pancreatic islet transplantation

In this paper the theoretical basis of ailoreactivity and its relevance to transplantation biology is discussed prior to a review of work showing that culture of adult mouse pancreatic islets for 7 days in 95% O2 and 5% CO2 facilitates successful grafting to nonimmunosuppressed allogeneic recipients. These allografts function by reversing both chemically induced and spontaneous diabetes. The fetal mouse pancreas is more immunogenic than adult islets, and even after a culture period of 10 days in 95% O2 and 5% CO2, BALB/c allografts are consistently rejected by nonimmunosuppressed recipient mice. The immunogenicity of fetal pancreas is thought to be due to the presence of contaminating lymphoreticular cells in the mesentery surrounding the fetal pancreas. Digestion of the fetal pancreas with collagenase allows the isolation of proislets that develop into functional islet tissue on transplantation. Fetal proislets are less immunogeneic than the whole fetal pancreas and may provide a source of tissue for clinical transplantation. Established islet allografts are relatively stable and are not rejected following nonspecific stimulation of the recipient's immune system or following passive transfer of either antibody or antibody and complement. After prolonged residence in the recipient a state of allograft tolerance develops and such grafts resist rejection by specific stimulation of the recipient. The administration of donor antigen in the form of uvirradiated cells enforces this state of allograft tolerance. DIABETES 31 (Suppl. 4):30-37, June 1982.


SUMMARY
In this paper the theoretical basis of ailoreactivity and its relevance to transplantation biology is discussed prior to a review of work showing that culture of adult mouse pancreatic islets for 7 days in 95% O 2 and 5% CO 2 facilitates successful grafting to nonimmunosuppressed allogeneic recipients. These allografts function by reversing both chemically induced and spontaneous diabetes. The fetal mouse pancreas is more immunogenic than adult islets, and even after a culture period of 10 days in 95% O 2 and 5% CO 2 , BALB/c allografts are consistently rejected by nonimmunosuppressed recipient mice. The immunogenicity of fetal pancreas is thought to be due to the presence of contaminating lymphoreticular cells in the mesentery surrounding the fetal pancreas. Digestion of the fetal pancreas with collagenase allows the isolation of proislets that develop into functional islet tissue on transplantation. Fetal proislets are less immunogeneic than the whole fetal pancreas and may provide a source of tissue for clinical transplantation.
Established islet allografts are relatively stable and are not rejected following nonspecific stimulation of the recipient's immune system or following passive transfer of either antibody or antibody and complement. After prolonged residence in the recipient a state of allograft tolerance develops and such grafts resist rejection by specific stimulation of the recipient. THEORY OF ALLOGENEIC REACTIVITY AND ITS RELEVANCE TO TRANSPLANTATION BIOLOGY Theory of ailoreactivity. It may seem, at first, a facile truism to maintain that an understanding of the transplantation barrier is tied to an understanding of the immune system. However, it is now very likely that the behavior of the major histocompatibility complex (MHC) in transplantation reac- tions simply reflects the function of this antigen complex as a control mechanism within the immune system. If we examine this notion in detail, we can see how a modification of the tissue prior to grafting offers a new approach to the regulation of tissue immunogenicity.
The following theoretical approach accepts the postulates of Burnet 1 that receptor diversity is generated by a random process, and that self-reactivity is forbidden by a mechanism learned during ontogeny and reinforced throughout the life of the animal. What we have done is develop a theoretical system based on a further two postulates that govern the process of clonal selection and expansion.
T-cells do not respond with antigen alone, but respond to antigen in conjunction with a second signal provided by an inductive molecule. 2 " 4 We refer to this inductive activity as costimulator (CoS) activity. These findings form the basis of the first postulate, which is: Two signals, antigen and CoS activity, are required for T-cell activation.
In symbolic terms we can write this postulate as where T (x) is a resting T-cell of receptor specificity x which is converted to the activated state T ( ' x) following contact with antigen x in the presence of CoS activity. CoS activity is a cellular product. 2 -3 Thus, a corollary of the first postulate is the notion that a stimulator cell is required for T-cell activation. A stimulator cell is defined as any cell that has the capacity to produce and release CoS activity (expression of the S + phenotype). Other cells that do not provide a source of CoS activity express the S~ phenotype. 5 The reaction sequence to describe the first postulate can now be written as: where antigen x provides signal (1) for the T-cell, and the S + cell provides a source of CoS activity, signal (2). The second postulate of the theory is: Immune responses are initiated when antigen x is presented to potentially re-r sponsive T-cells on the surface of an S + cell, and that a control structure c on the surface of this cell regulates the release of CoS activity; CoS activity is released when the control structure is engaged by the potentially responsive Tcell.
