Although transplantation of isolated pancreatic islets in the insulin-dependent diabetic is not yet a routine clinical procedure, an extensive research literature exists and limited clinical trials are now in progress. Thus, this experimental area of transplantation needs to be considered in the context of vascularized pancreatic transplantation for type I diabetes.

Transplantation of the vascularized pancreas has become increasingly respectable in terms of graft and patient survival, such that the results obtained now compare reasonably with those of other organ transplants. But the concept of transplanting an entire organ to cure a non-lethal condition when what is required is only a small fraction of the transplanted tissue would appear to be flawed. Although there is evidence that the complications of diabetes are caused by imperfect control of glucose metabolism and that improved glucose homeostasis would probably prevent their onset, there is also considerable evidence that transplantation of the endocrine pancreas at a late stage in the disease has relatively little effect on these complications. Therefore transplantation would have to be performed in the early years after presentation in order to influence the development of these microangiopathic complications of diabetes. Vascularized pancreatic transplantation with immunosuppression is never likely to be an acceptable therapy for the recently diagnosed young diabetic, but pancreatic islet transplantation could be, especially if this could be performed without the need for long-term immunosuppression, as might be possible (see below).

The idea of transplanting insulin-secreting tissue as a free graft for diabetes is older than the discovery of insulin, but little progress was made until relatively recently. In 1965 Moskalewski described the successful isolation of islets from the rodent pancreas using collagenase digestion, after which Lacy in 1967 perfected the isolation technique to produce sufficient islets for transplantation in rats. The enthusiasm raised by these early experimental results prompted several early attempts at human pancreatic islet allotransplantation and autotransplantation. Despite early claims to the contrary, it is now generally agreed that these early attempts failed, not least because the technique for islet isolation developed in the rat was not applicable to the human pancreas. These early clinical trials also emphasized the dangers of transplanting unpurified dispersed pancreatic tissue. Over the last 20 years techniques for isolation and transplantation of pancreatic islets have been developed in experimental animals. Significant technical advances in that period were the demonstration of successful transplantation of unpurified islet tissue in pancreatectomized dogs; the use of intraductal collagenase for islet isolation in the dog; and successful transplantation of purified islets in the dog. This work led to the successful development of techniques for islet isolation from the human pancreas. Further modifications continue to improve both the yield and purity of islets obtained from both human and other mammalian pancreases.

An alternative to the use of isolated adult islets is the transplantation of fetal pancreas, which arose out of observations by Coupland in 1960, namely that the endocrine tissue of the fetal pancreas is relatively well developed and survives transplantation while the exocrine tissue is poorly developed and undergoes atrophy after transplantation. However, it was the late Josiah Brown in 1974 who showed that transplantation of fetal pancreas would cure experimental diabetes in rodents.

Although many techniques for islet isolation have been described in the past, varying from microdissection to dispersion with modified food blenders, almost all groups working with human pancreas, as well as the pancreas of other large mammalian species, now use a variant of intraductal collagenase digestion. The principle behind the technique is the delivery of the collagenase into an exocrine pancreatic duct, which allows selective digestion of the interacinar connective tissue (Figs. 1–3) 732,733,734. Under optimal conditions, for example, collagenase delivered at 39 °C, the interacinar fibrous tissue is removed within 10 to 30 mins. The digestion process can be stopped by cooling at this stage, but the thicker intralobular fibrous tissue remains largely undigested and to release the islets it is necessary to employ the gentlest possible mechanical dispersion. There are now a number of semiautomated machines that use the above principles to disperse human, porcine, and canine pancreas, liberating islets mixed with a large number of exocrine fragments. No entirely successful technique has been developed to separate the exocrine tissue from the islet tissue. The most successful technique relies on the differential density of islet and exocrine tissue, which allows separation of the tissues by centrifugation on a Ficoll or albumin density gradient (Figs. 4 and 5) 735,736.

Although some endocrine tissues are relatively non-immunogenic, there is no doubt that isolated pancreatic islets transplanted as a free graft are highly immunogenic. The mechanisms of rejection may be somewhat different to those of a vascularized organ allograft and more akin to those seen in skin allograft rejection. Although prolonged survival of islet allografts has been obtained with a number of immunosuppressive protocols in several animal models, permanent survival of grafts has been difficult to achieve, even in rodent models where tolerance to a vascularized graft has been produced. However, the introduction of cyclosporin, particularly using high-dose parenteral regimens, has resulted in the long-term survival of allografts in rat and dog models. Cyclosporin-based protocols, usually as part of a triple or quadruple therapy regimen, are currently being used for immunosuppression in clinical trials of isolated pancreatic islet transplantation.

