Although allogeneic islet transplantation can successfully cure type 1 diabetes, it has limited applicability. For example, organs are in short supply; several human pancreas donors are often needed to treat one diabetic recipient; the intrahepatic site may not be the most appropriate site for islet implantation; and immunosuppressive regimens, which are associated with side effects, are often required to prolong survival of the islet graft. An alternative source of insulin-producing cells would therefore be of major interest. Pigs represent a possible alternative source of beta cells. Grafting of pig islets may appear difficult because of the immunologic species barrier, but pig islets have been shown to function in primates for at least 6 mo with clinically incompatible immunosuppression. Therefore, a bioartificial pancreas made of encapsulated pig islets may resolve issues associated with islet allotransplantation. Although several groups have shown that encapsulated pig islets are functional in small-animal models, less is known about the use of bioartificial pancreases in large-animal models. In this review, we summarize current knowledge of encapsulated pig islets, to determine obstacles to implantation in humans and possible solutions to overcome these obstacles.

Type 1 diabetes has been treated successfully by transplanting islets of Langerhans, the endocrine tissue that releases insulin[1]. Despite the clinical efficacy of islet transplantation, serious issues preclude its broad clinical application, including the side effects of chronic immunosuppressive regimens and a shortage of human donors. For example, it was recently estimated that only one pancreas is available per 333 patients with type 1 diabetes in the United States[2]. This situation is aggravated by the need of each recipient undergoing transplantation for 2-4 pancreases[1,3]. This shortage of pancreas donors therefore justifies the search for alternative sources of insulin-producing cells. Swine appear to be the major candidates for islet procurement because: (1) humans have been treated with porcine insulin for > 40 years (pig and human insulin differ by only one amino acid); (2) pigs have large litters with offspring that rapidly attain adult size and are therefore amenable to genetic engineering; (3) pig pancreases contain large islets that respond to glucose stimulation; and (4) since pigs are widely bred and slaughtered for food, the use of their islets to restore human health may be an option that could satisfy sociocultural and ethical concerns[4-6].

Unlike primarily vascularized organs, pancreatic islets are implanted without direct connection to the host vascular network, with 7 to 14 d needed to re-establish blood flow[7-12]. Thus, it was thought that pig islet xenografts could escape typical hyperacute rejection and acute vascular rejection[13]. In vivo, however, pig islets in non-immunosuppressed nonhuman primates are rejected by both humoral and cellular immune reactions[14-16]. Diffuse, presumably nonspecific IgG deposits were observed within islet-associated accumulations of platelets 12 h after transplantation. Deposits of large amounts of IgM and moderate to large amounts of C3, C5, and C9 were present on islet surfaces 2 to 3 d after xenografting[15-17]. Anti-galactosyl (anti-Gal) and non-Gal[18,19] IgM antibodies bind to islet surfaces soon after transplantation and activate the classical complement pathway, as well as promoting neutrophil infiltration. These humoral immune responses to pig islets are consistent with early T-cell-independent immune system activation and are reminiscent of mechanisms that operate during the hyperacute rejection of solid-organ xenografts[20]. Humans and nonhuman primates have preformed anti-pig antibodies that rapidly recognize the Gal epitope on islet endothelial cells. During pig islet-to-primate xenotransplantation, however, the expression of Gal epitopes is influenced by the age of the pig. Gal residues are present on 20% of fetal, but on only 5.1% of adult, islet β cells[21-23]. Since Gal expression persists after islet isolation[24,25], Gal remains a target for humoral xenorejection.

Islet xenografts that survive immediate blood-mediated inflammatory reactions and additional humoral damage will be subject to acute cellular xenograft rejection. Following transplantation of fetal pig islets under the kidney capsule, the cellular infiltrate in primates has been found to consist mainly of CD8 T cells (implicating the indirect pathway), whereas the cellular infiltrate in rodents was dominated by macrophages. T-cell infiltration precedes macrophage influx, with small numbers of CD3⁺ T cells observed 12 h after transplantation[14]. After 24 h, equal numbers of CD3⁺ T cells and neutrophils were observed, and after 72 h, CD3⁺ T cells dominated, representing 50% to 80% of all infiltrating cells. After 72 h, large numbers of macrophages were observed, with T cells localized at the periphery of and within transplanted islets. In addition, increased E-selectin expression on portal vein endothelial cells correlated with the infiltration of neutrophils, which caused tissue damage by releasing enzymes, active oxygen intermediates, and proinflammatory cytokines, and produced chemokines that attracted dendritic cells and T lymphocytes. Pig islet xenorejection seems theoretically easier to overcome, but because hyperacute and acute vascular rejections do not occur, the rapid destruction of pig islets within 72 h of transplantation into nonhuman primates demonstrates the strength of xenorejection.

