Vascularized Islet Cell Transplantation in Miniature Swine: Islet-Kidney Allografts Correct the Diabetic Hyperglycemia Induced by Total Pancreatectomy

Minor and fully major histocompatibility complex (MHC)-mismatched transplant donors (>1.5 years of age) and recipients (3–5 months of age) were selected from our herd of partially inbred miniature swine (8,9).

Islet donors were adult miniature swine, subjected to either total or partial pancreatectomy, as described elsewhere (31). For partial pancreatectomy, ∼30% of the pancreas (the head and the remainder of the body) was left intact. For total pancreatectomy, the pancreas was carefully dissected so as to not damage the duodenum or the pancreas. The excised pancreas was placed in crushed ice immediately, and modified University of Wisconsin solution was infused through the pancreatic duct. Pancreatic tissue was processed at the Joslin Diabetes Center (Boston, MA) (27). After islet isolation, the purity of the preparation was evaluated with dithizone staining. Islet cell viability was assessed using dual fluorescence staining with acridine orange (AO; Sigma) and propidium iodide (PI; Sigma). The islets were cultured for 3 days, counted, and then transplanted either beneath the renal capsule or via the superior mesenteric vein.

IKs were prepared in donors as previously described (31). Briefly, islets were suspended in 1.0 ml saline and injected beneath the renal capsule using a 20G angiocatheter. IKs were prepared 2–3 months before the intended date of allogeneic transplantation

Type 1 diabetes was induced surgically by total pancreatectomy 3–7 days before either composite IK or direct islet transplantation. Total pancreatectomy was performed using a similar procedure to that described above for islet donors. Blood supply to duodenum and bile ducts was well preserved, and no blood transfusions were required (Fig. 1). Thereafter, these animals received pancreatase (Pancrezyme; Daniels Pharmaceuticals, St. Petersburg, FL) supplementation in their diet to replace exocrine pancreatic enzymes.

Allogeneic transplantation of composite IKs was performed by the same technique as described previously for renal transplantation (10,11). Briefly, recipients underwent bilateral nephrectomy. The IK was harvested from the donor, flushed immediately with cold perfusion solution, and then placed on ice until transplantation. The IK was transplanted orthotopically into the recipient, using end-to-side aortocaval anastomoses. Urinary drainage was accomplished through a vesicoureteral anastomosis. Two semipermanent indwelling silastic central venous catheters were placed into the external jugular veins of the recipient. One catheter was used for the administration immunosuppression and the other to obtain blood samples. Total cold ischemic time was kept to <60 min. No blood transfusions were required.

Cyclosporine A (CyA) (Sandimmune) was generously provided by Novartis Pharmaceutical (East Hanover, NJ) and was mixed and administered as an intravenous suspension, according to specifications of the manufacturer. CyA was given daily as a single infusion (10 mg · kg⁻¹ · day⁻¹) for 12 consecutive days, beginning on postoperative day (POD) zero. Whole-blood trough levels were determined by a fluorescence polarization immunoassay (Abbott Laboratories, Dallas, TX), which measured the parent compound but not metabolites, and the results were expressed in nanograms per milliliter. FK506 was generously provided by Fujisawa Healthcare (Deerfield, IL) and was administered continuously through an infusor pump (Baxter Healthcare, Deerfield, IL) for 12 consecutive days, beginning on POD zero. FK506 doses were started at 0.15 mg · kg⁻¹ · day⁻¹ and adjusted to maintain blood levels between 30 and 40 ng/ml. Whole-blood levels were determined by microparticle enzyme immunoassay (Tacrolimus II, IMX System; Abbott Laboratories).

Islet function was assessed through fasting blood glucose (FBG) levels, glucose tolerance tests, and histological examination of biopsies. Glucose tolerance testing was performed by intravenous infusion of a glucose load (0.5 g/kg body wt) after a 24-h fasting period. Blood samples were obtained at 0, 5, 15, 30, 60, and 120 min postinfusion and analyzed for glucose and insulin concentrations.

Biopsies were obtained on PODs 30, 60, and >100. Biopsies were performed through a 2-inch flank incision. Subcutaneous tissue was laterally displaced to permit both macroscopic examination and wedge biopsy. Biopsies that included islet and renal tissue were divided into sections for staining (hematoxylin and eosin [H&E]) and immunohistochemistry.

