs and BMSCs at passage 3.

TDSCs were seeded on the surface of the sDCB-D region of sDCB at a density of 5×10⁴ cells per scaffold. Cells were cultured in a culture medium that was changed every 2 days. On the sixth day, 50 μM ascorbic acid (Sigma, St Louis, MO, United States) was added to the subsequent medium to promote the deposition of TDSC-derived ECM. After 2 weeks, the sDCB-D coated with TDSCs was decellularized with 0.5% Triton X-100. Furthermore, we set four different time points: 2, 5, 10, and 15 min to obtain the optimal decellularization time. The deoxyribonucleic acid (DNA) of each sample (n = 3 per group) was extracted using a tissue DNA isolation kit (Omega Biotek, GA, United States) and quantified using NanoDrop spectrophotometer (Thermo Fisher, United States). After decellularization, the sDCB-D-ECM region was formed, and the whole scaffold was named sDCB-ECM including two different regions: sDCB-B and sDCB-D-ECM. The surface morphologies of sDCB-D, sDCB-D with TDSCs and sDCB-D-ECM were characterized using the SEM (EVO 10, Carl Zeiss, Jena, Germany). In addition, the same samples from each group were fixed in 4% paraformaldehyde, embedded in paraffin, and longitudinally cut into 5 μm thick sections. Hematoxylin and eosin (H&E) and Masson’s trichrome staining were used to observe the cellular components and ECM of scaffolds. The biochemical compositions of sDCB and sDCB-ECM, including the stromal cell-derived factor 1 (SDF-1), biglycan (Bgn) and fibromodulin (Fmod), were quantified by the enzyme-linked immunosorbent assay (ELISA, SDF-1, DLDEVELOP, China; Bgn, Cloud-Clone, United States; Fmod, DLDEVELOP, China, n = 3 per group). Finally, the surface elastic modulus of sDCB-B, sDCB-D and sDCB-D-ECM was determined by an atomic force microscope (AFM, Cypher VRS, Asylum Research, MA, United States).

We evaluated the effect of each region of the scaffold separately on the stem cell attachment, viability, and proliferative capacity since the sDCB-ECM contained two different regions. Specifically, we used BMSCs to evaluate the effect of sDCB-B, and TDSCs to evaluate the effect of sDCB-D-ECM, based on the different regions where the scaffolds promoted regeneration. DCB was used as the control group. Briefly, DCB, sDCB-B and sDCB-D-ECM were placed in 24-well plates (Corning, NY, United States). BMSCs were seeded on the DCB and sDCB-B at a density of 5 × 10³ cells/cm² (n = 3 per group), and TDSCs were seeded on the DCB and sDCB-D-ECM at a density of 5 × 10³ cells/cm² (n = 3 per group) in culture medium. At 24 h after culture, to assess the effect of each region of the scaffold on cell attachment, first, the cell-seeded scaffolds were fixed in 4% paraformaldehyde and pretreated with 0.5% Triton X-100. Then, rhodamine-phalloidin (Abcam, United States) and 4,6-diamidino-2-phenylindole (DAPI, Sigma, United States) were used to stain the cytoskeleton and nucleus, respectively. Finally, the cells on the scaffolds were observed using a fluorescent microscope (Axio Imager Z2, Zeiss, Jena, Germany). Meanwhile, a Live/Dead assay kit (Beyotime, Shanghai, China) including calcein-AM and propidium iodide (PI) was used to assess the effect of each scaffold region on cell viability. The cell-seeded scaffolds were incubated for 30 min in a culture medium supplemented with 2 μM calcein-AM and 4 μM PI, and the scaffolds were observed by fluorescence microscopy (Axio Imager Z2, Zeiss, Jena, Germany). At 1, 3, 5, 7 days after culture, the proliferation of cells on the scaffolds was quantified using the Cell Counting kit-8 (CCK-8, Dojindo, Tabaru, Japan). The cells-seeded scaffolds were incubated in fresh medium without FBS, and then 10% CCK-8 solution was added to each well. After incubation at 37°C for 4 h, the scaffolds were removed and the absorbance of medium was measured at 450 nm using a spectrophotometer (Synergy H1, Bio-Tek, VT, United States).