One of the most exciting aspects of islet transplantation has been the possibility of abrogating the immune response to pancreatic islets by altering their immunogenicity before transplantation. This has been achieved experimentally by tissue culture for several days either in routine culture conditions or at room temperature, or by treatment with ultraviolet light, high oxygen tension, or antibodies specific for interstitial dendritic cells. Islet allografts treated in these ways have shown delayed or absent rejection in rodent and canine models. These studies have obvious clinical potential, allowing the treatment of diabetics by islet transplantation without the need for long-term immunosuppressive drug therapy. Another possible approach to the prevention of rejection is the microencapsulation of the islets in individual membranes that are impervious to larger molecules, for example antibiotics and lymphocytes, yet are permeable to nutrients and insulin secreted by the islets. Animal models suggest that this technique could allow transplantation without the need for immunosuppressive therapy, and might even allow the transplantation of islet xenografts from animals such as the pig.

Islets have been implanted in most possible sites in experimental rodent models with variable success, but studies in large animal models suggest that the most acceptable sites for clinical application are the liver (via the portal vein), the spleen (either via the splenic veins or by direct puncture), or beneath the kidney capsule (Fig. 6) 737. At present the favoured site for clinical trials is the liver, using the portal vein. This appears to be safe, provided that the preparation is relatively free of exocrine contamination. However, impure preparations could cause portal hypertension with disseminated intravascular coagulation and should not be used in clinical trials.

Transplantation of isolated pancreatic islets can produce virtually normal glucose metabolism in rodent models of experimental and autoimmune diabetes. But in larger animal models the blood sugar, although remaining within the normal range during normal nutrition, becomes abnormal in the presence of stress, probably due to a relatively reduced islet cell mass. Whether this abnormality would be important for prevention of the long-term complications of diabetes, which is the ultimate goal of islet transplantation, remains to be seen.

One disturbing feature of studies of autologous islet transplantation in the dog and monkey is that grafts have tended to fail after 1 to 3 years. This may be due to a process of exhaustion, related to a reduced islet cell mass and the site of implantation. Recent reports in the dog model suggest that longer function is possible, but the definitive answer will probably only come from full clinical trials.

As described above, the early trials of clinical islet transplantation were flawed by uncertainty about the viability and actual islet content of the tissue being transplanted. These experiments highlight a problem that has dogged islet transplantation research in general: how to express the actual quantity, the purity, the extent of exocrine contamination, and the viability of the isolated islet tissue in a standard format. Considerable progress has been made in producing more standardized methods of assessment that allow better comparisons between laboratories. Of particular importance has been the development of specific and rapid stains, such as dithizone, for the accurate identification of islet tissue. Over the past 5 years there have been further clinical trials from centres such as St Louis and Edmonton, where the mass of islet tissue, and the purity and viability of the tissue, have been better documented. Most transplants were performed in diabetic patients with renal failure who also underwent kidney transplantation at the same time, but more recent transplants have been performed in diabetic patients with stable long-term renal allografts. In current trials the intraportal route of implantation in the liver has been used, relying on cyclosporin-based immunosuppression with the addition of antithymocyte globulin to prevent rejection.

Some of the earlier transplants showed signs of function with elevated C peptide for a few weeks, but there was not sufficient function to allow the discontinuation of insulin. The islets used in these experiments were obtained from a single donor. More recently, several patients have had graft function sufficient to allow insulin to be discontinued for a period of several weeks to over 1 year, but islets from more than one donor were used in all but one of these cases. Thus this represents a major advance in islet transplantation but it is still far from becoming a routine therapy. Nevertheless, the situation could be compared to the earliest steps in kidney transplantation 40 years ago, when a few grafts functioned only for a few weeks or months after transplantation, but from which a successful method of treatment has grown.

A major advantage of fetal pancreas transplantation over adult islet transplantation is that minimal processing of the islet tissue is required for transplantation. Although there is a potentially larger supply of human fetal pancreas than human adult pancreas (at least in those countries that permit abortion), several fetal pancreases are required in experimental models of diabetes to correct the diabetic state. Furthermore, the moral and ethical issues of using fetal tissue for transplantation are considerable. Another advantage of fetal pancreatic transplantation is the tremendous potential for growth of the fetal pancreas, unlike adult islets. However, immature fetal islet tissue does inhibit normal insulin secretion in response to glucose. For these reasons a long period of inadequate function is to be expected after transplantation of fetal pancreas into adult recipients, a prediction confirmed in rodent experiments.

In comparison to adult islet transplantation, progress in fetal pancreas transplantation is less advanced. Functioning fetal pancreas transplants, both isografts and allografts, have been described and studied in detail in rodent models but, apart from successful fetal pancreatic transplantation in the pig, it has not been possible to repeat these findings in other large animal models. Some successful human cases have been reported from China, but none elsewhere. The fetal pancreas can be transplanted intact in the mouse, or in a few segments in the rat. The tissue is then small enough to survive as a free graft, the most successful site for implantation being under the kidney capsule (Fig. 7) 738. In larger animals the fetal pancreas is too large to survive as a free graft and must be dispersed, either by sectioning the pancreas into small particles, or by digesting it with collagenase. The former provides less usable tissue, whereas it is possible to disperse the tissue fully using collagenase. After a period in tissue culture, which appears to favour islet tissue survival, the dispersed islet tissue forms ‘proislets’. This process has proven particularly successful in the preparation of pig fetal tissue.