Rejection was monitored through measurement of plasma creatinine and FBG levels, as well as histological assessment of biopsy tissue. The clinical end points used in this study were either graft survival >100 days or a moribund animal with uremia or hyperglycemia. Biopsied tissues were examined by a pathologist, and rejection of kidneys was diagnosed according to a standardized grading system (12).

Tangential sections were taken to include islet and renal tissue. Formaldehyde-processed specimens were examined using H&E staining and insulin-specific antibodies. Insulin staining was performed on 4.0-μm sections cut from a paraffin-embedded block. Sections were incubated at room temperature in a humidified chamber for 20 min with 10% normal goat serum in PBS (pH 7.4) and then for 60 min with guinea pig anti-porcine insulin antibody (Dako, Carpinteria, CA) diluted 1:10 in PBS. They were next incubated with a 1:200 dilution of biotinylated goat anti-guinea pig antibody (Vector Laboratories, Burlingame, CA) for 60 min. The tissue-bound primary antibodies were detected by an avidin-biotin-peroxidase complex (Dako) that was visualized by staining with 0.02% hydrogen peroxide containing 0.3 mg/ml 3,3′-diaminobenzidine in 0.05 mol/l Tris buffer. Sections were counterstained with Gill’s single-strength hematoxylin.

Blood was drawn from the external jugular vein on PODs 30, 60, and >100. For separation of peripheral blood lymphocytes, freshly heparinized whole blood was diluted ∼1:2 with Hank’s balanced salt solution (HBSS; Gibco BRL, Grand Island, NY), and the mononuclear cells were obtained by gradient centrifugation using lymphocyte separation medium (Organon, Teknika, Durham, NC). The mononuclear cells were washed once with HBSS, and contaminating red cells were lysed with ammonium chloride potassium buffer (B&B Research Laboratory, Fiskeville, RI). Cells were then washed with HBSS and resuspended in tissue culture medium. All cell suspensions were kept at 4°C until used in flow cytometry assays.

Cell-mediated lymphocytotoxicity (CML) assays were performed as previously described using our standard CML tissue culture medium (10). Briefly, lymphocyte cultures containing 4 × 10⁶ responder and 4 × 10⁶ stimulator peripheral blood mononuclear cells (PBMCs) (irradiated with 2,500 cGy) were incubated for 6 days at 37°C in 7.5% CO₂ and 100% humidity. Bulk cultures were harvested, and effectors were tested for cytotoxic activity on ⁵¹Cr (Amersham, Arlington Heights, IL)-labeled lymphoblast targets. Effector cells were incubated for 5.5 h with target cells at E/T ratios of 100:1, 50:1, 25:1, and 12.5:1. Three target cells were tested in each assay: swine leukocyte antigen (SLA)-matched PBMCs to the effectors, donor-matched PBMCs, and third-party PBMCs. Supernatants were then harvested using the Skatron collection system (Skatron, Sterling, VA), and ⁵¹Cr release was determined on a gamma counter (Micromedics, Huntsville, AL). The results were expressed as PSL (percent specific lysis), calculated as:

Mixed lymphocyte response (MLR) assays were performed by plating 4 × 10⁵ responder PBMCs in triplicate in 96-well flat-bottom plates (Costar, Cambridge, MA). Cells were stimulated with 4 × 10⁵ stimulator PBMCs irradiated with 2,500 cGy. Cultures were incubated for 5 days at 37°C in 6% CO₂ and 100% humidity in our standard MLR culture medium (13). ³H-thymidine was added for an additional 6 h of culture. Wells were then harvested onto Mash II glass fibers and counted for β emission.

Four animals were completely pancreactectomized within 1 week before receiving a minor-mismatched composite IK. After total pancreatectomy, the FBG of the recipients markedly increased, reaching levels >280 mg/dl (Fig. 2A). No insulin therapy was given.

Donors were selected by weight to achieve the required number of IE per kilogram for the intended recipient. IKs containing ∼5,000 IE/kg recipient body wt were transplanted, with a 12-day course of CyA at 10 mg · kg⁻¹ · day⁻¹. The recipients’ native kidneys were removed on the day of IK transplantation. After IK transplantation, FBG levels fell immediately to normal levels (<120 mg/dl) by POD 1. IK grafts were permanently accepted, and recipients maintained normoglycemia for at least 90 days in three of four cases (Fig. 2A, C, and D), without the need for insulin therapy. The fourth animal had to be sacrificed because of complications from a catheter infection on POD 67; however, this animal had also maintained normoglycemia after IK transplantation (Fig. 2B).