We can now use these postulates to draw up a theoretical model of the events that occur when foreign antigen x or allogeneic lymphoid cells enter the immune system. Only the theory of alloreactivity will be dealt with here.
In the resting immune system, spontaneous T-cell activation does not occur, because T-cells do not express a functional receptor for self-antigens, and c, the control structure is a self-antigen. Thus, the resting situation in the immune system can be written as: S c + T c ( , -» -ve where S c is an S + cell that carries the control structure c and T c ( ) is a T-cell of the same genotype that may also express c on its surface but does not express a functional receptor for c, which is a 'self component.
Let us assume that genes specifying the control structure are polymorphic within a population. These genes can be designated as c, c,, c 2 , etc. Thus, within the population we will have different lymphoid cell phenotypes, S c , T c , S C1 , T Ci ; etc. The rule for the T-cell is that T cannot carry a functional receptor for c, but may carry a receptor for c,, c 2 , etc.
Consider the interaction between T-cells, T Ci , and allogeneic S + cells, S c .
That is, in this situation T-cell activation will occur without the involvement of exogenous antigen. Such allogeneic interactions are simply abnormal immune reactions that occur spontaneously because T Cl cells carry a functional receptor for c, the control structure on the allogeneic S + cell.
There is another important characteristic of c-type antigens. Allogeneic T-cells respond to these antigens in an unrestricted manner (see above).
The specificity of cells responding to other antigens, whether they be minor histocompatibility antigens or other foreign antigens, will always be restricted by the control structure of the S + cells required for activation.
We can use experimental alloreactivity to define the nature of c, the control structure. C structures will be that individual or group of alloantigens that is highly immunogenic for T-cells in vitro. Experimentally, we know that the major histocompatibility complex (MHC) and the Mis locus behave in such a way. 6 Response to a tissue aUograft. The above analysis shows that the control structures (c) of the immune system behave as MHC antigens in the mixed leucocyte reaction. Let us now consider how these antigens behave in the case of tissue transplantation. The in vivo response to a tissue allograft is more complex than the in vitro unidirectional mixed leucocyte reaction because of the greater range of potential cellular interactions. The transplanted tissue will contain antigen-bearing S~ parenchymal cells. The tissue will also carry lymphoreticular cells, either as fixed tissue elements or as passenger leucocytes. These cells provide a source of antigen-bearing S + cells. They also provide a source of responsive leucocytes (R) that are activated by lymphokine and are not antigen specific. 7 Cells of this type contribute to the inflammatory component of the allograft response. 8 -9 The simplest way to analyze the allograft situation is to consider separately the types of interaction that may occur in the recipient's immune system and in the transplanted tissue itself.
Antigen-bearing S + cells from the grafted tissue will be carried to lymph nodes draining the transplantation site by way of the lymphatic system. In the node, control structure antigens of the donor (c genotype) will be presented directly to the recipient's T-cells (c, genotype) on the surface of S + cells. The allogeneic interactions detailed above will result in T-cells which are specifically reactive to alloantigen of the donor.
In the case of a solid transplant, the graft will consist of tissue parenchymal cells including vascular endothelium. Reactions in the transplanted tissue will be regulated primarily by the interaction of recipient T-cells with graft antigens. Activated T-cells have the potential to express cytotoxic activity and in this way damage graft endothelium and parenchymal cells. In addition, binding of these cells to antigens on target cells will trigger lymphokine release and so activate a nonspecific inflammatory process in the transplant. It is likely that graft damage will result from both a direct cytotoxic effect directed against the cells of the transplant combined with a severe nonspecific inflammatory reaction. 10 The idea that leucocytes carried along with the transplanted tissue (passenger leucocytes) are important in allograft rejection has been a recurrent theme in discussions of transplantation biology over the last 20 odd years. 3i11~13 Our theoretical treatment once again returns to this concept and, in an explicit way, states that passenger leucocytes constitute the major barrier to the transplantation of foreign tissues across MHC differences; transplantation antigen itself is not the barrier to the grafting of foreign tissues. Firstly, the generation of graft-specific T-cells is only brought about by direct stimulation of recipient T-cells by donor stimulator cells, and these are cells of lymphoreticular origin. Secondly, one component of the destructive process initiated in the tissue transplant is a postulated nonspecific inflammatory reaction that results from the activation of leucocytes within the transplanted tissue and the recruitment of host leucocytes following lymphokine release from specifically activated T-cells. Thus, passenger leucocytes play a major role in both the activation of, and the expression of cell mediated response of the recipient to allografted tissues.
