Ontogenic development of B-lymphocyte function and tolerance susceptibility in vivo and in an in vitro organ culture system.

The maturation of B-lymphocyte function during fetal development was studied in vivo and in an in vitro organ culture system. The results indicated that the progenitors for 2,4-dinitrophenol (DNP)-specific B cells are present as early as 14 d of gestation in liver and possibly as early as 15 d in spleen. In addition, it was found that the organ culture system supports the development of B lymphocytes as measured by an increase in both the percentage of surface immunoglobulin-positive cells and the frequency of clonable DNP-specific B cells after culturing. The majority of anti-DNP-secreting clones resulting from the antigenic stimulation of fetal B cells produced only the IgM isotype, and the ability to secrete the IgG isotypes increased as a function of gestational age. Because fetal DNP precursors from spleens and livers that had been incubated in organ culture resulted in a greater proportion of clones secreting IgG compared with age-matched uncultured controls, it was concluded that the maturation with regard to the ability to secrete IgG can occur in vitro. In studies relating to the ontogenetic development of tolerance susceptibility, it was found that up to one-half of the DNP-specific B-cell precursors from livers and spleens less than 18 or 19 d of gestation were resistant to tolerogen treatment for 24 h as if in a pretolerant phase. However, if tolerogen were present for 3--5 d during organ culture there was near total elimination of potential DNP clones. This finding suggested that the 24-h induction period was insufficient for affecting the DNP-specific precursors in livers and spleens from the earlier gestational ages, and that a proportion of precursors could subsequently form DNP clones in the splenic focus assay after the removal of tolerogen.

Growth of B-lymphocyte colonies in vitro.

In semisolid agar cultures containing mercaptoethanol, cells from the spleen, lymph nodes, marrow, peritoneal cavity, thoracic duct, and blood of normal mice generated clusters and colonies of up to 3,000 cells. Colony numbers and growth were markedly enhanced by the addition of sheep red cells. The frequency of colony-forming cells in the spleen or lymph nodes was 0.5-2.0%, and cluster forming cells were approximately five times more numerous. The mononuclear cells comprising these colonies had the electronmicroscopic morphology of immature lymphoid and plasma cells. The majority of the cells possessed Fc receptors, 61-69% reacted with anti-mu-serum and 4-11% with anti-gamma2-serum. Analysis of single cells from individual colonies indicated a higher frequency of the cells with membrane immunoglobulin and a clonal pattern of anti-mu or anti-gamma-reactivity. The clonal nature of colonies was supported by an analysis of NIP-binding cells in colonies grown from CBA spleen cells enriched for NIP-binding cells. Mass-harvested colony cells synthesized immunoglobulin in short-term liquid cultures. It is concluded that the colonies are clones of functionally active B-lymphoid cells.

Colonies formed in agar from human breast cancer and their identification as T-lymphocytes.

Single cell suspensions prepared from human breast cancer specimens by collagenase digestion were cultured in soft agar with phytohemagglutinin-stimulated human lymphocyte-conditioned medium (PHA-LCM). In 6 of 10 different tumors, PHA-LCM-dependent clonal growth was develop. After 12-14 days of incubation, two morphologic types of colony containing 20-500 cells were recognized. Both were composed of lymphocytes of T-cell nature, as judged by cell morphology in smears, cytochemical properties, capacity to form rosettes with sheep erythrocytes, and electron microscopic appearances. Contamination of the tumor cell suspensions by blood could be excluded as a source of the colony-forming lymphocytes, and the incidence of colony-forming cells correlated well with the degree of lymphocyte infiltration of the tumors. Some of the colonies in agar were expanded further in liquid culture in the continuous presence of PHA-LCM. These clones were apparently high in proliferating capacity as compared with the proliferating activity of peripheral T-cell clones obtained from normal blood. These clones were considered to be highly activated T-lymphocytes and to be stimulated to grow in vitro by the T-cell growth factor contained in PHA-LCM. The direct cloning and expansion of such activated T-lymphocytes infiltrating the tumors will be useful for studies on the functional characteristics of these cells.

Progression from insulitis to beta-cell destruction in NOD mouse requires L3T4+ T-lymphocytes.

