Background The intramedullary cavity is a widely distributed well-vascularized microenvironment capable of sustaining grafts, and is a potential site for islet transplantation. The bone marrow offers sufficient space that may also be suitable for bioartificial pancreas (BAP) implantation. Objective To evaluate the feasibility of bone marrow as an implantation site for BAPs. Materials and Methods A calcium phosphate cement chamber satisfies the criteria for immunoisolation. Mouse insulinoma cells were suspended with agarose gel and enclosed in a calcium phosphate cement chamber to create a BAP, which was implanted in the intramuscular space in diabetic swine or the intramedullary cavity in diabetic dogs. Blood glucose and C-peptide concentrations were determined perioperatively. Results In the swine, the mean ? SD blood glucose concentration decreased from 413 ? 24 mg/dL to 285 ? 47 mg/dL, and was maintained in the range of 285 to 336 mg/dL for 15 days. It increased to 368 to 450 mg/dL after the BAPs were implanted in the intramuscular space. In the dogs, the blood glucose concentration decreased from 422 ? 32 mg/dL to 247 ? 52 mg/dL, and was maintained in the range of 247 to 347 mg/dL after the BAPs were implanted in the intramedullary cavity. The C-peptide concentration increased from 6.1 ? 2.8 pmol/L to 104.7 ? 16.4 pmol/L when the BAPs were implanted in the intramedullary cavity. Conclusion This study indicates superior effectiveness of BAPs implanted in the intramedullary cavity compared with the intramuscular space. This observation may be attributed to the greater oxygen tension in the bone marrow. The BAPs in direct contact with the circulatory system receive sufficient blood flow for function and survival. This preliminary study demonstrates that the intramedullary cavity may be an implantation site for BAP transplantation.
