Skeletally mature Japanese White Rabbits weighing approximately 3.2 kg (ranging from 2.8 to 3.6 kg) were used in the experiments. Animal care was in accordance with the guidelines of the animal committee of Tokyo Medical and Dental University. The operation was performed under anesthesia induced by intramuscular injection of 25 mg/kg ketamine hydrochloride and intravenous injection of 45 mg/kg sodium pentobarbital.

Synovium with the subsynovial tissue was harvested from the left knee of the rabbits under anesthesia. The synovium was digested in a 3 mg/ml collagenase D solution (Roche Diagnostics, Mannheim, Germany) in αMEM (Invitrogen Corp., Carlsbad, CA, USA) at 37°C. After 3 hours, digested cells were filtered through a 70-μm nylon filter (Becton Dickinson, Franklin Lakes, NJ, USA), and the remaining tissues were discarded. The digested cells were plated at 5 × 10⁴ cells/cm² in 150 cm² culture dishes (Nalge Nunc International, Rochester, NY, USA) in complete culture medium, αMEM containing 10% FBS (lot selected for rapid growth of bone marrow derived MSCs, 100 units/ml penicillin, 100 μg/ml streptomycin, and 250 ng/ml amphotericin B; Invitrogen Corp.), and were incubated at 37°C with 5% humidified CO₂. After 3 to 4 days, the medium was changed to remove nonadherent cells, and the adherent cells were cultured for 7 days as passage 0 without refeeding. The cells were then trypsinized, harvested and resuspended to be used for transplantation. We already reported that these cells had characteristics of MSCs 202122.

The cells that were transplanted in animals to be sacrificed at day 1 were labeled for cell tracking by the fluorescent lipophilic tracer 1,1'-dioctadecyl-3,3,3',3'-tetramethylindocarbocyanine perchlorate (DiI) (Molecular Probes, Eugene, OR, USA). For labeling, the cells were resuspended at 1 × 10⁶ cells/ml in αMEM, and DiI was added at 5 μl/ml in αMEM. After incubation for 20 minutes at 37°C with 5% humidified CO₂, the cells were centrifuged at 450 × g for 5 min and washed twice with PBS 2023, and the cells were then resuspended in PBS for the transplantation.

Full-thickness osteochondral defects (5 mm × 5 mm wide, 3 mm deep) were created in the trochlear groove of the femurs of adult rabbits. The distal end of the femurs were then removed, and were precultured in serum-free Dulbecco's MEM (Invitrogen) supplemented with 100 units/ml penicillin (Invitrogen), 100 μg/ml streptomycin (Invitrogen), and 250 ng/ml amphotericin B (Invitrogen) for 24 hours. To determine the length of time needed for cell attachment to the defect, the cartilage defect of the femoral condyle was faced upward. Passage 0 autologous synovial MSCs, precultured for 7 days, were used for the transplantation.

The defect was filled with DiI-labeled synovial MSC suspension, which consisted of 10⁷ cells in 100 μl PBS, and was left stationary for 2.5, 5, 7.5, 10, and 15 minutes. The femurs were then turned with the defect side down for 10 minutes. This allowed the nonadhered cells in the defect to discard the defect in the culture medium (Figure 1a). The nonadhered cells in the medium were collected, as were the nonadhered cells attached to the dishes after trypsinization. The total number of nonadhered cells positive for DiI was counted. Finally, the adhered cell number attached to the cartilage defects was calculated by subtracting from 10 × 10⁶ cells.

Thirty-six rabbits were used for the in vivo transplantation study. Autologous synovial MSC transplantation was performed 7 days after the harvest. Under anesthesia, the right knee joint was approached through a medial parapatellar incision, and the patella was dislocated laterally. Full-thickness osteochondral defects (5 mm × 5 mm wide, 3 mm deep), whose size were critical for rabbit knees 24, were created in the trochlear groove of the femur.

The animals were divided into three groups for transplantation. For the control group, the cartilage defect was left empty. For the intra-articular group, 10⁷ DiI-labeled autologous synovial MSCs in 100 μl PBS were injected into the knee joint after the capsule was closed. For the local adherent group, the defect was filled with the cell suspension of 10⁷DiI-labeled autologous synovial MSCs in 100 μl PBS and held stationary for 10 minutes with the defect upward. In no groups were the defects patched, and a periosteum or artificial membrane was not used. All rabbits were returned to their cages after the operation and were allowed to move freely. Animals were sacrificed with an overdose of sodium pentobarbital at 1 day and 4, 12, and 24 weeks after the operation (n = 3 at each time point).

The cartilage defects were examined macroscopically for color, integrity and smoothness. Osteoarthritic changes and synovitis of the knee were also investigated. Macroscopic pictures of the femoral condyles were taken for evaluation using MPS-7 (Sugiura Laboratory Inc., Tokyo, Japan), a dedicated medical photography platform. Digital images were taken using a Nikon Coolpix 4500 digital camera (Nikon, Tokyo, Japan).