From our theoretical viewpoint, it should be possible to greatly diminish the immunogenicity of a tissue graft if one can completely remove passenger leucocytes, including fixed lymphoreticular cells, from the tissue prior to transplantation. A transplant treated in this way would still carry MHC antigens on graft parenchymal cells. Because these cells express the S~ phenotype, this antigen can only be immunogenic when presented to the recipient's immune system on host stimulator cells, and this reaction does not generate a graft-specific T-cell response. It could, however, result in the production of helper cells that would amplify graft-specific antibody responses. In some cases, such antibody may protect the transplant from the damaging effect of a cytotoxic T-cell response. 14 In 1975, we investigated the effect of organ culture on the immunogenicity of thyroid tissues. 15 The idea that organ culture might reduce tissue immunogenicity was not new. There were several reports in the thirties and forties suggesting that clinical benefit was observed when parathyroid tissue was held in organ culture for a period prior to transplantation to patients with hypoparathyroidism. 16 -17 However, these studies were not genetically controlled, and without the support of an adequate theoretical base, enthusiasm for these experiments quickly waned. Our interest in organ culture was stimulated by the report from Summerlin et al. 18 -subsequently not confirmed 19 -that organ culture prior to transplantation could facilitate the grafting of skin to normal allogeneic recipients. Against the above theoretical background, the effect of organ culture could be explained if blood cells in the tissue died or were inactivated during the culture period.
The organ culture technique has proved spectacularly successful in the case of thyroid transplantation. Both cultured and uncultured mouse thyroid can be transplanted under the kidney capsule of isogeneic thyroidectomized recipients, where their ability to concentrate radioactive iodine can be used as a measure of graft function. 15 Thyroid allografts transplanted across a major histocompatibility barrier (H-2 incompatible) are acutely rejected within 15-20 days of grafting. Organ culture of thyroid in an atmosphere of 95% O 2 , 5% CO 2 for 14 days extends allograft survival 15 and after a culture period of 3-4 wk allografts show no evidence of rejection over an observation period of 100 days. 20 Pretreatment of the tissue donor with cyclophosphamide (300 mg/kg) 4 and 2 days prior to the harvest of tissues, reduced the period of organ culture required to facilitate allograft survival; after this treatment tissue cultured for 7-10 days could be transplanted to normal allogeneic recipients where it continued to function for an observation period of 250 days. Histological examination at the completion of this study showed the presence of normal thyroid and parathyroid tissue under the kidney capsule of the allogeneic recipients. 3 What is the explanation of this phenomenon? Do tissues lose antigen in organ culture, or does the effect result from the postulated loss of passenger leucocytes? During organ culture, there is a rapid degeneration of the vascular bed and blood elements within the cultured tissue. 21 However, the tissue does retain recognizable antigen, and can be rejected when as few as 10 3 viable peritoneal cells of donor origin are injected into the recipient at the time the cultured tissue is transplanted. 22 Established allografts are also rejected when a second uncultured thyroid of donor origin is transplanted to the recipient. 3 This effect is specific since an uncultured third party transplant can be rejected, but its rejection has no effect on the integrity of the established cultured allograft. These studies also show that the cultured allograft does not induce a state of unresponsiveness in the recipient, a conclusion confirmed by the finding that lymphocytes from animals bearing long-term cultured allografts show normal reactivity in vitro to the antigens of the donor strain. 23 These studies on thyroid tissue transplantation have been confirmed by Bach and his colleagues, 24 -25 who have also concluded that the immunogenic stimulus in the transplanted tissue is provided by antigen on metabolically active leucocytes.
