PBMSCs were isolated from 3-month-old New Zealand white rabbits. As previously reported, this procedure was performed (Fu et al., 2015). Briefly, rabbits were pretreated with granulocyte colony-stimulating factor (G-CSF) once daily for 5 days (50 μg/kg i.h.) and AMD3100 was administered on the sixth day (5 mg/kg i.h.) 1 h before blood sampling. The blood samples were harvested using a sterile syringe from the central ear artery and transformed into a centrifugal tube. The nucleated cells were isolated by density gradient centrifugation, seeded in a 10 cm dish at a density of 10⁶ cells/mL, and cultured with a complete medium consisting of low-glucose Dulbecco’s modified Eagle’s medium (DMEM) and 16% fetal bovine serum (FBS). After 10–14 days, colonies were formed. In this study, we use polyclonal PBMSCs mixed from multiple clones. PBMSCs at the third passage were utilized for further experiments. Crystal violet staining was used to assess colony formation.

The multipotent differentiation potential of PBMSCs was evaluated through adipogenic, osteogenic, and chondrogenic induction in vitro, as previous reported (Fu et al., 2014). Adipogenesis, osteogenesis, and chondrogenesis differentiation were validated by Oil red O staining, alizarin red staining, and alcian blue staining, respectively.

The surface markers of PBMSCs at passage 3 were analyzed using flow cytometry (FCM). Accordingly, MSCs were harvested with trypsin/EDTA and then incubated with CD29, CD34, CD44, and CD45 primary antibodies, respectively, for 30 min, followed by secondary fluorescein antibody for 30 min. Finally, cells were fixed in flow buffer, washed, and suspended in 0.2 ml of PBS for FCM analysis using Cell Quest software (BD Biosciences, San Jose, California). Parallel blank tubes were used as controls.

After being fixed with 4% paraformaldehyde for 24 h, the DCBM was dehydrated with gradient alcohol. Then it underwent wax dipping and embedding. The wax block with samples encapsuled was then cut into 4 μm sections. After dewaxing, the section was stained with hematoxylin for 5 min and eosin for 3 min, as HE staining. And another section was soaked in hematoxylin for 5 min, ponceau for 5 min, 2% glacial acetic acid for a while, 2.5% phosphomolybdic acid aqueous for 5 min and aniline blue for 5 min, as Masson staining.

The surface and section microstructure of DCBM was characterized by SEM. Before SEM imaging, dried samples were cut with a razor blade in liquid nitrogen and sputter-coated with platinum. SEM was used to observe the microstructure of the surfaces and sections of the samples at an accelerating voltage of 15 kV.

Ten mg of sample was cut into pieces, digested with 1 ml of metalloproteinase K (0.2 mg/ml) at 60°C for 24 h, and centrifuged at 15 kg for 30 min. The DNA content assay was performed according to the instructions of the kit (Quant-iTTM PicoGreen^® ds DNA Assay Kit), and a cortical bone that was not undecellularized was set as a control group.

An Instron (Norwood, MA) 5,543 testing frame was utilized to analyze the tensile properties of the hydrated constructs. Testing samples (n = 7) were hydrated for at least 30 min prior to testing. The specimen was fixed with a specific fixture, loaded on a biomechanical testing instrument, preloaded with 0.1 N, loaded at a rate of 0.5 mm/s and recorded the stress-strain curve to calculate the tensile elasticity modulus. The normal meniscus of rabbits was used as the control group.

To evaluate the cytotoxicity of the scaffold to PBMSCs, CCK-8 was used. A certain quality of DCBM scaffold was weighed and cut into pieces. After sterilization with ethylene oxide, the scaffold was added into a 6-well plate, as well as 10 times the quality of serum-free medium was added. The supernatant was collected as an extract medium 3 days later. The extract medium of the scaffold was used to culture the PBMSCs in a 96-well plate, as well as the normal medium was used to culture the same batch of cells as control. At the end of culture, 110 μl of CCK-8 working solution (volume ratio CCK-8: DMEM = 1:10) was added to each well after 2-h incubating at 37°C, the OD value of 450 nm was detected in a microplate reader.