Thus, an immunosuppressive regimen is mandatory for the long-term survival of pig islets in primates[26,27]. Although several immunosuppressive strategies have successfully suppressed alloimmune responses, T-cell-mediated xenoimmune responses have proven more resistant to immunosuppressive therapy[28,29]. This may be due to the greater molecular incompatibility between donor and recipient, which activates particularly the innate immune response[30].

Until recently, the maximum reported duration of pig islet survival (insulin-positive cells, no function) following transplantation under the kidney capsule in nondiabetic cynomolgus monkeys and immunosuppression with anti-thymocyte globulin, anti-interleukin-2R mAb, cyclosporine, and steroids was 53 d[31]. In March 2006, however, two studies reported that neonate or adult pig islets xenotransplanted into primates survived and functioned for > 180 d[26,27]. More recently, the transgenic expression of a human complement-regulatory protein (hCD46) on porcine islets was shown to enhance the survival of islets xenotransplanted into cynomolgus monkeys with streptozotocin (STZ)-induced diabetes for > 12 mo[32]. In addition, transplantation of galactosyl knock-out neonate pig islets was found to significantly enhance normoglycemia rates in diabetic primates, likely due to the decreased susceptibility of these xenografts to innate immunity mediated by complement and preformed xenoantibodies[33]. These results, however, required treatment with a heavy immunosuppressive regimen, in particular an anti-CD154-specific mAb, an antibody that induced thromboembolic events precluding its clinical use[34]. Despite the unacceptability of these immunosuppressive regimens in humans, these results are very encouraging since an alternative, nontoxic regimen combined with xenotransplantation of pig islets may induce normoglycemia in diabetic patients.

A bioartificial pancreas, in which islets of Langerhans are encapsulated within a semipermeable membrane, may be an alternative therapeutic device for patients with insulin-dependent (type 1) diabetes mellitus. It may constitute a safe and simple method of transplanting islets without the need for immunosuppressive therapy. Since the semipermeable membrane protects the islets from the host immune system, the islets are likely to survive and release insulin for a long period of time, thereby controlling glucose metabolism in the absence of immunosuppressive medication. Nevertheless, several important questions are associated with the transplantation of immunoisolated adult pig islets as a “bioartificial pancreas”[35] (Table 1).

The choice of material to use for cellular encapsulation is a crucial parameter because failure of microencapsulated islet grafts is usually regarded as a consequence of insufficient biocompatibility, inducing a nonspecific foreign body immune reaction against the microcapsules and resulting in progressively fibrotic overgrowth of the capsules. This overgrowth interferes with adequate nutrition of the islets and consequently causes islet cell death. There are major distinctions between water-soluble polymers, such as alginate, and water-insoluble polymers, such as poly (hydroxyethyl methacrylate-methyl methacrylate)[35]. However, a major obstacle in using water-insoluble polymers for encapsulation of cells is the requirement for an organic solvent, which usually interferes with cellular function[62]. Despite their solubility in aqueous solutions, alginate-based capsules have been shown to remain stable for several years in small and large animals as well as humans[51,63-67]. The method of alginate capsule formation is based on the entrapment of islets in alginate droplets, which are transformed into rigid beads by gelification in a divalent cation solution, mostly Ca²⁺. In most studies to date, alginate beads were coated with a second layer to reduce the porosity of the capsule membrane[35,68]. In the pig-to-primate model, alternating layers of poly-l-lysine and a polyornithine were used to surround the alginate core[40,51]. The latter type of layer, however, has been associated with polyamino acid cytotoxicity and mechanical instability of the microcapsules, limiting their application[35,69,70]. Several groups have recently reported that encapsulation in simple alginate microbeads can protect pig pancreatic cells against xenorejections in diabetic mice[46,47,71]. Although several chemical formulations of alginate (e.g., high-mannuronic/guluronic; high/low viscosity, with or without additional peptide sequences) have been proposed for islet immuno-isolation, we found that high-mannuronic alginate was the most suitable to obtain selective impermeability for molecules over 150 kDa (as an IgG) before and after transplantation, and optimal biocompatibility to avoid nonspecific inflammatory response associated with surrounding angiogenesis, resulting in sufficient oxygen tension (about 40 mmHg) for the survival and function of encapsulated islets[72]. This type of alginate was biocompatible not only in a small-animal model (Wistar rat recipients) but also in a pig-to-primate model of xenotransplantation under the kidney capsule and skin for up to 6 mo (see below)[25,43,48].