It has been suggested that vascular endothelium may be a major source of allograft immunogenicity. 26 Vascular endothelium degenerates during organ culture, and its destruction could therefore account for the reduction in tissue immunogenicity achieved by organ culture. Endothelial cells express a high density of K-and D-type antigens but may lack l-region-type antigens. 27 Thus, the endothelium will provide a target for the attack of activated recipient Tcells. The question that arises is whether endothelium is a major source of stimulating activity for the recipient T-cells. Cyclophosphamide treatment of animals 4 and 2 days prior to the harvest of tissues causes a profound drop in the capacity of spleen cells to stimulate allogeneic T-cells in culture. 20 This treatment with cyclophosphamide has no microscopically obvious effect on the endothelium of tissues such as the thyroid. 21 However, after this treatment alone, approximately 30% of thyroid allografts function normally over an observation period of 100 days. 21 If donor endothelium was a major source of tissue immunogenicity for the recipient, we would not expect to observe such an effect. The fact that only a proportion of allografts show extended survival under these conditions could be accounted for by the incomplete removal of donor leucocytes with stimulating activity following cyclophosphamide pretreatment alone. It would appear likely, therefore, that tissue endothelium is not a major source of allograft immunogenicity. Conclusion. Our theoretical treatment is based on the postulate that a control structure regulates cellular interactions within the immune system. This complex, which includes a group of cell surface molecules, influences the specificity of T-cells activated to foreign antigen. Genes that specify the control structures of the immune system must, of their nature, form the MHC of the species, and the products of these genes are MHC antigens. The theory clearly shows the MHC antigens of themselves are not the major barrier to the allografting of tissues. This is because MHC antigens only induce a strong allograft-specific response when presented to the recipient's immune system as part of the surface of viable S + cells of donor origin. The theory emphasizes the role 'passenger leucocytes' play in both the induction and the expression of allograft immunity, and shows how treatment of the tissue prior to grafting can reduce its immunogenicity for the recipient. Experimentally, organ culture of tissues such as the thyroid gland in 95% oxygen has been shown to eliminate the immunogenicity of this tissue for allogeneic animals without removing MHC antigens from tissue parenchymal cells.

TRANSPLANTATION OF ADULT ISLETS
Organ-culture-induced modification of islet structure. Cultivation of pancreatic islets in 95% oxygen facilitates allograft survival as it does in the thyroid allograft model. 28 " 30 A number of major structural changes occur during the period in culture.
Normal islets of Langerhans consist of endocrine cells and a network of capillaries. 31 ' 32 After 7 days of cultivation in 95% oxygen, capillary endothelium degenerated and was not observed in any islets examined. 21 Small intercellular spaces were seen by electron microscopy to contain only debris from dead cells. The endocrine cells appeared to be normal. The loss of vascular endothelium during cultivation is similar to that seen when thyroid tissue is maintained in organ culture. 21 Mouse endothelial cells express K/D-region antigens but not l-region antigens. 27 The islets of Langerhans express K/D-region antigens that are detectable by complement-mediated lysis 32 but not by immunoferritin labeling. 27 Beta-cells do not express l-region antigens, 27 -32 whereas cells bearing a high density of l-region antigen are found in mouse thyroid tissue. 27 Morphologically, these cells resemble splenic dendritic cells and are lost from the tissue upon cultivation in high oxygen. Very low numbers of l-region antigen bearing cells were observed in uncultured islet tissue. Transplantation of cultured islet tissue: histoiogic study. The transplantation of uncultured BALB/c (H-2 d ) islets under the kidney capsule of CBA/H (H-2 k ) mice stimulated a violent allograft response by day 4 post-transplant. At this time the grafted area was lightly to heavily infiltrated by mononuclear cells including numerous small lymphocytes, lymphoblasts, and macrophages. The islets were generally intact at this stage but some were infiltrated by small pockets of lymphocytes. By day 7, the center of the islets had become disorganized and graft rejection was complete by day 14, when only scattered fragments of islet tissue were seen at the site of transplantation. 28 Mouse islets cultivated as clusters of 50 islets, unlike single islets, 28 ' 32 survive well in an atmosphere of 95% O 2 , 5% QQ 2 29, 33 T n e clusters maintain their histoiogic structure, and aldehyde-fuchsin staining of this tissue showed the presence of beta-cells following organ culture for a period of up to 20 days.
