Islet autotransplantation studies in pigs to assess viability and implantation site as a prelude to xenotransplantation

Abstract: Before pig-to-human xenotransplantation trials can begin, it must be demonstrated that porcine islets can engraft and function in pig recipients. Collagenase dispersion of the pancreas followed by density gradient centrifugation is effective in preparing viable discrete islets from species such as rodents and dogs, but these methods frequently results in marked fragmentation of porcine islets. Thus, we hypothesized that by limiting the digestion period and eliminating the purification step to minimize islet fragmentation, a viable preparation suitable for transplantation would be obtained. Purification is known to be unnecessary in dogs and humans as long as a site able to accommodate impure tissue is used. In this study, we autotransplanted impure islets prepared by two different methods (collagenase/semi-automated or mechanical chopping) without further purification into either the peritoneal cavity (N=46), spleen (N=5), or liver (N=5) of 56 pancreatectomized pigs. The semi-automated method (N=45) yielded a mean of 426,000 ± 41,696 islets, and the mechanical chopping method (N=11) yielded 165,438 ± 19,611 islets (P<0.05). Intraportal transplantation of impure islets was uniformly lethal due to portal hypertension. Intraperitoneal transplants were well tolerated with no pigs dying periooperatively. None of the mechanically prepared islet grafts transplanted in the peritoneal cavity functioned (N=11). Of the grafts prepared by the collagenase/semi-automated method, none transplanted into the spleen functioned; however, 13 of the 35 (37%) transplanted into the peritoneal cavity did function (30 day insulin-independent survival). Of these, four recipients were normoglycemic (blood glucose <200 mg/dl), and glucose disposal during an intravenous glucose tolerance test was similar to normal pigs. Although islets prepared by the semi-automated/collagenase digestion function in less than half of recipients, the intraperitoneal site generally is less receptive than other sites of engraftment. Unlike pigs, humans can tolerate the transplantation of impure islets intraportally. Thus, the proportion that functioned in this study likely under estimates the potential function rate using the intraportal site in humans. This study demonstrates by in vivo testing the viability of an impure porcine islet preparation and supports its use in preclinical xenograft models and human xenograft trials.