Bone marrow as a site for pancreatic islet transplantation

To the editor:

We read with great interest the study by Cantarelli et al in which the authors definitively demonstrated the ability to cure streptozotocin-induced diabetes in mice by transplanting islet isografts into the bone marrow (BM) cavity of the femur.¹  We agree with the authors that this site has many potential technical advantages over the traditional intraportal site.²  However, we have 2 queries for the authors. First, the full justification for use of this site is not articulated, and we would ask the authors to elaborate on their theoretical rationale for this implantation site. Our own work in this area was explicitly designed to investigate the BM as a potential site for grafts of nonhematopoietic tissue²  and was based upon an immunologic theory proposing that the BM maintains a proportion of suppressive clones to facilitate the avoidance of autoimmunity.^3,4  Indeed, our results suggest that the BM does impart some degree of immunologic privilege as a transplantation site, based upon histologic demonstration of intact insulin- and glucagon-positive islet allografts within the marrow cavity of the rat 3 weeks after transplantation without evidence of islet allograft rejection and without immunosuppression.²  Using similar protocols, rat islet allografts transplanted at other sites (portal vein, kidney capsule) are uniformly rejected, functionally and histologically, in less than a week.^5,6  Second, we would ask the authors to elaborate further on their methods used for the implantation process. Their method is identical to that described in our previous work, except for the use of the marrow cavity of the mouse femur in their work versus the rat tibia as we described.²  In our studies, there was a concern regarding leakage of part of the islet “plug” from the marrow cavity, which can prevent the achievement of an adequate graft volume. For this reason, we chose to use histology to demonstrate proof of principle that islets could be successfully transplanted and engraft within the bone marrow site. Did the authors add any measures beyond the methods from our original paper to prevent leakage of the implanted cells? It is likely that drilling upward from the knee into the distal femur of a mouse would result in some leakage, as did drilling downward from the knee into the proximal tibia of a rat. Regardless, we have found that the procedure is technically much easier in larger animals with a larger bone and marrow cavity, and we have used this technique to successfully transplant neonatal pig islet cells into the intramedullary bone marrow of the pig tibia. The findings of both Cantarelli et al and ourselves support further work examining the utility of the BM as a site for transplantation, with an enormous potential for further research and clinical application.