Allografts of 7-day cultured islet tissue showed no evidence of rejection when grafted under the kidney capsule of nonimmunosuppressed recipients. Such allografts have been examined histologically at 14, 28, 84, and 420 days after transplantation. 28 ' 30 At 14 days posttransplantation, the tissue was well vascularized and the beta-cells in the graft stained strongly with aldehyde-fuchsin. Although small pockets of lymphoid cells were sometimes seen (V38 grafts) these had dissipated by day 84. At 420 days, histoiogic examination of the grafts showed a remarkable arrangement of blood vessels and aldehyde-fuchsin staining beta-cells ( Figure 1) with no endocrine cell being more than one cell removed from a blood vessel.
Islet transplantation in different strain combinations. Prolongation of allograft survival can be obtained in strain combinations other than BALB/c to CBA. Uncultured and 7day cultured BALB/c islet clusters were transplanted to B-10 G (H-2 q ), ATL (H-2K s l k K d ), C57BI/6 (H-2 b ) and DBA/2 (H-2 d , minor antigen differences) mice. The B1O.G recipients behaved in a similar way to CBA mice, as a 7-day culture period reduced graft immunogenicity and facilitated islet survival. At 28 days following transplantation, the B1O.G mice rejected % uncultured grafts. In contrast % cultured islet allografts survived allotransplantation in this combination (Table 1). In the BALB/c to C57BI/6 strain combination 1 of 10 uncultured allografts was intact at 28 days posttransplantation. After 7 days of organ culture, 6 /n allografts survived for 28 days. When uncultured BALB/c islets were allografted to ATL mice, 7 /9 grafts showed no signs of rejection at 28 days while all 11 of the cultured grafts remained intact. In this strain combination, the uncultured islets were poorly immunogenic. The uncultured BALB/c islets were also poorly immunogenic when grafted across a minor histocompatibility barrier into DBA/2 recipients and 4 /s grafts were completely intact at 28 days post-transplant. All 6 of the cultured islet clusters survived completely intact (Table  1) in this combination.
These findings show the general applicability of the organ culture procedure in preparing islet tissue for grafting. Seven days cultivation in 95% O 2 and 5% CO 2 was not sufficient to ensure survival of 100% of BALB/c allografts in C57BI6 mice although the survival was significantly better than with uncultured tissue. However, this treatment was effective in the case of BALB/c tissue grafted to CBA or B10.G recipients. In the case of minor antigeneic differences no pretreatment of the tissue was required to facilitate allograft survival. Reversal of diabetes by allogeneic islet transplantation without immunosuppression. CBA mice made diabetic by the administration of a single dose of streptozotocin received either 350 uncultured or 7-day cultured BALB/c islets. Within 7-10 days of transplantation, the blood sugar levels of transplanted animals returned to the normal range. By 2-3 wk posttransplantation, rejection of the uncultured islets commenced and the blood sugar levels returned to pretransplant levels. Animals receiving cultured islet allografts maintained their nonfasting blood sugar levels either  differences only) * 5 /s intact at 420 days in or near the normal range. Removal of the kidney containing the islet graft resulted in a rapid return to the diabetic condition. 34 The animals lost between 13 and 29% of their body weight by 14 days after the induction of diabetes. Those that received cultured allografts regained their premorbid body weight by 60 days after transplantation. Animals grafted with uncultured islets showed initial weight gain but, with the return of hyperglycemia, weight loss commenced. Examination of the allografts carried by these animals 50 days after transplantation showed the tissue had been rejected. 34 It was found that animals grafted with 300 cultured islets had blood glucose levels on or slightly above the upper limit of the normal range. Those animals also excreted small amounts of glucose in their urine. 34 When animals were grafted with 350 islets, they had good control of blood sugar levels and were aglycosuric ( Figure 2). The above studies have shown that allotransplantation of adult islet tissue reverses chemically induced diabetes without a requirement for recipient immunosuppression. Reversal of insulitis-associated diabetes and spontaneous type I diabetes. It has been suggested that the inflammatory response seen in animals with insulitis induced by subdiabetogenic doses of streptozotocin 35 -36 may be the result of an autoimmune response to damaged islet tissue since infiltration of the islets by mononuclear cells occurred after destruction of 85% of the islet mass. 37 Recent studies 38 in the spontaneous diabetic BB rat have indicated that the response that destroyed the original pancreas may also destroy the islet tissue of a new graft. The cultured allografts will always contain damaged tissue, and this damaged tissue may restimulate immune processes which could destroy the graft.