Before seeding with cells, scaffolds were sterilized, washed in sterile PBS, and treated with DMEM overnight. Each scaffold was seeded with 10 μl of 1.2*10⁴/μl PBMSCs every 10 min for 4 times. Cells were allowed to adhere for 4 h and then the scaffolds were transferred to the incubator and the medium was changed every day.

To evaluate the viability of PBMSCs seeded in the material, a Live/Dead Assay kit (Invitrogen) was used. The cell/scaffold constructs were washed with sterile PBS and incubated in PBS solution with 2 mM calcein AM and 4 mM ethidium homodimer-1 for 1 h at room temperature, according to the instructions of the kit. After another wash with sterile PBS, constructs were observed using a confocal microscope (Leica).

We utilize the SEM to observe the microstructure of the cell-scaffold composites and the adhesion of PBMSCs cultured in vitro on the scaffolds. Cell-scaffold composites were acquired in 1, 2 and 3 days after seeding the cells. Samples were fixed in 2.5% (v/v) glutaraldehyde for 24 h, washed with PBS, dehydrated in graded alcohol and dried with CO₂ critical point drying. After coating with gold, cell-scaffold composites were observed using the SEM.

PBMSCs of passage 3 were seeded in a 6-well plate. Moreover, when the cells were more than 90% confluent, they were replaced with serum-free starvation for 12 h. A 200 μL gun tip was used to scratch along the straight line drawn in advance; then, the scratched cells were washed by PBS. A DCBM extract medium was used to culture the PBMSCs, while the normal medium was used in the control group. In addition, to see the effect of TGF-β3 on cell behaviors, a DCBM extract + TGF-β3 (concentration) group was set, in which 10 ng/ml TGF-β3 was added into the extract medium. All scratches were observed using a microscope at 0, 8, and 24 h after the intervention.

In order to evaluate the effect of DCBM on the differentiation of PBMSCs and the induction ability of TGF-β3 at the molecule level, we detected the deference of GAG content. After seeding, the cell-scaffold composites in two experimental groups were cultured in DMEM or DMEM with TGF-β3 (10 ng/ml), while an equal quantity of PBMSCs was seeding in a 12-well plate in the control group and DMEM was used to culture it. The mediums were changed every 2 days. After 3, 7 and 14 days of culture, the GAG in the scaffolds of the 3 groups were evaluated according to the GENMED quantitative detection kit. A previous study described the method and principle in detail (Barbosa et al., 2003).

To evaluate the effect of DCBM on the differentiation of PBMSCs and the induction ability of TGF-β3 at the gene level, we decided to detect the expression level of a cartilage-related gene using polymerase chain reaction (PCR). The experimental grouping and culture are the same as in biochemical assays for GAG. After 3, 7 and 14 days of culture, the cells were harvested, washed with PBS and blew with 1 ml Trizol for 5 min. After adding 200 μl of chloroform, shaking for 15 s, the liquid was centrifuged in 14 kg for 15 min. After centrifugation, 400 μl liquid in the upper aqueous layer was carefully aspirated and mixed with 400 μl isopropanol. After centrifuging the sample at 14 kg for 10 min 75% alcohol was used to dissolve the white feather-like RNA precipitation adhered to the bottom wall of the EP tube. After another centrifugation at 15 kg for 5 min, DEPC water was used to dissolve the RNA.

Reverse transcription was performed using Promega’s GoScript Reverse Transcription system and Bio-Rad’s DNA Engine PCR machine. Moreover, amplification was performed using a Roche LightCycler96 PCR instrument.

The detected gene and primer sequence are listed in Table 1 .

The preparation procedure is described in 2.2. After sterilization, the scaffold was washed in sterile PBS, and immersed in DMEM overnight. Then 10 μl of 2.5×10⁴/μl PBMSCs suspension was added to the DCBM scaffold every 10 min for 4 times. Cells were allowed to adhere for 4 h, and then, medium with 16% FBS and 10 ng/ml TGF-β3 was added to the plate. The plate was transferred to an incubator and cultured for 1 week, and the medium was changed every 2 days.