The identity of the cells responsible for beta-cell destruction in type I (insulin-dependent) diabetes is still uncertain. L3T4+ T-lymphocytes have a role in the initiation of insulitis and in damaging transplanted allogeneic islets in nonobese diabetic (NOD) mice. The role of L3T4+ T-lymphocytes in destruction of beta-cells of the NOD mouse was studied in cyclophosphamide (CY)-induced diabetic NOD mice with a rat anti-L3T4 monoclonal antibody (MoAb). After administration of CY, most untreated animals became diabetic, whereas all antibody-treated animals remained normoglycemic. Insulitis was still present in MoAb-treated animals, but immunocytochemical staining showed rat antibody blocking the L3T4 antigen on T-lymphocytes. This study provides further evidence that L3T4+ T-lymphocytes are critical to the process of beta-cell destruction in NOD mice. The means by which L3T4+ cells exert their effect remains to be clarified.

Effect of culture conditions on fetal mouse pancreas in vitro and after transplantation in syngeneic and allogeneic recipients.

Organ culture of fetal mouse pancreas under conditions designed to either maximally preserve islet cell survival or reduce immunogenicity was used to test the efficacy of islet graft function in diabetic syngeneic or fully allogeneic recipients. Organ culture in either 90% O2, or at low temperature (22 degrees C), or in a combination of 90% O2 and 22 degrees C for 14 days not only failed to reduce immunogenicity in fully allogeneic recipients but also diminished endocrine cell survival in grafts in syngeneic recipients. Monitoring of in vitro insulin secretion provided a better guide to future graft function than did the insulin content of the cultured tissue. It is concluded that culture conditions that have been shown to reduce immunogenicity (high O2 concentration and low temperature) are not synergistic when used together and are also potentially damaging to endocrine cells.

Induction of insulitis in athymic (nude) mice. The effect of NOD thymus and pancreas transplantation.

The NOD mouse is a model for human insulin-dependent diabetes mellitus. The disease is thought to have an autoimmune etiology because it is T-cell dependent and is characterized by mononuclear cell infiltration in and around the pancreatic islets of Langerhans. The mechanism by which autoreactive T-cells are generated is not fully understood, but it has been postulated that there is a breakdown in self-tolerance induction during intrathymic T-cell maturation. The aim of these studies was to determine whether transplantation of NOD thymus into diabetes-resistant mouse strains would generate islet-reactive T-cells. Neonatal thymus was pretreated either by irradiation or culture in 2-deoxyguanosine (dGua) and then transplanted into athymic BALB/c, CBA, and C57BL/6 nude mice. Generally, insulitis was not seen in the CBA or C57BL/6 recipients, but was found in 56% of BALB/c mice transplanted with an irradiated NOD thymus and in 46% BALB/c mice with a dGua-treated thymus. Similar experiments in which a NOD fetal pancreas was transplanted into nude BALB/c mice before NOD thymus transplantation showed a similar frequency and severity of insulitis in both the host pancreas and grafted NOD pancreas. This suggests that NOD islets are no more prone than the host islets to autoimmune attack and do not exacerbate insulitis. Overall, the data suggest that a defect of thymic origin (and correlating with the thymic epithelium) in the NOD mouse can lead to the development of autoreactive T-cells and specific islet cell damage. Autoreactivity appears to be restricted to the H-2Kd allele that is shared by NOD and BALB/c mice.

Insulin secretion by fetal human pancreatic islets of Langerhans in prolonged organ culture.

Fetal human pancreata, obtained from legally-induced abortions, were placed in organ culture to study their capacity to produce insulin over periods in vitro of between 18 and 40 days. Specimens obtained by hysterotomy usually showed increasing insulin secretion as measured by the insulin content of the media at each twice-weekly medium change. In most instances, relatively little insulin was produced during the first 7-10 days, but thereafter insulin secretion rapidly increased. In contrast, most specimens from prostaglandin-induced abortions showed high levels of insulin in the medium early in the culture period but little thereafter, indicating rapid release of insulin from damaged tissue. In some instances, after early insulin release by damaged tissue, some recovery of islet function occurred and insulin secretion again increased. The presence of differentiated endocrine cells was confirmed histologically. Tissue from each fetus was placed in a number of petri dishes, and for each individual fetus a qualitatively similar pattern of secretion was noted in each dish. These data suggest that representative and nondestructive monitoring of islet function is possible and may be important before such tissue is considered for use in islet transplantation in insulin-dependent diabetes.

Effect of ischemia and temperature on fetal mouse pancreas. Insulin production in vitro, and function after isotransplantation.