To examine this possibility, cultured BALB/c islet clusters were grafted to CBA mice suffering from insulitis due to multiple doses of streptozotocin at various times after the induction of the disease. 39 All the grafted mice became normoglycemic within 7 days and remained so for the 100 days of the experiment. A group of nongrafted control diabetic mice remained hyperglycemic for this period.
The pathogenesis of chemically induced and spontane- ously arising diabetes may differ, and because of this we were very keen to transplant a case of spontaneous diabetes. We were given this opportunity, when as a part of another study, we came across a case of spontaneous diabetes in a CBA mouse. 39 This animal received 8 clusters of cultured BALB/c islets at 6 wk of age. Prior to transplant, the animal had a blood glucose level of 19 mM and a 24-h urine glucose excretion of 0.9 mmol. By 1 wk posttransplant, the animal became normoglycemic and aglycosuric (Figure 3), and remained so for the duration of the observation period (> 60 days). 39 FIGURE 3. Nonfasting blood glucose levels and body weights of a spontaneous diabetic CBA mouse grafted with 400 cultured BALB/c islets. The shaded area represents the 95% confidence interval for nonfasting blood glucose levels in normal CBA mice. Squares, blood glucose; triangles, body weight.

60
In the above two experimental models, the cultured allografts did not succumb to the same fate as the original pancreas. Thus, it may be possible to achieve successful allograft survival in man without the need for prolonged immunosuppression. However, it is important to be aware that some forms of diabetes do have an autoimmune etiology that can affect the transplanted tissue. 38 Xenotransplantation of islet tissue. Organ culture of thyroid prior to transplantation has been shown to facilitate xenograft survival in nonimmunosuppressed recipients. 25 Cultured xenogeneic islets have also been successfully transplanted from rats to mice that have been treated with antilymphocyte serum (ALS) at the time of transplantation. 40 We have not been successful in prolonging survival of xenografts of cultivated mouse islet tissue grafted to rats. By 4 wk posttransplant, most of the islet tissue was disorganized and only scattered fragments of islets could be seen. One injection of rabbit anti-mouse lymphocyte serum or rabbit anti-rat lymphocyte serum at the time of transplantation did not affect graft survival. While cultivation alone may enhance survival of xenografts it is likely that immunosuppression will be needed to obtain prolonged survival.

TRANSPLANTATION OF FETAL TISSUE
Introduction. It is clear from the above discussion that allotransplantation of adult islet tissue can reverse either chemically induced diabetes or spontaneous diabetes. Moreover, extensive recipient immunosuppression is not required to facilitate al log raft acceptance. While these findings are most encouraging from a clinical standpoint, technical problems associated with the isolation of islets from the adult pancreas at present prohibit the use of this procedure in man. For this reason, attention has turned to the fetal pancreas as a source of islet tissue for transplantation. In rodents, a single fetal pancreas isograft will reverse diabetes, 41 and Mandel et al. 42 have shown that a fetal pancreas isograft will prevent the development of vascular lesions associated with diabetes. We have found the fetal pancreas to be a much more difficult tissue to manage because of its greater immunogenicity.
Organ culture of fetal pancreas. At 17 days of gestation, the fetal mouse pancreas consists of a small number of immature islets, some duct tissue, and a considerable amount of acinar tissue at various stages of differentiation. In addition, careful examination of histologic sections has identified primitive lymphoid tissue closely associated with fetal pancreas explants. 30 After 10 days in culture, the fetal pancreas condenses into a compact mass. Compared with the uncultured tissue, the relative proportion of islet tissue is increased and the amount of lymphoid tissue is substantially reduced. After 17 days of organ culture, acinar tissue has completely degenerated and all that remains is well formed islets, duct tissue, and connective tissue. However, histologic examination of isografts at varying times after transplantation has shown a decline in islet survival after 20 days in culture. This was reflected in the slow and somewhat variable recovery of function after transplantation of 20-day cultured tissue to diabetic recipient mice, as compared with the function of uncultured tissue. 41 This deterioration in the survival and function of isografts of fetal mouse pancreas following pro-longed organ culture in an oxygen-rich atmosphere is probably attributable to oxygen toxicity. Immunogenicity of adult islets and fetal pancreas. In contrast to the results with the adult mouse islets, allografts of fetal mouse pancreas that have been cultured for 10 days are acutely rejected. At 2 wk posttransplantation, the tissue shows obvious signs of rejection with damaged islets visible. At 4 wk after transplantation, all that remained at the graft site was scar tissue and a mononuclear cell infiltrate. 