Glucose tolerance testing revealed normal responses to intravenous glucose loads of 0.5 g at 2 months after IK transplantation (Fig. 2E, representative data shown). In contrast, glucose levels in pancreatectomized control animals reached >450 mg/dl (Fig. 2E). Upon gross examination of the IK allografts, no edema or spot hemorrhages were visible. As seen in Fig. 2F and G, a biopsy on POD 128 in a representative animal revealed numerous insulin-producing β-cells with minimal cellular infiltrate (Fig. 2F, H&E staining; Fig. 2G, insulin staining). Renal structure and function were followed throughout the transplant period. Creatinine levels remained <1.5 mg/dl throughout the clinical course (Fig. 2A–D), and biopsies revealed intact kidney parenchyma with minimal cellular infiltrate (renal parenchyma in Fig. 2F). The recipient animals survived long term (>3 months) and demonstrated a significant increase in weight during the experiment (40, 63, and 8% weight gain on the day of IK graftectomy compared with day of IK transplant in each animal). The IKs were explanted on PODs 67, 90, 128, and 150, and follow-up blood glucose levels were monitored to assure the completeness of pancreatectomies. The blood glucose levels rose immediately after IK explantation in each animal, indicating that the total pancreatectomies were complete in all cases.

Three pancreatectomized animals received direct minor-mismatched islets transplanted beneath the renal capsule (n = 1) or injected via the portal vein (90% of islets) and beneath the renal capsule (10% of islets) (n = 2). A total of 7,000 IE/kg recipient body wt were injected into each animal, followed by a 12-day course of CyA (10 mg · kg⁻¹ · day⁻¹). This increased number of islets was ∼40% more than that used in the preparation of IKs to compensate for any islet cell proliferation that might have occurred during the 2- to 3-month period after IK preparation. Recipient FBG decreased on POD 1; however, the levels subsequently increased and the islets were rejected by histological criteria by POD 23 (Figs. 3 and 4A and B) in the two animals that underwent direct islet transplantation. The third animal was maintained on exogenous insulin during postoperative weeks 2–4 to determine whether prolonged survival of the animal by insulin therapy would provide direct islet allografts additional time to engraft. This animal’s FBG was maintained between 180 and 250 mg/dl with insulin therapy; however, blood glucose levels immediately increased after cessation of insulin at week 5 (Fig. 3). Autopsy results revealed mononuclear cell infiltrates and fibrosis beneath the renal capsule (the site of islet implantation) (Fig. 4C). No functioning islets were found on histological evaluation (Fig. 4D).

Two pancreatectomized animals received direct minor-mismatched islets and renal allografts, isolated from the same donor and transplanted on the same day, with a standard 12-day course of CyA. Three days before transplantation, donor pancreata were partially removed through a 70% pancreatectomy, and islets were harvested for a 3-day culture. On the day of transplantation, renal harvest was performed on the islet donor and recipients underwent bilateral nephrectomies before renal transplantation. Immediately after revascularization of the donor renal allograft, cultured donor islets (7,000 IE/kg recipient body wt) were injected beneath the renal capsule. In both animals, FBG levels decreased on POD 1 (Fig. 5A). However, thereafter, both animals became markedly hyperglycemic. One moribund animal was killed on POD 14 (Fig. 5A), while the other animal was maintained with exogenous insulin from POD 5 until POD 29 (Fig. 5B). The target FBG level in this period was established between 180 and 250 mg/dl. Regular and/or lente insulin was titrated between 4 and 10 units/day to maintain target FBG levels. The animal required less insulin (from 10 units/day to 6 units/day) to maintain target FBG levels during week 4, and insulin therapy was discontinued on POD 29. However, FBG levels never fell below 180 mg/dl after the cessation of insulin and levels increased markedly to 600 mg/dl on POD 31 (Fig. 5B). Creatinine levels remained normal in the animal that died on POD 14 (3 days after cessation of CyA); however, creatinine levels increased slightly in the animal that demonstrated a prolonged survival on insulin therapy (Fig. 5B). Histological analysis of the autopsy specimen on POD 31 revealed few remaining islets and a diffuse mononuclear cell infiltrate in the site of islet transplantation (Fig. 5C). Furthermore, an interstitial mononuclear cellular infiltrate was seen in the renal parenchyma (Fig. 5D), a finding not noted in our previous experience with minor-mismatched renal transplantation alone in miniature swine.

