Integrin a10b1-selected equine mesenchymal stem cells (EQSTEM) have previously shown to decrease cartilage degradation and bone sclerosis in an equine posttraumatic osteoarthritis (OA) model. Recently we demonstrated that human integrin a10b1-selected MSCs are able to home to a cartilage defect in rabbits and directly participate in the regeneration process. Here we investigated clinical outcome and mechanism of action of EQSTEM in a non-randomized, partially blinded case-control study using the carpal osteochondral fragment OA model. Day 18 after surgery, horses (females) either received EQSTEM (20x106 male MSCs) intra-articularly (treatment group, n=8) or were untreated (n= 9). Lameness (AAEP) was assessed, and synovial fluid sampled weekly up to 70 days after surgery. After euthanasia (Day 71) carpi were evaluated by CT and MRI, macroscopic pathology, and histology. MSCs were traced by qPCR of the Y-chromosome. Lameness (p=0.022) and response to flexion (p=0.004) were significantly improved over time after EQSTEM treatment. No joint flares were seen after the treatment. The treated horses had significantly less macroscopic cartilage pathology (p=0.019) and lower histological cartilage OA score (p=0.037). Treated horses showed a better healing of the osteochondral fragment on CT (p=0.07), although this was not significant. The immunomodulatory factors prostaglandin E2 (p=0.018), interleukin-10 (p=0.035) and interleukin-6 (p=0.08) increased in synovial fluid after treatment. MSCs were detected in the synovial fluid until day 35 and in the cartilage after euthanasia. EQSTEM holds a promising potential to be a Disease Modifying Osteoarthritis Drug (DMOAD) in treatment of OA in horses.