30 Organ culture is much more effective in preparing adult islet tissue for transplantation than it is in the case of the fetal pancreas. It is likely that this difference is attributable to the major lymphoid component associated with the fetal tissue. This lymphoid tissue is not actually in the pancreas itself, but appears to represent developing lymph nodes associated with the mesentery that surrounds the rather diffuse pancreas in situ. 30 Such an explanation would account for the development of a large lymph node-like structure following isotransplantation of the uncultured fetal pancreas. Organ culture before transplantation damages this lymphoid tissue and lymph nodes do not develop after isotransplantation. However, we have observed lymphoid remnants in association with 10-day cultured fetal pancreas and it is possible that this residual lymphoid tissue triggered the rejection process. It is possible to successfully transplant fetal pancreas to nonimmunosuppressed recipients when a longer (17-20-day) period of organ culture is used. 30 However, as indicated above, the functional capacity of this tissue decreased as the culture period was prolonged. 41 Fetal islet precursors can be isolated from the fetal pancreas by modification of the procedure described by Hellerstrom et al. 43 Fetal pancreas explants are subjected to a brief collagenase digestion, washed, and gently stirred. The digested fetal tissue is then cultured for 4 days in 5% CO 2 in air at 37°C. Histologically, the 4-day cultured tissue consists of differentiating islet tissue ('proislets') and some duct endothelium. Pancreatic exocrine tissue degenerates during culture and is totally absent by 4 days. 44 Four-day cultured fetal mouse proislets are not highly immunogenic and can be successfully allotransplanted across a major histocompatibility barrier. At 2 and 4 wk posttransplantation, 6 /g and 5 /s 4-day cultured proislet allografts, respectively, contained differentiated islet tissue without any evidence of mononuclear cell infiltration. The remaining allografts were rejected, and occasionally a mononuclear cell infiltrate was identified in the presence of apparently intact islet tissue. Six of six 4-day cultured proislet isografts examined histologically at 2 and 4 wk after transplantation also consisted of new differentiated islets and some duct tissue. 44 Isotransplantation of proislets derived from 8 fetal pancreases will reverse streptozotocin-induced diabetes within 40 days (unpublished data).
It is evident that the immunogenicity of fetal pancreas allografts is not attributable to the intrinsic immunogenicity of the fetal proislets but to some other component of the fetal pancreas. This finding supports our earlier postulate that the immunogenicity of fetal pancreas allografts is probably due to the mass of lymphoid tissue which is associated with the fetal pancreas explant. 30 In view of the difficulty in conditioning fetal pancreas for allotransplantation, the isolation of fetal proislets offers a promising approach to successful clinical allotransplantation of fetal islet tissue.

SUSCEPTIBILITY OF THE ESTABLISHED ALLOGRAFT TO REJECTION
Effect of specific and nonspecific stimulation of the immune system. Organ culture before transplantation reduces immunogenicity of the tissue but does not destroy tissue antigens; cultured allografts are consistently rejected when the recipient is immunized with lymphoid cells of donor type at the time of transplantation. 3i30i33i4S While cultured allografts can be transplanted without a need for suppression of the recipient's immune system, such grafts are constantly under the threat of rejection. Nonspecific stimulation of the recipient's immune system could raise the level of costimulator activity and trigger irreversible rejection. In the situation of clinical human transplantation, transfusion of blood could trigger rejection if the transfused blood carried histocompatibility antigens similar to those in the graft. These potential threats to the established islet allograft were examined in the following studies.
To examine the effect of nonspecific stimulation of the immune system CBA mice which had been carrying functional BALB/c islet allografts for 150 days were given complete Freund's adjuvant intraperitoneally. This provided a nonspecific stimulation of the recipient's immune system. Blood sugar levels were observed for 35 days, and none of the animals showed any evidence of a return to the diabetic state. 46 When the allografts were removed by nephrectomy, the animals became diabetic. Thus, nonspecific stimulation of the recipient's immune system does not trigger rejection of an established cultured allograft.