Two pancreatectomized animals received fully-mismatched IKs containing ∼5,000 IE/kg recipient body wt, with a 12-day course of FK506, which has been shown to facilitate the induction of tolerance to fully mismatched renal allografts in our miniature swine (13). Native kidneys were removed on the day of IK transplantation. In each case, FBG levels rose to >280 mg/dl after pancreatectomy. The composite IKs were accepted, and normoglycemia was immediately restored. However, both animals exhibited a transient period of hyperglycemia that lasted, approximately, from POD 5 to POD 18. These episodes resolved spontaneously within a week after the cessation of FK506, suggesting that the transient rise in blood glucose may have been due to FK506-associated islet toxicity (14,15). The clinical, pathological, and histological outcome in these animals was otherwise similar to that seen in minor-mismatched IK transplantation with a 12-day course of CyA. All IKs were permanently accepted, and recipients maintained normoglycemia for at least 120 days without insulin therapy (Fig. 6A and B). Glucose tolerance testing of IK recipients was normal 2 months after IK transplantation (Fig. 6C). Gross examination of the grafts revealed no edema, spot hemorrhages, or vasculopathy. Biopsy results revealed numerous insulin-producing β-cells with a minimal cellular infiltrate (Fig. 6D). Additional immunosuppression or insulin therapy required was not required in either case, and creatinine levels were maintained below 1.5 mg/dl throughout the experiment (Fig. 6A). The blood glucose levels rose immediately after IK explantation in each animal.

IK recipients developed in vitro evidence of donor-specific tolerance within 30 days after IK transplantation. Figure 7A demonstrates the CML responses on POD 30 in both animals. Strong anti-third party reactivity, but no anti-donor reactivity, was observed. MLR analysis also revealed anti-donor specific unresponsiveness (Fig. 7B), which was maintained for the duration of the experiment, confirming that these animals were tolerant.

We report here that vascularized islet cells, when transplanted as part of a composite IK, cured surgically induced diabetes across minor- and fully MHC-mismatched barriers. The present study demonstrates that 1) nonvascularized islet transplants isolated from a single donor are not capable of curing surgically induced diabetes and that 2) a similar number of prevascularized islets, within composite IK grafts, can cure surgically induced diabetes, as well as induce tolerance, after a short course of immunosuppression. Recipient animals that were hyperglycemic after complete pancreatectomy maintained normoglycemia indefinitely after IK transplantation. Upon removal of the IK, these animals returned to their pretransplant, hyperglycemic state, indicating that the pancreatectomies were indeed complete and that the transplanted vascularized islets within the composite IK were responsible for the normal regulation of blood glucose levels. An additional important note is the fact that each IK contained 5,000 IE/recipient body wt on the day of the IK transplantation. These recipients gained body weight, which effectively diluted the ratio of IE per kilogram recipient body weight after IK transplantation. Therefore, by the conclusion of the experiment, the recipients maintained normoglycemia with ∼3 × 10³ IE/kg recipient body wt.

Our previous studies of allogeneic and xenogeneic thymic tissue transplantation have demonstrated the role of prevascularization on graft survival (18,19), a principle that has already been noted for transplants of kidney tissues in a rodent model (28). The present studies provide further evidence of the critical nature of a graft’s vascular supply. Nonvascularized islets neither engrafted nor controlled blood glucose levels in transplanted recipients. While nonvascularized islet transplants transiently decreased FBG levels during the first 2 days after transplantation, this decrease was most likely secondary to the release of insulin from damaged ischemic islet β-cells. Histological and physiological data indicated that such grafts were rapidly rejected within the first weeks after transplantation. Vascularized grafts appeared less susceptible to this nonspecific graft loss and did not sensitize the host to donor antigen—processes that are likely to have limited the success of nonvascularized islet grafts. In fact, it has been shown that transplants of composite kidney-islet grafts were more successful in avoiding autoimmune destruction in BB rats than freely grafted islets (30).

While previous studies have reported increased islet cell survival when injected via the portal vein (2,20), our studies in miniature swine were unable to confirm this finding.

