Similar to our findings, some authors reported accelerated rejection of graft cells in mice when previously injected with allogenic MSCs in a model of bone marrow transplant. In another model of corneal rejection using two strains of mice and systemic injection of human MSCs, the MSCs prevented rejection by aborting the early inflammatory response. Discrepancies between these studies and ours might be caused by differences in the experimental parameters, including the animal model, the time of injection, the immunosuppression, and the mode of delivery, among others. All the above studies have been performed in normal- or low-risk transplant settings. As stated in the introduction, the primary problem to be solved in corneal transplantation involves high-risk recipients. In our high-risk transplant model, in which immune ocular privilege was undermined by the induction of neovascularization prior to graft surgery, we found that the use of systemic rabbit AD-MSCs prior, during, and at different time points after the surgery resulted in a lower survival rate of the graft compared with non-treated corneal grafts. Similarly, other authors, in a pig to rat model more similar to a high-risk human transplant, found that rat allogeneic MSCs administered topically did not prolong the survival of the corneal graft. In our high-risk study, intravenously administered MSCs were not found in the cornea, so these cells must ASP1517 systemically produce some factors that accelerate graft rejection. Similar to our results, some authors have reported a near absence of MSCs in the cornea after being systemically injected. Again, angiogenic factors such as VEGF and IL-8 secreted systemically by the AD-MSCs resident in the lungs could be the culprit in terms of neovascularization and higher rejection rates. In addition, it is known that MSCs are able to induce tolerogenic dendritic cells and to increase regulatory T lymphocytes. We administered syngeneic rabbit MSCs intravenously 7 days Epoxomicin before surgery to induce a systemic tolerogenic environment, and on days 0, 3, and 14 to maintain it. Other studies in low-risk settings found positive results when the cells were administered after surgery, or when the cells were injected 1 day before and the day of the surgery. In these cases, the mechanism of action could not be tolerogenic, because they achieved their positive results with post surgical or peri surgical administration. The models we present in this study are more similar to human corneal transplants. The similar size of the eyes allows for standard corneal transplant surgery, including the number of sutures, the type of sutures, the type of trepanation, and the size of trepanation, whereas using other, smaller animal models imposes the need to adapt the surgical technique. In addition, corneal allografts in rats are normally rejected within two weeks of transplant, whereas in mice syngeneic transplant allows 100% survival of the graft at 42 days, and allogenic transplant promotes 60% rejection earlier than a month post-transplant. In our experiments with rabbits, more than 50% of the corneal allograft often survived, even without immunosuppression, for over 2 months, closer to what is seen in humans if no immunosuppression is given. However, in both of our rabbit models, there were considerable differences in the clinical signs of rejection from our experience in humans. The surgery is more difficult due to the higher elasticity of the rabbit cornea, the higher retraction after trepanation, and a smaller anterior camera that provokes more frequent anterior synechiae and post-transplant ocular hypertension.