The effects of ischemia at varying temperatures on the survival of fetal islet endocrine cells was investigated by placing 17-day-old fetal mouse pancreata in organ culture after 2, 4, or 6 h at either 4 degrees C, 22 degrees C, or 37 degrees C. Insulin secretion by the cells in vitro, the content of insulin in the cultured pancreata, and the ability of the cultured islets to reverse diabetes in syngeneic streptozotocin-diabetic mice were assessed. Fetal pancreas subjected to 2-6 h of ischemia at either 4 degrees C or 22 degrees C showed neither loss of insulin secretory capacity in vitro nor loss of ability to produce large functional grafts, and behaved identically to tissue not subjected to deliberate ischemia. In contrast, after 2 h of ischemia at 37 degrees C, although some grafts functioned, their insulin content was reduced despite apparently normal prior insulin production in vitro, but 4 or 6 h at 37 degrees C resulted in total loss of functional islet tissue. However, despite retention of functional capacity and the ability to produce large grafts with high insulin content after cold or room temperature ischemia, some loss of insulin storage capacity in vitro was noted by islets subjected to ischemic periods longer than 2 h even at 4 degrees C. Thus, fetal pancreas can withstand prolonged periods of ischemia provided its temperature is reduced, and functional recovery can be demonstrated after transplantation.

Administration of silica particles or anti-Lyt2 antibody prevents beta-cell destruction in NOD mice given cyclophosphamide.

The cellular pathway of beta-cell destruction in type I (insulin-dependent) diabetes is still undefined. L3T4+ T-lymphocytes have a role in both the initiation of insulitis and in recurrent disease in transplanted allogeneic islets in nonobese diabetic (NOD) mice. The roles of macrophages and Lyt2+ T-lymphocytes in beta-cell destruction were studied in cyclophosphamide-induced diabetic NOD mice with silica particles and a rat anti-Lyt2 monoclonal antibody. After administration of cyclophosphamide, 10 of 26 untreated mice and 1 of 21 anti-Lyt2-treated mice became diabetic. Insulitis was significantly reduced in anti-Lyt2-treated mice, and immunocytochemical staining showed a lack of Lyt2+ cells. Only 1 of 19 silica-treated mice became diabetic, compared to 8 of 19 control mice. This study demonstrates that both Lyt2+ T-lymphocytes and macrophages are necessary, but not sufficient, for beta-cell destruction in NOD mice. Therefore, we propose that macrophages present beta-cell antigen to L3T4+ cells, which induce cytotoxic Lyt2+ cells to specifically destroy beta-cells.

Prevention of cyclophosphamide-induced diabetes by anti-V beta 8 T-lymphocyte-receptor monoclonal antibody therapy in NOD/Wehi mice.

Walter & Eliza Hall Institute nonobese diabetic (NOD/Wehi) mice exhibit a low incidence of spontaneous diabetes mellitus, but one large dose of cyclophosphamide (CY) can lead to a rapid progression to overt diabetes. Macrophages and Lyt-2+ and L3T4+ cells have been demonstrated to be involved in beta-cell destruction in this model. The role of a specific subset of T-lymphocytes expressing a particular T-lymphocyte-receptor segment was examined in CY-induced diabetic NOD mice with a mouse anti-V beta 8 T-lymphocyte-receptor monoclonal antibody (F23.1). After administration of CY, only 4 of 51 treated mice became hyperglycemic compared to 23 of 47 untreated mice, 13 of 26 mice treated with an isotype-matched control ascites, and 4 of 6 mice given antibody-negative ascites. Insulitis was significantly reduced in the F23.1-treated group, and immunocytochemistry revealed the absence of V beta 8 expression on cells in the lymphoid organs and insulitis of these mice. This investigation revealed that V beta 8+ cells were implicated in CY-induced diabetes in NOD/Wehi mice.

Cyclophosphamide-induced diabetes in NOD/WEHI mice. Evidence for suppression in spontaneous autoimmune diabetes mellitus.