Thirty-two adult New Zealand white rabbits weighing about 2.5 kg were selected. After the medial meniscus was removed, the DCBM + PBMSCs complex and DCBM scaffolds were implanted in situ in right and left knee, respectively, in 16 rabbits. The rest 16 rabbits were treated with incision of the skin and joint capsule, but not the meniscus, in the right knee, as the sham-operated group, and the left knee with the meniscus removed as the control group. The cell inoculation method was the same as mentioned above. The amount of PBMSCs seeded on a single material is 10⁶. After cell seeding, the complex was cultured in DMEM with 16% FBS and 10 ng/ml TGF-β3 for 1 week and then implanted.

After anesthesia and routine preparation, the knee was approached through a medial parapatellar incision. A total meniscectomy was performed by resecting the medial meniscus sharply along the periphery and detaching it from its anterior and posterior junction. The medial collateral ligament, which is important for the postoperative stability of the knee joint, was reserved. Both the anterior and posterior horns and periphery of the scaffolds were reattached to the respective root attachments and appropriate adjacent synovium with absorbable No. 4-0 sutures. For the posterior root, attachment of the medial meniscus is adjacent to the posterior cruciate ligament. A self-made threading apparatus and extracapsular knot-ting technique were used to fix the posterior horn of the scaffold with the ligamentous structures. The joint capsule, periarticular tissue, and skin were closed with No. 3-0 Vicryl sutures. Figure 1A showed the picture of rabbits’ knee after meniscus removal, Figure 1B showed the pictures of rabbits’ knee after the meniscus graft implanted and Figure 1C showed the meniscus graft.

The animals were sacrificed at 12 and 24 weeks after the operation. The femur and tibia of the knee joint were dissected out, and the newly formed meniscus tissue was freed and photographed.

Nine aspects of meniscus fusion, position, anterior and posterior angle position, size, tear, surface, shape, tissue, and synovium were evaluated and scored (Chiari et al., 2006). A higher score indicated a worse prognosis.

In addition, the newly formed meniscus was prepared as a pathological section according to conventional procedures. HE staining, toluidine blue staining, safranin O staining, Sirius red staining, MASSON staining, and immunohistochemical staining of collagen types I and II were performed on the meniscus, and the semi-quantitative histological evaluation of the meniscus-like tissue was performed using the Ishida score (Ishida et al., 2007).

Similar to the meniscus, the cartilage was also evaluated in gross evaluation through International Cartilage Repair Society (ICRS) cartilage lesion classification (Mainil-Varlet et al., 2003) and in histology through the Mankin grading system (Baklaushev et al., 2019).

Adherent cells were observed after 5–7 days of primary cell culture. Among them, the spindle-shaped cells proliferated relatively quickly. After culturing for about 2 weeks, almost all cells were spindle-shaped, and they moved after 10–15 days. The colonies were formed and the cell morphology was uniformly spindle-shaped with time going by (Figures 2A,B).

Many calcium nodules were observed in osteogenic induced cells after alizarin red staining (Figure 2C). Oil red O staining showed that lipid-rich vesicles are formed in the cells after adipogenic differentiation (Figure 2D). Alcian blue staining showed a blue cell sphere after chondrogenic induction (Figure 2E). The results showed that PBMSCs could perform osteogenic, adipogenic, and chondrogenic differentiation under induction.

Flow cytometry analysis showed that the PBMSCs had a positive expression of CD29 and CD44 and a negative expression of CD34 and CD45 (Figure 2F).

HE staining showed no cell residues in the DCBM scaffold, and a large number of round or oval pore-like structures were retained in the material, mainly composed of hollows after decalcification and decellularization (Figure 3A). Masson staining showed similar collagen structure, void structure and distribution. The blue reaction mainly localized to the osteoid tissue and collagen fibers, and the red for lamellar bone formation (Figure 3B).

SEM scanning results showed that the surface of the DCBM scaffold is smooth (Figure 3C) with a few pores (Figure 3D), while the section of it is rough (Figure 3E) and full of pores (Figure 3F).

After decellularization, the DNA content of DCBM scaffold was 34.48 ± 12.14 ng/mg, which was significantly lower than before (200.75 ± 34.47 ng/mg) (p < 0.05) (Figure 3G).

