In this study, to form aggregates of synovial MSCs, the hanging drop technique was used [8
]. This is a simple method; expensive or specific tools are not required. Three days after cultured in the drop, the aggregate, consisting of 250,000 MSCs, became approximately 1 mm in diameter, large enough to be visible and solid enough to aspirate with a pipette. Aggregates of MSCs sank faster in the suspension medium than dispersed MSCs and helped to avoid loss of MSCs from targeted cartilage defect. The use of aggregates was practically convenient for transplantation of MSCs.
In the previous report, the number of apoptotic or necrotic cells was greater in aggregates prepared with 100,000 or 250,000 human bone marrow MSCs, which was examined by flow cytometry, measuring propidium iodide uptake and annexin V labeling [10
]. We examined the viability of aggregates of MSCs by TUNEL staining and confirmed that cells positive for TUNEL staining were observed; the number was small compared to the previous report. This difference may have been due to the difference of methods. Microarray analysis showed up-regulation of genes with ontology for regulation of cell death. The microarray data are available at the public database (GEO accession# GSE 31980). These results suggest that aggregation of 250,000 MSCs affect the viability of cells. However, we thought that aggregates of MSCs could be used as a source for cartilage regeneration because most cells which are cultured in drops for three days are viable.
Aggregation of synovial MSCs changed the gene expression profile dramatically without any special tools or chemical factors. This is possibly due to environmental changes, including cell-to-cell contact, hypoxic condition and low nutrient condition. Aggregation of human synovial MSCs increased expressions of several chondrogenesis-related genes and the most up-regulated gene was BMP2, which was also up-regulated in bone marrow MSCs [8
In this study, we compared in vitro
chondrogenesis potential of synovial MSCs after hanging drop culture with that of MSCs after monolayer culture. We used 1,000 ng/ml BMP7
for in vitro
chondrogenic differentiation assay. We previously examined the dose effect of BMP6 between 0 to 500 ng/ml for in vitro
chondrogenesis of bone marrow MSCs. Cartilage pellets increased in size along with the concentration of BMP6
, and a maximal effect was at 500 ng/ml [17
]. Our preliminary experiments showed that 1,000 ng/ml BMP6
induced larger cartilage pellets than 500 ng/ml BMP6
in bone marrow and synovial MSCs. We obtained similar results with BMP7
. Real time RT-PCR analysis showed higher expression levels of COL2A1
, aggrecan and SOX9
for pellets derived from MSC-aggregates after hanging drop culture compared to those of MSCs in a monolayer culture. Furthermore, the wet weight of pellets derived from MSC-aggregates after hanging drop culture was heavier than that of pellets derived MSCs in a monolayer culture. These indicate that chondrogenic potential increased in aggregates of MSCs after hanging drop culture.
In this study, we used an osteochondral defect model of rabbits, which have a higher, self-renewal capacity than bigger animals and humans. Therefore, the results obtained here should be critically evaluated. However, we prepared negative controls, which healed poorly at 4 and 12 weeks. We previously confirmed that the osteochondral defect created in the trochlear groove of the femur, similar to this study, was not repaired without any treatments 24 weeks after surgery [6
]. These findings indicate that this rabbit model is useful to evaluate the effects of the treatments for cartilage regeneration.
For in vivo
analysis of cartilage regeneration by transplantation of aggregates of synovial MSCs in rabbits, successful cartilage regeneration was observed in the cases of a relatively small number of transplanted aggregates of MSCs, and the worst results were observed when the highest number of aggregates of MSCs was transplanted. These results were not what we expected, because we previously reported that better cartilage regeneration was obtained when higher cell densities of MSCs were embedded in collagen gel [3
Why were poor results obtained when more than a certain number of aggregates were transplanted? We listed three possible reasons. First, nutrition to maintain transplanted MSCs was depleted and the environment around transplanted MSCs worsened when too many aggregates were transplanted. As shown in Figure , in the case of 40 aggregates that were cultured for seven days in a well of 96-well plates, medium color changed to yellow. This means that adjustment of pH could not be controlled. Second, TUNEL positive cells increased when too many aggregates were transplanted. The number of TUNEL positive cells was higher when too many aggregates were transplanted (Figure ) than before transplantation (Figure ) and after a suitable number of aggregates were transplanted (Figure ). Third, transplantation of too many aggregates prevented chondro-progenitor cells from moving to the osteochondral defect from bone marrow and from synovial fluid.
We confirmed that transplanted aggregates of synovial MSCs were directly differentiated into chondrocytes by transplanting MSCs derived from GFP transgenic rabbit. This result suggests that aggregates of synovial MSCs were involved in the reparative process. However, as shown in Figure , in the case of aggregates of GFP positive MSCs being transplanted, regenerated cartilage consisted of both GFP positive cells and GFP negative cells. MSCs existed in synovial fluid [18
] and these MSCs contributed to the repair of cartilage injury [6
]. These results suggest that some host MSCs were also involved in the reparative process. In addition, host MSCs may have been involved in the anti-inflammatory process. In our rabbit osteochondral defect model, inflammation like a synovitis was not severe even in the control group. Therefore, we could not confirm the anti-inflammatory effect of MSCs. It would be interesting to investigate the anti-inflammatory effect of transplantation of aggregates of synovial MSCs and host MSCs in other arthritis models.
As previously reported, in bone marrow MSCs [10
], aggregates of human synovial MSCs expressed anti-inflammatory genes TSG6
is secreted by synoviocytes, mononuclear cells and chondrocytes under inflammatory conditions and has an anti-inflammatory effect. Overexpression of TSG6
or administration of recombinant TSG6
inhibited inflammation and joint destruction in a murine collagen induced arthritis model [20
is reported to have an anti-apoptotic effect as well as an anti-inflammatory effect [24
]. However, their roles in joint homeostasis are unknown.
In this study, transplantation of low numbers of aggregates, in other words, low density of aggregates to the volume of the cartilage defect, showed better regeneration (Figures and ). This is favorable for clinical application. We have performed clinical trials of autologous human synovial MSCs transplantation for cartilage defects. In the experiences of 12 patients, approximately 50 million synovial MSCs at passage 0 were transplanted for approximately 280 mm2 cartilage defects (unpublished data). In a rabbit model, we transplanted synovial MSC-aggregates into the osteochondral defects without any loss of cells, and 10 MSC-aggregates (2.5 × 106 cells) per 25 mm2 defects were needed for better cartilage regeneration. According to these data, we can prepare a sufficient amount of human synovial MSCs at passage 0.
In this study, we did not use scaffolds for transplantation of aggregates of synovial MSCs. We were able to adhere aggregates of synovial MSCs on the osteochondral defect without scaffolds; however, the use of scaffolds or materials to improve survival of transplanted cells is attractive. One of the methods is the use of a fibrin glue, which has an effect of improving survival of transplanted cells [26
]. In addition, cell transplantation of MSCs with a fibrin glue can probably be performed under arthroscopic surgery. Further studies are needed to improve cell transplantation procedures.