The dissected distal femurs were immediately fixed in a 4% paraformaldehyde solution. The specimens were decalcified in 4% ethylenediamine tetraacetic acid solution, dehydrated with a gradient ethanol series, and embedded in paraffin blocks. Sagittal sections 5 μm thick were obtained from the center of each defect and were stained with toluidine blue. Sections dedicated for fluorescent microscopic visualization of DiI-labeled cells were not stained with toluidine blue, and nuclei were counterstained by 4',6-diamidino-2-phenylindole dihydrochloride.

Histological sections of the repaired tissue were analyzed using a grading system consisting of five categories (cell morphology, matrix staining, surface regularity, cartilage thickness, and integration of donor with host), which were modified from the repaired cartilage score described by Wakitani and colleagues 25, so that overly thick regenerated cartilage could not be overestimated (Table 1). The scoring was performed in a blinded manner by two observers, and there was no significant interobserver difference.

The study was approved by our Institutional Review Board, and informed consent was obtained from all study subjects. Human synovium and cartilage were harvested during total knee arthroplasty with medial compartment osteoarthritis. Synovial tissue was minced into small pieces, digested in a collagenase solution, and then filtered. Nucleated cells were cultured for 14 days. Passage 3 cells were used for further analyses 15.

Osteochondral fragments at the lateral femoral condyle were diced with a bone saw. The cartilage defects 2.5 mm in diameter were created and filled with 800 × 10³ DiI-labeled human synovial MSCs in 8 μl PBS. After 5, 10, 20, and 30 minutes, the cartilage defects were turned down for 10 minutes. After trypsinization, the DiI-positive cells in the dish were counted, and number of the cell attached to the cartilage defects was calculated by subtracting from 800 × 10³ cells.

The sections of the human osteochondral fragments were deparaffinized, washed in PBS, and pretreated with 0.4 mg/ml proteinase K (DAKO, Carpinteria, CA, USA) in Tris–HCl buffer for 15 minutes at room temperature. Endogenous peroxidases were quenched using 3% hydrogen peroxide in methanol for 20 minutes at room temperature. The sections were rinsed three times in PBS for 5 minutes and were briefly blocked with 5% normal horse or rabbit serum (Vector Laboratories, Burlingame, CA, USA) to avoid nonspecific binding of the antibody. The sections were then incubated in mouse monoclonal anti-human intercellular adhesion molecule 1 (ICAM-1) antibody (1:50 dilution; SANBIO BV, Uden, Netherlands) or in goat anti-human vascular adhesion molecule 1 (VCAM-1) antibody (1:100 dilution; R&D Systems, Wiesbarden, Germany) at room temperature for 1 hour. After rinsing in PBS, the tissues were incubated with biotinylated horse anti-mouse or rabbit anti-goat IgG secondary antibody (Vector Laboratories) for 30 minutes at room temperature. After incubation for another 30 minutes with Vectastain ABC reagent (Vector Laboratories), the slides were counterstained with Mayer hematoxylin, dehydrated, and mounted in a xylol-soluble mount (Vitro-Club; Langenbrinck, Emmendingen, Germany).

Three million DiI-labeled human synovial MSCs were incubated in 2 ml PBS including 1% BSA with 10 μg/ml neutralizing antibody for human ICAM-1, VCAM-1, activated leukocyte-cell adhesion molecule, or mouse IgG1 isotype control antibody (R&D Systems) for 30 minutes at 37°C with 5% humidified CO₂ 26. After the supernatant was discarded, 800 × 10³ cells resuspended in 8 μl PBS were placed on the cartilage defect of osteocartilage fragment and held stationary for 10 minutes. The cartilage defects were then turned down for 10 minutes.

Human synovial MSCs at 500 × 10³ in 10 μl PBS were placed on eight-well chamber glass slides (BD Bioscience) and washed by PBS at 1 minute and 10 minutes, and were then fixed with 99.5% acetone for 15 minutes. The glass slides were stained with mouse monoclonal anti-human ICAM-1 antibody (1:10 dilution with PBS in 5% goat serum; R&D Systems) for 2 hours. After rinsing with PBS three times, the slides were stained with goat anti-mouse IgG secondary antibody labeled with Alexa fluor 568 (Invitrogen) for 1 hour. The nuclei were stained with Hoechst 33342 (Invitrogen). The number of ICAM-1-positive cells and nuclei was counted in three high-power fields.

To assess differences, the Kruskal–Wallis test and the Mann–Whitney U test were used. P < 0.05 was considered significant.

To clarify the minimum time for an adequate number of synovial MSCs to attach to the cartilage defect by the local adherent technique, we performed an ex vivo sequential analysis using rabbit synovial MSCs and rabbit cartilage (Figure 1a). The number of attached cells increased in a time-dependent manner, and more than 60% of the cells attached in 10 minutes (Figure 1b).

