RESULTS In vitro study of human OA chondrocytes  Chondrocytes from OA patients were isolated and cultured in vitro (Figure 1D). The related gene expression of chondrocytes was analyzed by real-time quantitative PCR (Figure 1E?H). The results showed that chondrocyte marker genes Col II, Nkx3.2, RUNX2 and Sox9 in OA chondrocytes were significantly downregulated, indicating the degenerative status of OA chondrocytes. In addition, the expression level of Runx2 was significantly increased in OA chondrocytes. Therefore, the OA chondrocytes presented an inflammatory, degenerative, hypertrophic status. Effects of p65shRNA and BMP4 transfection on human OA chondrocytes OA chondrocytes were transfected by AAV containing p65shRNA or BMP4. The success of the transfection was confirmed by real-time quantitative PCR and western blot/ELISA. The results show that p65 was significantly downregulated in p65shRNA transfected cells and significantly upregulated in BMP4 transfected cells. When AAV-BMP4 was implanted with gelfoam into the muscle pocket of nude rat, more bone formation could be found on X-ray, which was further confirmed by micro-CT analysis (Figure 2A?H). Combination effect of BMP4 and p65 and optimization of the mixed ratio in vitro The cell numbers were the same for the combined group and the three unmixed cell groups, which were composed of non-transfected OA chondrocytes, p65shRNA trans-fected cells, and BMP4 transfected cells. On days 14, 21, and 28, cell culture medium and cell pellets were collected for analysis.  The concentrations of IL-1β, IL-6, and IL-10 in the culture medium were measured to reveal the modulation of inflammation. As shown in Figure 2C (day 21), OA chondrocytes presented a pro-inflammatory status with high levels of IL-1β and IL-6, but a low level of IL-10. Compared to the non-transfected OA chondrocytes, the p65shRNA transfected cells presented an improved inflammatory status. For the combined group, the pro-inflammatory factors were significantly decreased, but slightly higher than the p65shRNA alone group. Thus, transfection with both p65shRNA and BMP4 could suppress the inflammatory reaction of OA chondrocytes. The combined group showed slightly less suppression than the p65shRNA group, which resulted from the proportion of BMP4 transfected cells.  The effects of p65shRNA and BMP4 on OA chondrocytes were further analyzed based on the promotion of anabolic activity. Growing from the same number of cells, the cell pellets in different groups clearly had various sizes after 21 days (Figure 2A). The non-transfected OA chondrocytes formed very small aspheric pellets, while the transfected cells formed large pellets. The combined group was the largest among the cell groups. These observations indicate that combined p65shRNA and BMP4 transfection could promote the proliferation or extracellular matrix (ECM) synthesis of OA chondrocytes, and the combination of the two viruses exert a synergistic effect. The amount of ECM synthesized in each cell group was then assessed by Alcian blue, Safranin O, and type II collagen immunohistochemical staining (Figure 2B?E). The non-transfected control group had the lightest staining. In p65shRNA and BMP4 transfected cells, the staining strength increased, and the combined group had the heaviest staining. The quantified statistical analysis of staining (Figure 2E) shows that both p65shRNA and BMP4 transfection increased the ECM synthesis activity of OA chondrocytes.  After confirming the significant effect of the combined strategy, we further optimized the repair effects by altering the mixture ratio of the two transfected cells. We tested five different mixture combinations (Figure 3). After 21 days, cell pellets in different groups demonstrated various sizes; the 1:1 group had the largest sizes, the 5:1 and 1:5 groups had the second largest, and the 50:1 and 1:50 groups had the smallest (Figure 3H). With Alcian blue and Safranin O staining, the 1:5 group showed the heaviest staining, while the 1:1 group showed the second heaviest (Figure 3H). This indicated that BMP4 promoted the synthesis of proteoglycan. For type II collagen immunohistochemistry staining, the 1:1 group had significantly heavier staining than the others, indicating that the presence of p65shRNA BMP4 promoted the synthesis "

of type II collagen, which is a marker of hyaline cartilage formation. The statistical analysis (Figure 3I?K) confirmed our observations of the staining images. BMP4 contributed more to ECM synthesis than p65shRNA. However, with the same extent of inflammation modulation by p65shRNA, BMP4 synthesized the largest amount of type II collagen. Therapeutic effects of combined AAV-p65shRNA and AAV-BMP4 injection in the rat OA model in vivo The synergistic effect of AAV-p65shRNA and AAV-BMP4 in vivo was also tested by injecting the processed virus into the joints of SD rats with OA. As mixing an equal amount of p65shRNA and BMP4 transfected OA chondrocytes presented the best re-generative effects in vitro, we injected equal amounts of AAV-p65shRNA and AAV-BMP4 virus particles (0.5×1010) into the rat joints. The rats were sacrificed 12 weeks after injection, and hind limb knee joints were harvested. By gross observation, we found that the OA control rats had severely destroyed joints (Figure 4). Injection of AAV-p65shRNA virus particles (1×1010) greatly inhibited the inflammation but did not generate more cartilage when compared to the BMP4 and combined groups. With AAV-BMP4, there was regenerated hyaline-like cartilage. For the combined group, joints showed less inflammation and obvious hyaline-like cartilage formation. Therefore, the same as the in vitro results, the application of both viruses combined also presented the strongest regenerative potential in vivo.  We further performed histological staining of the cartilage specimen (Figure 4). It showed that the OA rats had no cartilage left. The rats injected with virus showed obvious"

"

cartilage regeneration, and the combined group had the most robust ECM synthesis. At higher magnification, many chondrocytes were found to be distributed in the cartilage layer in samples from the AAV-BMP4 and combined groups; only the combined group had chondrocytes distributed all around the whole cartilage layer. Twelve weeks after virus injection, robust cartilage regenerative potential was observed in all virus-injected groups; the combined group had the best effect and BMP4 exhibited a better repair effect than p65shRNA."

"