Abstract Background There are no proven therapies that modify the structural changes associated with osteoarthritis (OA). Preclinical data suggests that intra-articular recombinant human BMP-7 (bone morphogenetic protein-7) has reparative effects on cartilage, as well as on symptoms of joint pain. The objective of this study was to determine the safety and tolerability as well as dose-limiting toxicity and maximal tolerated dose of intra-articular BMP-7. The secondary objectives were to determine the effect on symptomatic responses through 24 weeks. Methods This was a Phase 1, double-blind, randomized, multi-center, placebo-controlled, single-dose escalation safety study consisting of 4 dosing cohorts in participants with knee OA. Each cohort was to consist of 8 treated participants, with treatment allocation in a 3:1 active (intra-articular BMP-7) to placebo ratio. Eligible participants were persons with symptomatic radiographic knee OA over the age of 40. The primary objective of this study was to determine the safety and tolerability of BMP-7 including laboratory assessments, immunogenicity data and radiographic assessments. Secondary objectives were to determine the proportion of participants with a 20%, 50%, and 70% improvement in the WOMAC pain and function subscales at 4, 8, 12, and 24 weeks. Other secondary outcomes included the change from baseline to 4, 8, 12, and 24 weeks for the OARSI responder criteria. Results The mean age of participants was 60 years and 73% were female. All 33 participants who were enrolled completed the study and most adverse events were mild or moderate and were similar in placebo and BMP-7 groups. The 1 mg BMP-7 group showed a higher frequency of injection site pain and there was no ectopic bone formation seen on plain x-rays. By week 12, most participants in both the BMP-7 and placebo groups experienced a 20% improvement in pain and overall the BMP-7 group was similar to placebo with regard to this measurement. In the participants who received 0.1 mg and 0.3 mg BMP-7, there was a trend toward greater symptomatic improvement than placebo. The other secondary endpoints showed similar trends including the OARSI responder criteria for which the BMP-7 groups had more responders than placebo. Conclusions There was no dose limiting toxicity identified in this study. The suggestion of a symptom response, together with the lack of dose limiting toxicity provide further support for the continued development of this product for the treatment of osteoarthritis.