The effect of an antigen-specific stimulation was examined in a group of CBA mice whose diabetes had been reversed by allotransplantation of cultured BALB/c islet tissue 100 days previously. These animals were challenged by the injection of 10 5 peritoneal cavity cells from mice of BALB/c donor strain. Two animals promptly became diabetic and their transplants were seen to have been rejected when examined histologically. The remaining two animals showed no evidence of rejection when challenged with 10 6 and 10 7 donor peritoneal cavity cells. 46 When the grafts were removed by nephrectomy the animals became diabetic, demonstrating the dependence of blood sugar control in these animals on the graft tissue. 46 This finding demonstrated that antigen specific stimulation can trigger graft rejection although a proportion of the animals failed to reject their grafts following challenge. This may be due to the development of specific tolerance or allograft enhancement and will be considered in more detail below. Susceptibility of grafts to antibody-mediated rejection. As outlined above, organ culture in an oxygen-rich atmosphere reduces the immunogenicity of a tissue allograft. During the culture period, the vascular endothelium is destroyed. 21 When the cultured tissue is transplanted, revascularization with host vascular endothelium occurs. It has been suggested that antibody and complement-mediated hyperacute rejection results from an interaction of antibodies with the vascular endothelium of the graft. 47 When CBA mice carrying revascularized BALB/c islet allografts were injected with CBA anti-BALB/c alloantiserum alone or alloantiserum and rabbit complement, none of the animals showed any evidence of a return to the diabetic state, even though high titers of circulating cytotoxic antibody were demonstrated up to 7 days after the injection of al-loantiserum. 48 Histologic examination of the graft tissue after removal showed normal granulated beta-cells. It would appear, therefore, that revascularization of a graft with host endothelium may not protect the graft from cell-mediated damage, although it can protect the graft from antibody and complement-mediated attack. Establishment and induction of toierance in allografted animais. The above findings indicated that mice that had been carrying islet allografts may develop tolerance to donor transplantation antigens. The development of unresponsiveness has been studied in more detail in the thyroid allograft model. 3 The results of this study 49 clearly showed that a state of partial tolerance developed in recipients of preconditioned thyroid allografts. After prolonged presence of the cultured allograft in the recipient, a proportion of allografts were not rejected when the recipient was challenged with peritoneal cells of donor origin. This proportion may be influenced by the way the challenge dose is administered. The adaptation of the transplant to its host cannot be explained in terms of antigeneic modulation of the graft. Grafts taken from animals that resist rejection were destroyed by an allograft response when transplanted to naive young CBA animals that were then challenged with BALB/c peritoneal cells. Clearly, the long-standing graft does contain potentially recognizable antigen. One must therefore conclude that the recipient animal is modified, and appears to be tolerant of donor strain antigen. This effect was similar to that seen in the case of an established islet allograft. 46 However, thyroid grafted animals do not become unresponsive to alloantigen until the transplant has been in the recipient for a considerable period of time (>100 days <350 days).
The state of tolerance was not due to the deletion of antigen-reactive cells. Recipients of long-standing allografts were hyporesponsive to in vivo challenge with donor antigen. However, in vitro mixed leucocyte reactivity to donor antigens was essentially normal. Although some recipients of long-standing thyroid allografts appear to be tolerant of donor antigen, this tolerance is not complete. Animals that failed to reject their primary allograft, even after repeated antigeneic challenge with donor peritoneal cells, responded quite acutely to a second uncultured BALB/c allograft. This inflammatory reaction can severely disrupt allograft integrity. At this stage we cannot be certain of the mechanism underlying this partial tolerance.
When mice that have been carrying islet allografts for 30 days are challenged with 10 6 donor spleen cells, the grafts were rejected. However, if these mice were given 10 6 ultraviolet-irradiated donor spleen cells weekly for 3 wk before the injection of 10 6 followed 1 mo later by 10 7 donor spleen cells, the grafts did not reject. This demonstrated a marked difference in the immunogenicity of uv-irradiated and living spleen cells and is consistent with the finding that uv-irradiated cells are unable to stimulate in a mixed lymphocyte culture. 5 The finding that grafts also reject if 10 6 donor spleen cells are given on day 60 showed that the injection of uv-irradiated cells induces a form of allograft tolerance. That is, allograft tolerance can be induced in adult animals without a requirement for recipient immunosuppression. 49 Although we have no firm evidence concerning the nature of this tolerance or exactly how it is induced, it would appear that the way antigen is delivered to the immune system profoundly influences its response, and that the slow, constant presentation of alloantigen to the recipient can induce a state of tolerance.