The tensile test results showed that the tensile modulus of DCBM was 259.41 ± 90.88 MPa, while the tensile modulus of rabbit meniscus was 321.62 ± 89.09 MPa. The tensile modulus of DCBM scaffold was slightly lower than that of the rabbit meniscus; however, the difference is not statistically significant (p > 0.05) (Figure 3H).

The cell viability of the two groups evaluated by CCK-8 increased with time on the first 5 days. On the 6th to 7th days, cell proliferation entered a plateau phase. Compared with the normal medium group, the DCBM extract group showed similar cell viability at every assessing timepoint (Figure 4A).

The live/dead staining showed PBMSCs could adhere to the surface of the scaffold well, indicating good cell viability. Over time, the number of living cells has increased significantly. From 1st day to 3rd day, only a few dead cells were seen on the scaffold (Figure 4B).

The SEM results confirmed the live/dead staining results. On 1st day, a few cells can spread and adhere to the surface of the scaffold (Figure 4C). On 2nd day, the cells on the surface of the scaffold increase significantly (Figure 4D). On 3rd day, layers of PBMSCs were on the surface of the scaffold (Figure 4E).

Migration occurred in all three groups and the remaining scratch area was marked and calculated (Figure 4F). The migration speed of PBMSCs in the extract + TGF-β3 group was significantly faster than in the extract and control groups. At 24 h after the scratch, the area difference was statistically significant (p < 0.05). The left 2 groups showed no significant difference at each time point (p > 0.05) (Figure 4G). The results showed that the extract did not influence the migration of PBMSCs while the TGF-β3 promoted it.

At 12 weeks after the operation, the synovial limbus of the meniscus in the DCBM + PBMSCs group was closely connected with the joint capsule and the articular cartilage surface of the femoral condyle and the tibial plateau was smooth. So were the meniscus in the DCBM and sham operation groups. There was no obvious new tissue formation of the meniscus in the control group, and the cartilage surface of the femoral condyle and the tibial plateau was not smooth, with visible abrasion.

At 24 weeks after the operation, the new meniscus tissue in the DCBM + PBMSCs group was mature and similar to the normal meniscus. The synovial limbus was tightly connected with the joint capsule, the articular cartilage of the femoral condyle and tibial plateau was slightly torn, and no obvious osteophytes were found. In the DCBM group, the meniscus was closely connected with the joint capsule, the cartilage surface of the femoral condyle and the tibial plateau was not smooth, and a small amount of osteophyte was formed on edge. The meniscus in the sham operation group was normal meniscus shape, and the articular surfaces of the femur and tibia were normal. No new meniscus tissue was observed in the control group, and the cartilage surface of the femoral condyle and the tibial plateau was rough, with severe wear and obvious osteophytes (Figure 6A).

Quantitative evaluation of meniscus showed that at 12 weeks, the general meniscus score of the DCBM + PBMSCs group was 21.50 ± 1.66, while the score of the DCBM group was 22.13 ± 1.69. There was no statistically significant difference between the two groups (p > 0.05). However, scores of DCBM + PBMSCs and DCBM group were higher than those of the sham operation group (p < 0.05), indicating that at 12 weeks, the difference between the newly formed meniscus and the normal meniscus still existed. At 24 weeks, the scores in the DCBM + PBMSCs group and the DCBM group were 15.50 ± 1.66 and 18.13 ± 1.69, respectively. The scores of the DCBM + PBMSCs group were significantly lower than those of the DCBM group (p < 0.05), indicating that the newly formed meniscus in the DCBM + PBMSCs group had a better condition than that in the DCBM group (Figure 6B).

The overall ICRS scores showed that at 12 and 24 weeks after the operation, the scores in the DCBM + PBMSCs group and the DCBM group were lower than that of the control group (p < 0.05), but higher than that of the sham operation group (p < 0.05). At each time point, the DCBM + PBMSCs group had a lower score than the DCBM group (p < 0.05) (Figure 6C).

These results indicated that DCBM augmented with PBMSCs had a better meniscus repair and cartilage protection effect.

