Novel organ culture model for a complete synovial joint: creation and application

Abstract

Disorders affecting articular cartilage are amongst the most common problems in orthopaedics. Osteoarthritis, the end stage of the disease of articular cartilage, reduces the quality of life for tens of millions of people in the world, and has a profound impact on the economics of industrialized countries. Despite progress in articular cartilage research, the problem is still far from being defeated. Various models e.g. in vitro cartilage explants and in vivo animal models, have been established for cartilage research, but each has its own limitations. Thus, a novel ex vivo isolated joint organ culture model was developed. Bovine metatarsophalangeal joints were chosen as a suitable synovial joint because it consists of a hinge-type joint that is similar to the human knee joint, and has a large cartilage surface that provides enough space for multiple sampling in the same joint. The joints were isolated aseptically and placed into culture media. The viability of chondrocytes, glycosaminoglycan (GAG) content of cartilage matrix, cartilage morphology and water content of matrix were evaluated under different culture conditions, i.e. static, static with flowing media, and dynamic with different durations of the movement period. The model was used to investigate the effect on the sharp scalpel cartilage injury of adding serum to the culture medium by culturing the whole joint explants in serum-supplied or serum-free media. The feasibility of investigating the early phases of chondrocyte implantation in this model was also studied: circular holes of 2.5 mm diameter were created by making a pilot hole with a 2.0 mm drill followed by using a fresh 2.5 mm biopsy punch. Allogeneic isolated chondrocytes at different passages were aggregated as cell pellets and implanted in the holes to evaluate their integration ability and the response from the recipient cartilage. Results from the static model showed that, after 28 days culture, the chondrocytes were still alive with 66.5%, 80.9% and 46.9% viability in the superficial, middle and deep zones, respectively. The GAG content of the static model decreased 19.2% after the first week of culture and then lost another 15.0% during the third week. Paradoxically, at end of the 4th week the GAG level rebounded to some extent and increased 19.0% relative to the previous week. Interestingly, the cell viability of all three zones improved if the culture fluid was flowing as seen with the experiments carried out with stirred media or dynamic movement of the articular surfaces. (e.g. for the stirred media after 28 days of culture the chondrocyte viability was 80.6%, 92.4% and 70.4% for the superficial, middle and deep zones respectively.) The GAG content was maintained at a constant level in the contact area of the dynamic model, but decreased as in the media-stirred model and non-contact area of the dynamic model to a similar extent to that observed with the static model. In the injury model, the GAG content fell approximately 10.8% straight after the scalpel cut, but no further loss was observed if the joint was cultured in the serum-supplied media. In contrast, if the injured joint was cultured in the serum-free media, the GAG content continued to fall week by week and finally dropped by 41.7% at the end of the 4th week. In the chondrocyte implantation model, the majority of the host chondrocytes around the circular defect were alive (78.5 % viability). Viewed from the surface, the dead cells were all within 20 μm from the cut edge. The implanted chondrocytes, which were aggregated as cell pellets, began to transform their shapes and spread to the surrounding surface of the recipient cartilage, but did not appear to integrate with the host tissue during the first 2 weeks of culture. The results supported the validity of this ex vivo joint model and demonstrated that the chondrocytes subjected to flow of the media or dynamic loads survived well over a 4 week period. Of importance was the finding that there was no measured loss of the matrix GAG content when the joints were under dynamic load compared to all of the non-loaded conditions. This whole joint model could be of value in providing a more natural and controllable platform where research involving the normal processes or pathologic mechanisms of articular cartilage can be investigated, as well as the early response to newly developed pharmacological agents and cartilage tissue engineering constructs.