Osteochondral defects were created in rabbit knees. For the control group, the cartilage defect was left empty. For the intra-articular group, synovial MSCs were injected into the knee joint after the capsule was closed. For the local adherent group, the defect was filled with the synovial MSC suspension and faced upward for 10 minutes according to the ex vivo analyses.

At 1 day, the cartilage defects were overlaid with blood clots, and there seemed to be no obvious differences among the control, intra-articular, and local adherent groups macroscopically (data not shown).

At 4 weeks, the cartilage defect in the control group still showed reddish tissue (Figure 2b, image a). In the intra-articular group, the defect was covered with whitish tissue in some areas, but the reddish area remained in other areas (Figure 2b, image b). In the local adherent group, the defect became whitish and glossy in the entire area (Figure 2b, image c).

At 12 weeks, in the control and intra-articular groups, the reddish regions decreased in size but still remained locally (Figure 2b, images d and e). In the local adherent group, the border between repaired tissue and neighboring cartilage appeared less distinct (Figure 2b, image f).

At 24 weeks, the cartilage defect area in the control group decreased but still remained (Figure 2b, image g). In the intra-articular group, the defects were covered with whitish tissue but the margins were still distinct (Figure 2b, image h). In the local adherent group, the peripheral lesion of the defect appeared to integrate into the surrounding native cartilage (Figure 2b, image i).

In all three groups there were no obvious features of hydrarthrosis or synovial proliferation. Mild spur formation was observed on the edge of the trochlear groove of the femur in some samples of the control group, but there were no osteoarthritic changes of the femorotibial joint in any groups.

At 1 day, the defect in the control group was filled with blood clots (Figure 3a, images a and b). In the intra-articular group, DiI-positive synovial MSCs were observed in the defect (Figure 3a, images c and d); the cells were very sparse when examined at higher magnification, however, even in the selected area where relatively dense DiI-positive cells were observed in lower magnification (Figure 3a, images e and f). In contrast, in the local adherent group, there were more DiI-positive synovial MSCs along with the osteochondral defect, with the cellular layer 20 cells deep (Figure 3a, images g and h). DiI-positive cells were denser in the local adherent group (Figure 3a, images I and j) than in the intra-articular group.

At 4 weeks, the defect in the control group was filled with fibrous tissue and the cartilage matrix formation was poor (Figure 3b, image a). In the intra-articular group, although more cartilage matrix could be observed than in the control group, the height of the repaired tissue was lower than that of the surrounding cartilage (Figure 3b, image b). In the local adherent group, the defect was filled with abundant cartilage matrix. In addition in the local adherent group, remodeling of the cartilage into the underlying bone was observed in deep areas (Figure 3b, image c).

At 12 weeks, in the control and intra-articular groups, the defects were filled with fibrous tissues and were poorly healed (Figure 3b, images d and e). In the local adherent group, the cartilage matrix at the defect still remained, and the border between regenerated cartilage and subchondral bone moved upward. Integration between native cartilage and regenerated tissue appeared to be improved (Figure 3b, image f).

At 24 weeks, in the control and intra-articular groups, the cartilage defects were still not healed (Figure 3b, images g and h). In the local adherent group, the regenerated cartilage matrix was well developed. The subchondral bone moved further upward, and the thickness of the regenerated cartilage was similar to that of the neighboring cartilage. The borders between the native and regenerated tissue were well integrated (Figure 3b, image i).

The histological scores of the local adherent group improved continuously through 24 weeks and were always better than those of the control group and the intra-articular group at each point (Figure 3c).

The results described above were obtained using rabbit MSCs. We investigated whether human MSC exhibited the same capacity as rabbit cells to adhere to cartilage with the same kinetics. The defects of cartilage obtained from humans were faced upward, filled with 800 × 10³ DiI-labeled human synovial MSCs, and the position maintained for 5 to 30 minutes.

Macroscopically, the cartilage defect looked yellowish at time 0, slightly reddish at 5 minutes, and red at 10 minutes and thereafter (Figure 4a). The cell number attached to the cartilage defect increased rapidly at 5 minutes, and then started to rise slowly (Figure 4b). It should be noted that more than 60% of the human synovial MSCs already adhered to the cartilage defects at 10 minutes, indicating similarity between rabbits and humans.

It is expected that adhesion molecules are involved in cell attachment. Ten minutes after filling human synovial MSCs in human cartilage defect, adhered cells expressed ICAM-1 and VCAM-1 (Figure 4c). Neutralizing antibodies for ICAM-1, VCAM-1, and activated leukocyte-cell adhesion molecule, separately or together, inhibited attachment of human synovial MSCs to human cartilage defects (Figure 4d). When human synovial MSCs were plated on grass slides, ICAM-1-positive cells significantly increased between 1 minute and 10 minutes (Figure 5a,b).

We finally examined the morphological change of human synovial MSCs during a 10-minute period after plating on a culture dish. Most cells looked thick and round at 1 minute. They became thinner, larger, and polygonal at 10 minutes (Figure 5c).