The HE staining of meniscus showed a small amount of chondroid-like cells in the DCBM + PBMSCs group and the DCBM group at 12 weeks after the operation, but the new collagen fibers were in a mess distribution. At 24 weeks after the operation, the meniscus fibrocartilage-like structure was formed in the DCBM + PBMSCs group, and its morphology was more similar to the native meniscus than the DCBM group.

Toluidine blue staining showed that at 12 weeks after the operation, the newly formed tissue in the DCBM + PBMSCs group was lightly colored, with no obvious difference between the medial and lateral parts. So was the DCBM group. At 24 weeks after the operation, the medial part of the tissue was stained deeply, as well as many cartilage lacuna structures were seen in the DCBM + PBMSCs group. The regional metachromaticity of the staining indicated that more cartilage-specific matrix GAGs were secreted medially. However, similar results did not appear in the DCBM group.

Immunohistochemistry of collagen I and II showed that at 12 weeks after the operation, collagen I in the DCBM + PBMSCs group and DCBM were strongly positive, and the expression of collagen II was only positive in the medial part. At 24 weeks after the operation, collagen I staining in the DCBM + PBMSCs group showed that the lateral side of the meniscus was strongly positive and the medial side was weakly positive. A similar result was not obvious in the DCBM group. The expression of collagen II in the medial part of the meniscus was stronger in the DCBM + PBMSCs group than in the DCBM group, and the expression of collagen II in the lateral part was negative in both groups.

The results of Sirius red staining showed that at 12 weeks after the operation, strong red or yellow refraction (collagen I fibers) and a small amount of green refraction (collagen III fibers) were seen in the newly formed meniscus in the DCBM + PBMSCs group. The DCBM group showed strong red refraction (collagen I fibers). At 24 weeks after the operation, the DCBM + PBMSCs group showed stronger refraction and the distribution of collagen fibers was similar to that of the normal meniscus. The DCBM group showed moderate-strength collagen I fibers refraction, while the distribution of collagen fibers was different from the fibrous arrangement of the normal meniscus (Figure 7A).

The Ishida score showed that the repair of meniscus in the DCBM + PBMSCs group was better than that in the DCBM group in both 12 (p < 0.05) and 24 weeks (p < 0.05). In addition, in both groups, the repair effect was better in 24 weeks than 12 weeks (p < 0.05). Ishida score in DCBM + PBMSCs group had no statistic difference with sham-opperated group in 24 weeks (Figure 7B).

HE staining of cartilage showed that there was no obvious cartilage degeneration at 12 weeks after the operation in the sham-operated group, the DCBM + PBMSCs group, and the DCBM group. In the DCBM group, a small number of chondrocytes were enlarged in morphology. In the meniscectomy group, the cartilage degenerated slightly, and the cartilage surface was uneven and cracked. At 24 weeks after the operation, the cartilage structure and cartilage thickness in the DCBM + PBMSCs group were nearly the same as those in the sham-operated group. The cartilage in the DCBM group and the meniscectomy group had different degrees of degeneration.

The results of toluidine blue staining showed that the DCBM + PBMSCs group and DCBM group were similar to the sham operation group at 12 weeks after the operation. The staining of the cartilage surface in the meniscectomy group became lighter, indicating that the cartilage matrix was damaged. At 24 weeks after the operation, the surface of the cartilage in the DCBM + PBMSCs group was slightly lighter, and the coloration of the cartilage in the DCBM group was significantly lighter.

The results of collagen Ⅱ staining showed that at 12 weeks after the operation, only the positive staining in the control group was unevenly distributed, and the surface layer was weakly positive. At 24 weeks after surgery, the results of the DCBM + PBMSCs group were similar to the sham-operated group, and both were uniformly positive. Cartilage in the DCBM group showed weak positivity with uneven distribution. The meniscectomy group had a negative result, and the articular cartilage thickness was significantly reduced (Figure 8A).

The Mankin score of cartilage histology results also showed that on 12 and 24 weeks after the operation, the femoral and tibia cartilage degeneration in the DCBM + PBMSCs group was more serious than that in the sham operation group (p < 0.05). However, it was significantly better than that in control group (p < 0.05) and DCBM group (p < 0.05).