This study also demonstrated that the success of the IK graft is not due to antigen load or specific help by a donor kidney, as cotransplantation of a donor kidney did not facilitate engraftment of nonvascularized islets. The improved survival of composite IK grafts appears to be based almost entirely on the prevascularization of the islet tissue. Creation of autologous IKs allowed for ischemia, reperfusion, and revascularization to all occur in a nonallogeneic setting, whereas directly transplanted islets must encounter all of these barriers to engraftment in an allogeneic setting. Autologous IK creation permitted not only enhanced islet engraftment by avoidance of typical rejection responses observed during this period, but also facilitated the long-term induction of tolerance by minimizing the presentation of allogeneic antigens after transplantation. When renal transplantation was immediately followed by the subcapsular implantation of cultured islets isolated from the same donor, the composite grafts were unable to cure diabetes in pancreatectomized recipients. Furthermore, the renal parenchyma in these grafts demonstrated an increased infiltrate as compared with historical renal allograft controls performed across similar immunological barriers (21). This indicated that 1) the process of renal transplantation itself does not lead to enhanced islet survival, 2) nonvascularized islets may be more immunogenic than a kidney transplanted alone, and 3) the presence of nonvascularized islets may actually sensitize the host to donor antigen.

The present studies also indicated that immunosuppressive regimens that induced tolerance to renal allografts also induced tolerance to associated vascularized islets within the IK graft. However, such regimens are not capable of inducing tolerance to islets directly transplanted into recipients. Our previous studies have demonstrated that a 12-day course of CyA (10) or FK506 induces tolerance to partial MHC- and fully MHC-mismatched renal allografts, respectively (13). These same regimens were capable of inducing tolerance to vascularized islets across similar barriers in the present study. Although short-term high-dose FK506 has not been used for tolerance induction clinically, we have used it successfully for this purpose recently in baboons (P.A.V., R.N.B., S.Y., J.M.L., J.C.L., C.K., D.H.S., K.Y., manuscript in preparation), and similar doses have been used clinically for the treatment of ABO-incompatible renal transplantation (16) and GvHD (17).

We have suggested that the transient rise in blood glucose in our fully MHC-mismatched recipients may have been due to FK506-associated islet toxicity (14,15). This level has never induced diabetes in our miniature swine recipients of kidneys (13), and FK506 blood levels were controlled between 25 and 35 ng/ml in both cases. Thus, transplanted islets could be sensitive to this drug, especially in the induction period after transplantation. Even though the regimen of FK506 needed to induce tolerance might have a slightly diabetogenic effect, the benefit of using a short course of immunosuppression for the establishment of long-term tolerance is very attractive. In fact, the hyperglycemic period, which was insulin free, was only transient and easily reversed upon the discontinuation of FK506. In the future, strategies that are currently demonstrating success in preclinical and early clinical models of renal allograft transplantation, such as mixed chimerism (22,29), costimulatory blockade (23), and FN18 immunotoxin (24,25), may be used to induce tolerance to vascularized islet allografts and avoid any secondary islet toxicity.

Although the feasibility of islet transplantation has previously been demonstrated (1,2), to our knowledge, the successful transplantation of islets isolated from a single living donor has not achieved the success reported from the use of multiple islet donors. The IK strategy permitted the isolation of a sufficient number of islets, while leaving the donor normoglycemic and with intact exocrine function. Such a strategy may eventually permit successful living-related islet donation for diabetic patients with end-stage diabetic nephropathy. We can envision the laporoscopic preparation of IK grafts (26) several months in advance of kidney transplantation.

While we have chosen to initiate our studies in an allogeneic model to assess cellular responses without attendant complications from natural antibodies, this model may be even more applicable for IK transplants across xenogeneic barriers. In xenotransplantation, the preparation of IKs could provide an unlimited source of organs for transplantation into diabetic recipients requiring kidneys.

This work was supported by the following grants/grantors: NIH R01 AI 31046, the Juvenile Diabetes Foundation, and Fujisawa Health Care.

The authors would like to thank Drs. Paul S. Russell, James S. Allan, and Douglas R. Johnston for their helpful review of the manuscript; Scott Arn for herd management and quality control typing; and Lisa A. Bernardo for help in manuscript preparation. The authors would also like to thank Fujisawa Healthcare (Deerfield, IL) for generously providing FK506, Novartis Pharmaceutical (East Hanover, NJ) for providing CyA, and Schering-Plough Animal Health for providing Flunixamine.