Nonobese diabetic (NOD) mice spontaneously develop a lymphocytic infiltration of pancreatic islets (insulitis) that may progress to overt diabetes. Virtually all NOD/WEHI mice develop insulitis, but very few progress to diabetes. However, cyclophosphamide (CY) can promote the onset of diabetes in NOD mice, including the NOD/WEHI strain. The means by which CY produces diabetes was investigated in NOD/WEHI mice, in which it was hypothesized that active suppression mechanisms prevented the progression from insulitis to diabetes. A study of the time course of insulitis in the islets after CY was given showed that insulitis was initially reduced but rapidly increased over 16 days, and T-lymphocytes were predominant in the lesion. This suggested a compression of the normal time course of the disease seen in NOD mice. CY did not produce diabetes in any of 11 non-NOD strains studied. Fetal isografts in NOD mice given CY several days before were subjected to lymphocytic infiltration and beta-cell destruction. These findings suggested that CY was not directly beta-cell toxic and that altered beta-cells were not essential for beta-cell destruction. This was further demonstrated with subdiabetogenic doses of streptozocin, which significantly damaged beta-cells but did not increase the severity of insulitis or induce diabetes as did CY. Most important, the transfer of mononuclear cells from nondiabetic NOD mice to mice given CY prevented diabetes, which indicated that the likely effect of CY was via immunomodulation, possibly by allowing poised effector cells to act on beta-cells.(ABSTRACT TRUNCATED AT 250 WORDS)

Comparison of high- and low-diabetes-incidence NOD mouse strains.

The nonobese diabetic (NOD) mouse is a model of insulin-dependent diabetes mellitus. These mice develop insulinopenia and hyperglycemia secondary to beta-cell destruction, which is associated with insulitis and autoantibody production. We have two strains of NOD mice: a low-incidence strain (NOD/Wehi), in which less than 10% females and less than 1% males develop diabetes by 150 days despite intense insulitis, and a high-incidence strain (NOD/Lt), in which most females and many males develop diabetes by 150 days. This phenotypic difference has been maintained for 24 mo despite identical housing in our specific pathogen-free unit. Reciprocal skin grafting and allozyme electrophoresis have not identified a difference between the strains. Mixed-lymphocyte cultures were performed with splenic T-lymphocytes cultured with equal numbers of irradiated stimulator splenocytes for 3-6 days. NOD/Wehi mice demonstrated a heightened syngeneic mixed-lymphocyte response (SMLR), averaging 19% of the allogeneic response to CBA/CaHWehi cells. The response to NOD/Lt stimulator cells was not significantly different from the syngeneic response. In contrast, NOD/Lt mice had an SMLR similar to that of BALB/cAnBradleyWehi control mice, averaging 5% of the allogeneic response. NOD/Lt cells also responded similarly to NOD/Wehi stimulator cells and briskly to allogeneic cells. The heightened SMLR in NOD/Wehi mice may reflect active generation of suppressor function, and this may account for the low incidence of diabetes.

Organ culture of fetal mouse and fetal human pancreatic islets for allografting.

In organ culture of fetal human and fetal murine pancreas under "conventional" conditions (10% CO2 in air), the islet cells of both species survive, proliferate and function but the acinar tissue rapidly degenerates. Fetal mouse islet cells also survive in 90% O2 and 10% CO2 but nonendocrine cells, including fibroblasts and macrophages, degenerate. Fetal mouse islets grown in 90% O2 show diminished immunogenicity when transplanted into recipients differing across the entire MHC, but a reduced allograft response by the host is frequently still present in the absence of immunosuppression. Fetal human islets, grown in 10% CO2 in air, produce insulin in vitro for prolonged periods, and as xenografts, differentiate under the kidney capsule of athymic mice, suggesting that under appropriate conditions both in vitro and in vivo, the fetal human islets can survive. However, fetal human pancreatic cells of all types are highly susceptible to high oxygen concentrations and are rapidly killed. Because of the susceptibility of fetal human pancreas to oxygen, conditions for the culture of fetal human islets for allotransplantation may need to be modified from those tolerated by fetal mouse islets. Fetal human islets may be a useful source of transplant material in human insulin-dependent diabetes, but it is likely that tissue matching and immunosuppression may be required in addition to modification of islet immunogenicity by prior organ culture.

Long-term survival of segmental pancreas isografts in NOD/Lt mice treated with anti-CD4 and anti-CD8 monoclonal antibodies.