The recruitment capacity of sDCB-ECM was evaluated using a 24-well chamber transwell chemotactic migration model (pore size: 8 μm, Corning, United States, n = 3 per group). The DCB was used as the control group, and two different regions of the sDCB-ECM were evaluated separately. First, DCB, sDCB-B or sDCB-D-ECM was put in the lower chamber, and 200 μL of BMSCs or TDSCs suspension (5 × 10⁵ cells/ml) was placed in the upper chamber after serum-starving for 12 h. Next, 1 ml L-DMEM supplemented with 1% FBS was added to the lower chamber. After incubation at 37°C for 12 h, non-migrated cells in the upper chamber were scraped with cotton swabs, and the cells that migrated through the pores were fixed with 4% paraformaldehyde. After staining with DAPI, BMSCs or TDSCs on the lower surface of the membrane were captured by a fluorescence microscope (Axio Imager Z2, Zeiss, Jena, Germany). Three random fields were obtained at ×100 magnification to calculate the mean number of cells in each well using ImageJ software (NIH, Bethesda, MA, United States).

At 7 and 14 days after culture, total RNA (n = 3 per group) was extracted by lysing the cells using the TRIZOL reagent (Invitrogen, Waltham, MA, United States) following the manufacturer’s protocol. The cDNA was synthesized from total RNA by reverse transcription using the GoScript Reverse Transcription System (Promega, Madison, WI, United States). The rat-specific primers for genes, including glyceraldehyde-3-phosphate dehydrogenase (GAPDH), collagen I (COL1), runt-related transcription factor 2 (RUNX2), alkaline phosphatase (ALP), osteocalcin (OCN), osteopontin (OPN), SRY-box transcription factor 9 (SOX9), aggrecan (ACN), scleraxis (SCX), tenomodulin (TNMD), thrombospondin-4 (THBS4) and tenascin-C (TNC), were synthesized by Sango Biotech (Shanghai, China). The primer sequences are shown in Table 1. The synthesized cDNA was amplified by PCR using a LightCycler 96 system (Roche, Munich, Germany). The 2^−△△Ct method was used to reflect the relevant expression level of the target gene, and the expression of the target gene was normalized to that of the housekeeping gene GAPDH.

At 7 and 14 days after culture, total proteins (n = 3 per group) were extracted by lysing the cells in Radio-Immunoprecipitation Assay buffer (Servicebio, Wuhan, China) containing protease inhibitors (Servicebio, Wuhan, China) on ice. The supernatants were obtained by centrifugation at 12,000 rpm for 10 min, and protein concentration were quantified by the BCA assay (Servicebio, Wuhan, China). Samples containing 20 μg of protein were loaded onto 10% SDS-PAGE and then were electrotransferred to polyvinylidene difluoride membranes (Millipore, MA, United States). After blocking with 5% bovine serum albumin for 30 min, membranes were incubated with primary antibodies against β-actin (1:1000, gb12001, Servicebio, China), RUNX2 (1:1000, bs-1134R, Bioss, Beijing, China), ALP (1:1000, gb11528, Servicebio, Wuhan, China), SOX9 (1:1000, gb11280, Servicebio, Wuhan, China), TNMD (1:1000, ab203676, Abcam, MA, United States) and THBS4 (1:1000, ab263898, Abcam, MA, United States) overnight at 4 °C. Next, the membranes were washed and incubated with a secondary antibody (1:3000, Servicebio, Wuhan, China) for 30 min at 25°C. After washing with Tris-buffered saline containing 0.1% Tween-20, the target protein bands were detected using a chemiluminescence imaging system (ChemiScope 6,300, Clinx, Shanghai, China) after reaction with chemiluminescence substrate (ECL, Servicebio, Wuhan, China), and their intensities were measured with AlphaEaseFC software (Alpha Innotech, TX, United States) and then normalized to β-actin.

All data were expressed as mean ± standard deviation (SD). An unpaired two-tailed Student’s t-test was used to compare the two groups. One-way analysis of variance was used to compare three or more groups. A value of p < 0.05 was considered to be statistically significant, and the statistical analysis was performed with SPSS Statistics software (version 24, IBM, NY, United States).

