The repair of deep osteochondral joint surface defects represents a significant unmet clinical need. Importantly, untreated lesions lead to a high rate of osteoarthritis. The current strategies to repair these defects include osteochondral autograft transplantation or "sandwich" strategies combining bone autografts with autologous chondrocyte implantation, with poorly documented long-term outcomes. In this study, we first investigated the capacity of juvenile osteochondral grafts (OCGs) to repair osteochondral defects in skeletally mature rats. With this regenerative model in view, we produced a new biological, bilayered and scaffold-free Tissue Engineered construct (bTEC) for the repair of a deep osteochondral defect of the rat knee. Cylindrical OCGs were excised from the femoral intercondylar groove of the knee of skeletally immature rats (5 weeks) and transplanted into osteochondral defects created in skeletally mature rats (11 weeks). To create bTECs, micromasses (μMasses) of human periosteum-derived progenitor cells (hPDCs) and human articular chondrocytes (hACs) were produced in vitro using previously optimized chemically defined medium formulations containing growth and differentiation factors including bone morphogenetic proteins. These two μMass types were subsequently implanted as bilayered constructs into osteochondral defects in nude rats. At 4 and 16 weeks after surgery, the knees were collected and processed for subsequent 3D imaging analysis and histological evaluation. Micro-computed tomography (μCT), H&E, and Safranin O staining were used to evaluate the degree and quality of tissue repair. The osteochondral unit of the knee joint in 5 weeks old rats exhibits an immature phenotype, displaying active subchondral bone formation through endochondral ossification and the absence of a tidemark. When transplanted into skeletally mature animals, the immature OCGs resumed their maturation process, i.e., formed new subchondral bone, established the tidemark, and maintained their Safranin O-positive hyaline cartilage at 16 weeks after transplantation. The bTECs (hPDCs + hACs) could partially recapitulate the biology as seen with the immature OCGs, including the formation of the joint surface architecture with typical zonation, ranging from non-mineralized hyaline cartilage in the superficial layers to a progressively mineralized matrix at the interface with a new subchondral bone plate. Cell-based TE constructs mimicking immature OCGs and displaying a hierarchically organized structure comprising of different tissue forming units seem an attractive strategy to treat deep osteochondral defects of the knee.