Spontaneously diabetic nonobese diabetic (NOD/Lt) mice were treated with anti-T-cell monoclonal antibodies (mAbs) at the time of grafting with vascularized segmental pancreas isografts. Recipients were either untreated or given anti-CD4 and/or anti-CD8 mAbs (0.5 mg/20-g mouse on each of 4 consecutive days), which reduced target cell levels to <5% of normal. Graft function was monitored by measuring blood glucose (BG) levels. Transplants were removed for histological examination when BG returned to >20 mmol/l for two consecutive readings. Isografts from 3- to 4-week-old prediabetic mice placed in untreated diabetic NOD mice ceased functioning in 9-13 days with a mean survival time (MST) +/- SD of 10 +/- 2. Treatment with anti-CD4 prolonged survival significantly (MST = 61 +/- 35 days, P < 0.05 compared with untreated control mice). Anti-CD8 treatment was less effective, but it still significantly improved graft survival (MST = 24 +/- 9 days, P < 0.05 compared with untreated control mice). Anti-CD8 plus anti-CD4 treatment was highly effective in inhibiting autoimmune destruction of the grafts (MST = 97 +/- 8 days). This clearly demonstrates that transient inactivation of most T-cells with anti-CD4 plus anti-CD8 mAbs effectively controls autoimmune disease in the isograft, despite recovery of CD4 and CD8 T-cells to normal levels. Although insulitis developed in the long-term grafts, insulitis scores did not increase between 33 and 100 days, and none of the mice progressed to IDDM in 100 days. Histology showed a predominantly peri-islet T-cell and macrophage infiltrate with ductal expression of the cytokines interleukin (IL)-4, IL-2, and interferon-gamma. There was little infiltrate or expression of cytokines within the islets. Thus, mAb treatment at the time of grafting allowed isograft survival and prevented progression from insulitis to beta-cell destruction.

Fetal islet xenotransplantation in rodents and primates.

Beta cell replacement in IDDM by transplantation of either isolated adult islets of Langerhans or of proliferating immature islet tissue from fetal pancreas are potential ways of curing this disease. Because of the dearth of human cadaver donors adult allogeneic islets are scarce and in most Western societies availability of human fetal tissue of suitable maturity is also uncommon. The use of xenogeneic islets from domestic species already widely used for human consumption, e.g. pigs, could overcome this scarcity but xenogeneic tissues are faced with major problems of graft rejection. Hyperacute rejection (HAR) is the main cause of destruction of immediately vascularised xenografts and is caused by the interaction of natural cross-reactive antibodies with donor endothelial cells. Neovascularized islet grafts do not have donor EC as the target for HAR and are not subjected to this problem but are still acutely rejected. The mechanism of this destruction is still poorly understood but is clearly T cell dependent. However, current immunosuppression that is usually adequate for control of allograft rejection generally does not prevent xenograft rejection. A better understanding of the ways in which xenoantigens are recognised and of the nature of the immune response they initiate is fundamental to the development of appropriate strategies for the safe and effective control of xenograft rejection. The studies summarized herein describe the response of mice and primates to a challenge with fetal pig pancreas grafts. The rejection response that develops is different from that seen against a challenge with fetal allogeneic islets. Although the xenograft response is highly T cell dependent the actual effectors of graft damage appear to be different from those that provoke allograft destruction and include macrophages and granulocytes, particularly eosinophils, and possibly non-classical T cells.

The presence of mast cells in agar cultures.

Using a specific stain and electron microscopy, small numbers of mast cells were detected in human bone marrow cultures. However, they were not detected in agar cultures containing murine bone marrow cells. In bone marrow cultures from three patients with acute myeloid leukemia the number of mast cells was elevated.

The in vitro production of chimeric murine thymus from nonlymphoid embryonic precursors.

A simplified technique for the production of chimeric thymus completely in vitro from the third pharyngeal pouch (TPP) of 10-day mouse embryos has been examined. The results are similar to those obtained by a more cumbersome technique published by others. This modified technique permits the production of large numbers of synchronized chimeric thymuses, so that sufficient numbers of thymocytes are available for multiple functional assays. TPP may also be maintained in culture, then grafted into normal or thymectomized, irradiated, fetal liver-reconstituted hosts where they undergo development into histologically normal lymphoid thymic tissue.

Fetal pig pancreas. Preparation and assessment of tissue for transplantation, and its in vivo development and function in athymic (nude) mice.

