Current repair of articular cartilage (AC) often leads to a lower quality tissue with an unstable hypertrophic phenotype, susceptible to endochondral ossification and development of osteoarthritis. Engineering phenotypically stable AC remains a significant challenge in the cartilage engineering field. This motivates new strategies inspired from the extracellular matrix proteins unique to phenotypically stable AC. We have previously shown that the bone morphogenetic protein antagonist gremlin-1 (GREM1) protein, present in permanent but not transient cartilage, suppresses the hypertrophy of chondrogenically primed bone marrow stem cells (BMSCs) in pellet culture. The goal of this study was to assess the effect of GREM1 on the in vitro and in vivo phenotypic stability of porcine BMSC-derived cartilage engineered within chondro-permissive scaffolds. In addition, we explored whether GREM1 would synergize with physioxia, a potent chondrogenesis regulator, when engineering cartilage grafts. GREM1 did not influence the expression of chondrogenic markers (SOX-9, COL2A1), but did suppress the expression of hypertrophic markers (MMP13, COL10A1) in vitro. Cartilage engineered with GREM1 contained higher levels of residual cartilage after 4 weeks in vivo, but endochondral bone formation was not prevented. Higher GREM1 levels did not significantly alter the fate of engineered tissues in vitro or in vivo. The combination of physioxia and GREM1 resulted in a higher sulfated glycosaminoglycan deposition in vitro and a greater retention of cartilage matrix in vivo than physioxia alone, but again did not suppress endochondral ossification. Therefore, while physioxia and GREM1 regulate BMSC chondrogenesis in vitro and reduce cartilage loss in vivo, their use does not guarantee the development of stable cartilage. Impact Statement A major challenge associated with bone marrow stem cell (BMSC)-derived cartilage is that the chondrocyte-like cells have a tendency to undergo hypertrophic differentiation, yielding tissue unsuitable for use in hyaline articular cartilage (AC) repair. This is motivating the development of new tissue engineering strategies to generate phenotypically stable chondrocyte-like cells from BMSCs. In this study, we aimed to engineer phenotypically stable cartilage grafts using BMSCs seeded onto solubilized AC extracellular matrix-derived scaffolds and treated with the bone morphogenetic protein antagonist gremlin-1. This article describes the effects of potential therapeutic strategies that could improve the phenotypic stability of chondrogenically differentiated BMSCs, and examined the use of this strategy both in vitro and in vivo.