The sDCB-ECM scaffold was prepared as illustrated in Figure 1. The SEM results showed that sDCB contained two different regions (sDCB-B and sDCB-D), and EDX elemental mapping demonstrated that there was a gradient of Ca and P elements in the sDCB (Figures 2A,B). The FT-IR results indicated that the absorptions from phosphate vibrations were not observed in the sDCB-D region. In contrast, the absorptions from phosphate vibrations were observed in the sDCB-B region (Figure 2C). Before modification, the collagen of demineralized bone was evident in sDCB-D (Figures 3A,B, left panels). After TDSCs were cultured for 2 weeks, a thin layer of cells with their ECM could be observed on the surface of the sDCB-D region (Figures 3A,B, middle panels). After decellularization, a thin layer of stem cell-derived ECM was left on the surface of the sDCB-D-ECM (Figures 3A,B, right panels). The results of SEM and histological staining showed that the stem cell-derived ECM was preserved on the surface of the sDCB-ECM after decellularization treatment (Figures 3A,B, right panels). After decellularization treatment at different time points, we found that 10 and 15 min of Triton X-100 treatment effectively removed cellular and nuclear material (p < 0.05, Figure 4A). For biochemical components of TDSC-derived ECM, the ELISA results indicated that SDF-1, Bgn, and Fmod contents in the sDCB-ECM were significantly higher than those in the sDCB (p < 0.05, Figures 4B–D). For the surface elastic modulus, the value of sDCB-B was significantly higher than that of sDCB-D and sDCB-D-ECM (p < 0.05, Figure 4E), and the value of sDCB-D was higher than that of sDCB-D-ECM (p < 0.05, Figure 4E).

The results of Live/Dead cell staining confirmed that sDCB-B and sDCB-D-ECM regions of sDCB-ECM were suitable for the growth of BMSCs and TDSCs (Figure 5A). Moreover, the results of cytoskeleton staining revealed that BMSCs cultured on sDCB-B showed a polygonal shape, while TDSCs cultured on sDCB-D-ECM showed a shape of spindle (Figure 5B). After 1 and 3 days of growth, there was no significant difference in the number of BMSCs between the sDCB-B and DCB groups (Figure 5C). However, the number of cells in the sDCB-B group was significantly higher than that in the DCB group at 5 and 7 days after culture (p < 0.05, Figure 5C). At 1 and 3 days after culture, the number of TDSCs in the sDCB-D-ECM group was higher than that in the DCB group (p < 0.05, Figure 5D). However, there was no significant difference in the number of cells between the DCB and sDCB-D-ECM groups after 5 and 7 days of culture (Figure 5D). For cell migration analysis, the results showed that the number of migrated BMSCs in the sDCB-B group was significantly higher than that in the DCB group (p < 0.05, Figures 6A,B), and the number of migrated TDSCs in the sDCB-D-ECM group was significantly higher than that in the DCB group (p < 0.05, Figures 6A,B).

For differentiation-related gene expression in BMSCs, the expressions of COL1, RUNX2, ALP, OPN and ACN at 7 days after culture were significantly higher in the sDCB-B group than in the DCB group (p < 0.05, Figure 7A). However, the expressions of SCX, THBS4 and TNC were significantly lower in the sDCB-B group than those in the DCB group (p < 0.05, Figure 7A). At 14 days after culture, the expressions of RUNX2, ALP, SOX9 and ACN were significantly higher in the sDCB-B group than in the DCB group (p < 0.05, Figure 7A). However, the expression of THBS4 was significantly lower in the sDCB-B group than in the DCB group (p < 0.05, Figure 7A). For differentiation-related gene expression in TDSCs, at 7 days after culture, the sDCB-D-ECM group showed higher SOX9 and ACN expressions than the DCB group (p < 0.05, Figure 7B). In contrast, the sDCB-D-ECM group showed lower ALP and OPN expressions than the DCB group (p < 0.05, Figure 7B). At 14 days after culture, the sDCB-D-ECM group showed higher SOX9, SCX and TNC expressions than the DCB group (p < 0.05, Figure 7B). In contrast, the sDCB-D-ECM group showed lower expressions of RUNX2 and ALP than the DCB group (p < 0.05, Figure 7B).

For protein expression in BMSCs, we observed that the expressions of RUNX2 and SOX9 were significantly higher in the sDCB-B group than those in the DCB group (p < 0.05, Figures 8A,B) at 14 days after culture, but the expression of TNMD was significantly lower in the sDCB-B group than that in the DCB group (p < 0.05, Figures 8A,B). However, there was no significant difference in the expressions of RUNX2, ALP, SOX9, TNMD and THBS4 between the two groups at 7 days after culture (Figures 8A,B). For protein expression in TDSCs, we found that the expression of THBS4 was significantly lower in the DCB group than that in the sDCB-D-ECM group (p < 0.05, Figures 8C,D) at 7 days after culture. In addition, TNMD expression was significantly lower in the DCB group than in the sDCB-D-ECM group (p < 0.05, Figures 8C,D) at 14 days after culture. However, there was no significant difference in the expressions of RUNX2, ALP and SOX9 between the two groups (Figures 8C,D).