The possibility of using xenogeneic islets for transplantation in insulin-dependent diabetes mellitus (IDDM) necessitates characterization of their potential for growth and functional differentiation. Fetal pig pancreas (FPP) of various gestational ages was examined with respect to morphology, ability to produce insulin before and during culture, and development and function in nude mice. Insulin-containing beta cells were present, but distinct islets were not apparent in FPP even in late gestation, and did not develop during culture. FPP remained viable and produced insulin for up to 30 days in vitro. Mitotic figures were seen in cultured tissue. Culture on a gelfoam raft resulted in more viable tissue than free-floating culture. Culture in a high concentration of O2 (90% O2/10% CO2) was detrimental compared with culture in 10% CO2 in air. Responses to static incubation in secretagogues showed that IMBX, theophylline, and tolbutamide all stimulated insulin secretion, but high glucose concentration (5 g/L), arginine, and leucine did not. The potential of this tissue for growth and its ability to regulate blood glucose levels appropriately were tested in athymic (nu/nu) mice. Pancreatic tissue from fetuses as young as 4 weeks gestation showed growth after transplantation into athymic mice, with representation of the major pancreatic endocrine cells demonstrated by selective immunochemical staining. The increase in the size of the grafts showed an impressive proliferative capacity, and histology confirmed mitotic activity and islet structure in the graft. The amount of endocrine tissue in grafts reflected the condition of the explants at the time of grafting, and prolonged culture times were detrimental to eventual graft size. Functional capability of the grafted FPP to release insulin in response to hyperglycemia was tested by transplantation into mice made diabetic with streptozotocin. Blood glucose levels, animal weights and survival, and the histological appearance of the tissue after graft nephrectomy indicated that either fresh tissue or tissue cultured for up to 8 days (Gelfoam; 10% CO2 in air) had better eventual graft function then FPP grown in 90% O2 or transplanted as a secondary graft following an interim period to allow gestational maturation in a nondiabetic nu/nu host. Return to euglycemia took 3-4 months after transplantation of FPP. The in vitro characteristics of FPP are similar to those reported for human fetal tissue, and since FPP is capable of growth and proliferation in vivo and has the ability to normalize hyperglycemia, further investigation of FPP to establish its suitability as a source of xenogeneic insulin-secreting tissue for human transplantation is warranted.

A comparison of corneal, pancreas, and skin grafts in mice. A study of the determinants of tissue immunogenicity.

A model of murine heterotopic allogeneic transplantation was used to study the rejection characteristics of three tissues--adult cornea, fetal pancreas, and fetal skin--for attributes that might explain their variation in rejection rates and help define the determinants of graft immunogenicity. Under identical conditions, tissues were transplanted to the renal subcapsular space and their base-line rejection rates compared. The expression of MHC class I and II and intercellular adhesion molecule-1 (ICAM-1), was determined for each tissue, as was their ability to produce interleukin-6, IL-3, interferon-gamma, and granulocyte-macrophage colony-stimulating factor in vitro. These studies were performed under basal conditions and after stimulation with concanavalin A-stimulated spleen cell supernatant (CAS) or INF gamma. Corneal grafts had a slow rejection rate compared with pancreas and skin. While all three tissues had low basal expression of MHC class II, both fetal skin and pancreas, but not adult cornea, were able to increase this under our experimental conditions. Pancreas and skin produced IL-6 under basal conditions and could be stimulated to increase production 2-3-fold but the cornea did not basally produce IL-6 and showed minimal upregulation. We postulate that delayed corneal rejection, compared with pancreas and skin, results from two compounding deficiencies: the relative lack of class II MHC-positive APC and the inability to overcome this deficiency by upregulating class II expression and producing accessory molecules for antigen presentation.

Reduction of the immunogenicity of fetal mouse pancreas allografts by organ culture in 90 percent oxygen.

BALB/c female streptozotocin-diabetic mice were transplanted under the left kidney capsule with syngeneic or allogeneic CBA 17-day fetal pancreases that had been organ-cultured for 23 days. Two isografted groups received a single graft per recipient, cultured normally (90% air, 10% CO2), or in an oxygen-rich atmosphere (90% O2, 10% CO2): 13/14 and 12/13 respectively became normoglycemic. The 3rd and 4th groups were transplanted each with 1 and 3 allogeneic fetal pancreases per recipient, which had been cultured in 90% O2: 5/12 and 2/13 mice respectively had surviving allografts after 104 days, and 2/5 of the former and both of the latter had become normoglycemic. Five out of the seven surviving allografts contained foci of infiltrating lymphocytes. After removal of the kidneys containing the grafts, surviving mice were retransplanted under the right kidney capsule, each with 2 CBA fetal pancreases which had been cultured for only 4 days under normal conditions--hence they were fully immunogenic. Rejection of the secondary allografts was delayed in the long-term, primary allograft acceptors, suggesting that a degree of tolerance had developed. This study demonstrates that it is possible for a single, cultured fetal pancreas to survive indefinitely and reverse diabetes after transplantation to a fully allogeneic, nonimmunosuppressed recipient.