As was the case for adult islets, initial hopes that fetal pancreatic tissue would be non-immunogenic have proved unfounded, and allografted or xenografted fetal pancreas is certainly rejected vigorously. Rejection can be prevented by various immunosuppressive protocols, although interpretation of these experiments is made difficult by the prolonged time it takes for the grafts to function after transplantation. Tissue culture in the presence of high oxygen tension has been used to reduce the immunogenicity of fetal pancreas grafts, as for adult islets. Long-term survival of allografted fetal pancreas grafts after culture has been described in mice, although this effect is markedly strain-dependent and has not been reproduced in the rat. Histological evidence that these culture techniques can prevent or delay rejection in large animals has also been presented, but there are no functioning models of fetal pancreas transplantation in large animals as a final confirmation.

There have been a remarkable number of trials of fetal pancreas transplantation in man, stretching back some 15 years. Numerically the greatest number have been performed in China and the Eastern bloc countries, and despite claims of success in some patients, objective evidence of function has not been provided. Clinical trials have been undertaken in Sydney and Denver, as well as sporadic attempts elsewhere. The sites that have been used for transplantation include muscle pockets, the kidney capsule, and the intra-abdominal omentum. A combination of tissue culture prior to transplantation with a cyclosporin-based immunosuppressive protocol after transplantation has been used to try to prevent rejection. To date there has been no evidence of function either in terms of C peptide production or decreased requirements for insulin therapy.

That clinical development of isolated pancreatic islet transplantation to the point where definite, albeit short-term, function has been produced in a few patients is certainly an exciting advance. However, the function of these transplanted human islets has been less satisfactory than that obtained in animal models with an apparently equivalent transplanted mass. The reason for this difference needs elucidation, and, if the problem is solvable, this may result in functional one-to-one grafts. The problems of rejection, recurrence of disease (which has not been discussed here), and long-term function will then be the next hurdles to overcome before finally applying the technique to young diabetics.

Fetal pancreas transplantation needs development in a large animal model to answer some fundamental questions before application to the human is likely to succeed. There are moral and ethical issues involved that make general application of the technique seem unlikely at the present time.

Brown J, Molnar IG, Clark W, Mullen Y. Control of experimental diabetes mellitus in rats by transplantation of fetal pancreases. Science 1974; 184: 1377–9.

Coupland RE. The survival and growth of pancreatic tissue in the anterior chamber of the eye of the albino rat. J Endocrinol 1960; 20: 69–77.

Gray DWR, McShane P, Grant A, Morris PJ. A method for isolation of islets of Langerhans from the human pancreas. Diabetes 1984; 33: 1055–61.

Gray DW, Morris PJ. Transplantation of isolated pancreatic islets. In: Groth CG, ed. Pancreatic transplantation. Philadelphia: W. B. Saunders, 1988: 363–90.

Horaguchi A, Merrell RC. Preparation of viable islet cells from dogs by a new method. Diabetes 1981; 30: 455–8.

Lacy PE, Kostianovsky M. Method for the isolation of intact islets of Langerhans from the rat pancreas. Diabetes 1967; 16: 35–9.

Mirkovitch V, Campiche M. Absence of diabetes in dogs after total pancreatectomy and intrasplenic autotransplantation of pancreatic tissue. Transpl Proc 1977; 9: 321–3.

Morris PJ, Gray DWR, Sutton R. Pancreatic islet transplantation. Br Med Bull 1989; 45: 224–41.

Moskalewski S. Isolation and culture of the islets of Langerhans of the guinea pig. Gen Comp Endocrinol 1965; 5: 342–53.

Noel J, Rabinovitch A, Olson L, Kyriakides G, Miller J, Mintz DH. A method for large-scale high-yield isolation of canine pancreatic islets of Langerhans. Metabolism 1982; 31: 184–7.

Sutherland DE. Pancreas and islet transplantation. I. Experimental studies. Diabetologia 1981; 20: 161–85.

Sutherland DE. Pancreas and islet transplantation. II. Clinical trials. Diabetologia 1981; 20: 435–50.

Warnock GL, Rajotte RV. Critical mass of purified islets that induce normoglycemia after implantation into dogs. Diabetes 1988; 37: 467–70.

Warnock GL, et al. Continued function of pancreatic islets after transplantation in type I diabetes. Lancet 1989; ii: 570–2.