The lack of revascularization of the encapsulated islets interferes with both the functional performance and the longevity of the grafts. Apparently, a site in which encapsulated islets are in close contact with the bloodstream is obligatory for clinical application. Unfortunately, it is difficult to find such a site because it must be of sufficient size to bear a large graft volume and be near blood vessels. Sites reported to allow successful nonencapsulated islet transplantation, such as the liver and spleen, do not meet these requirements because these sites are unable to tolerate the large volumes (> 16 mL) of capsules (of diameter > 600 μm) required for transplantation into primates. Therefore, most transplantations of encapsulated pig islets into primates were intraperitoneal[39,39,51]. Although this technique seemed relatively easy, the peritoneal site was not optimal. Indeed, recent studies in mice found that macrophages and lymphocytes are involved in the rapid degradation of encapsulated pig islets after their transplantation into the peritoneum[71,73-76]. The peritoneum is, indeed, a preferential site for inflammation and immunologic reactions[77] and peritoneal mesothelial cells facilitate the action of powerful innate immune mechanisms[77]. Studies in mice showed that immunosuppression had beneficial effects, improving the biocompatibility and prolonging the survival of encapsulated pig islets transplanted into the peritoneum[73,74,78]. This method of combining encapsulation and immunosuppression, however, remains incompatible with clinical applications. The biocompatibility of alginate-encapsulated pig islets depends on the implantation site. Encapsulated pig islets transplanted under the kidney capsule and under the skin demonstrated better biocompatibility than capsules transplanted into the peritoneum[48]. Indeed, a cellular reaction essentially composed of macrophages was observed 7 d after transplantation into the peritoneum. This finding is in good agreement with results showing that macrophages are recruited 7 d after transplantation of encapsulated pig islets into the peritoneum of mice and rats[73,74]. In addition, severe fibrosis surrounding intraperitoneally implanted capsules was observed 30 d after transplantation and was correlated with the loss of porcine C-peptide 7 d after implantation. In contrast, subcutaneous and kidney capsule implantation resulted in very weak cellular immune reactions against encapsulated pig islets, along with improved porcine islet viability; porcine C-peptide was detected in the sera of rats for 30 d after transplantation of encapsulated pig islets at both sites. These findings indicate that implantation into the subcapsular kidney and subcutaneous spaces improves the biocompatibility and in vivo survival of encapsulated pig islets, as well as enhancing pig islet function during the first 7 d after transplantation. The loss of the in vivo function activity of encapsulated pig islets transplanted into the peritoneum correlated with significant alterations in islet viability, a loss of insulin content, and significant reductions in insulin secretion after glucose stimulation. These findings may be associated with macrophage overgrowth of the area surrounding the capsules, creating a microenvironment of stress, with low oxygen tension, for pig islets[75,76]. Indeed, macrophage activation, as shown by NO production and the release of the cytokines L-1β and tumor necrosis factor-alpha, had a deleterious effect on islet function and viability[73,75,76]. We found that implantation into kidney subcapsular and subcutaneous spaces improved the biocompatibility of encapsulated pig islets and significantly reduced macrophage recruitment. This reduction in pig islet stress and improved islet viability maintained insulin level per islet and insulin secretion after glucose stimulation[48]. Subcutaneous tissue was recently shown to provide oxygen tension (20-40 mmHg) compatible with the function and survival of encapsulated islets[72]. Among the sites being tested for islet transplantation, with or without encapsulation, to improve the survival, engraftment and function of islets, are the brachioradialis muscle[79], striated muscle[80], the greater omentum[81], and the anterior chamber of the eye[82].

Immunosuppression remains the major limitation of allotransplantation or xenotransplantation of islets for type 1 diabetes. Extended survival of transplanted pig islets has recently been observed in primate models, but several questions and problems associated with immunosuppression remain to be resolved in terms of adjustment before clinical trials. A bioartificial pancreas made of encapsulated pig islets may overcome the two major hurdles to islet transplantation: the shortage of human organ donors and the requirement for immunosuppressive regimens.

The development of a bioartificial pancreas for preclinical/clinical studies requires the conjunction of integrated parameters such as the choice of a biocompatible material for encapsulation to maintain selective permeability. The encapsulation device should be designed to maintain mechanical properties and stability at an implantation site compatible with the viability and physiology of the encapsulated islets to control glycemic homeostasis.

Of the three major types of bioartificial pancreases (macroencapsulation, microencapsulation, and conformal encapsulation), the macroencapsulation system is the only method that has demonstrated the capacity to control diabetes in large animals and in preliminary clinical studies.

Several improvements must be made to reduce the size of the implant (by increasing islet concentration relative to the surface or volume of the implant), to improve oxygenation of islets (to limit islet death), and to develop a simple clinical procedure for bioartificial transplantation and easy access to a device allowing “re-alimentation” of the islets.