The purpose of this study was to investigate the efficacy of cartilage regeneration when using a mixture of transforming growth factor-beta1 (TGF-beta1)-producing human chondrocytes (hChon-TGF-beta1) and primary human chondrocytes (hChon) ("mixed cells"), compared with either hChon-TGF-beta1 or hChon cells alone. Specifically, mixed cells or hChon cells were first injected intradermally into the backs of immune-deficient nude mice to test the feasibility of cartilage formation in vivo. Both the mixed cells and the hChon-TGF-beta1 cells alone induced cartilage formation in nude mice, whereas hChon cells alone did not. To further test the efficacy of the cells in generating cartilage, an artificially induced partial thickness defect of the femoral condyle of a rabbit knee joint was loaded with hChon-TGF-beta1 cells with or without mixing additional untransduced hChon cells, and hyaline cartilage regeneration was observed at 4 or 6 weeks. The efficiency of complete filling of the defect and the quality of tissue generated after implanting were evaluated on the basis of a histological grading system modified from O'Driscoll et al. (J. Bone Joint Surg. 70A, 595, 1988). Significantly, mixed cells (14.2 +/- 0.9) produced significantly better results than hChon-TGF-beta1 (9.0 +/- 1.7) or hChon (8.0 +/- 1.8) cells alone. Histological and immunohistochemical staining of the newly repaired tissues produced after treatment with either mixed cells or hChon-TGF-beta1 cells alone showed hyaline cartilage- like characteristics. These results suggest that the implantation of mixed cells may be a clinically efficient method of regenerating hyaline articular cartilage.