In preparation for assessment of percutaneous biopsies in our clinical pancreas transplant program, a working knowledge of the histopathologic changes after transplantation was obtained in a longitudinal open biopsy study of 16 dogs receiving bladder-drained whole pancreas allografts. Edema, extravasation of polymorphs, and lymphocytes associated with focal parenchymal injury were early, invariable, and probably nonspecific findings. The initial feature of unmodified rejection was the appearance of capillary and small vein endothelial changes with mainly perivascular inflammatory cell infiltration. Acinar cell loss occurred early and was progressive, whereas islets and ducts were relatively preserved, indicating that acinar tissue may be more vulnerable to lytic necrosis when damaged. Functional rejection, determined by fasting urinary amylase levels, was at a stage of extensive and irreversible necrosis. Functioning grafts in immunosuppressed dogs had minor and transient endothelial changes with absence of class II antigen staining of parenchymal cells.

The effect of total-lymphoid irradiation on survival of canine pancreas and kidney allografts was studied. TLI had a marked immunosuppressive effect as measured by in vitro immune responses and reduced circulating leukocytes. Despite the changes, median graft survival times for animals treated with 800 cGy (9 days) or 1800 cGy (9.5 days) were not significantly different from untreated control animals (7 days). The addition of low-dose antithymocyte globulin (10 mg/kg/day) on post-transplant days 0, 2, 4, 6, 8, and 10 had no measurable synergistic effect. Similarly, median segmental pancreas allograft survival times after 1700-2200 cGy of TLI treatment (16.5 days) were only marginally longer than those of untreated controls (9 days). The only animal to maintain a graft for greater than 200 days was matched to the donor in mixed lymphocyte culture (MLC). This animal was able to reject a third-party skin graft after 8 days while a graft from the original donor was still surviving after 21 days when the pancreas graft failed from a chronic-type rejection. These results indicate that TLI alone or in combination with ATG will not be predictably effective as a method of prolonging allograft survival. The role of matching major histocompatibility complex antigens in TLI treatment requires clarification.

A canine model of whole pancreas transplantation with pancreaticocystostomy was studied for reproducibility and long-term graft function with oral cyclosporine. The feasibility of the operative technique was established in three dogs and the graft histology at 5 days was studied. Seven pancreatectomized dogs were transplanted without immunosuppression; acute rejection was evident at a median of 10 days (range: 7-12 days). Another 14 non-pancreatectomized dogs were given oral cyclosporine (25 mg/kg per day) resulting in prolonged graft survival (P less than 0.01) with a median (actuarial) survival of 91 days (range: 8-159 days); five dogs had vascular thrombosis or graft rejection and eight dogs died with functioning grafts. Early technical loss in two dogs (8.3%) was due to arterial thrombosis. It is concluded that the model of whole pancreas transplantation was reproducible in dogs, that long-term graft function can be achieved on oral cyclosporine, and that duct patency can be maintained. Graft infarction, either primary or due to rejection, continues to complicate this model of pancreas transplantation.

Despite improved pancreas transplant graft survival, early diagnosis of rejection remains a clinical challenge. Using a canine model of whole pancreas transplantation with bladder exocrine drainage we have shown fine needle aspiration biopsy (FNAB) to provide an earlier diagnosis of rejection than fall in urinary amylase (UA). The aim of this study was to assess the ability of anti-rejection therapy to reverse rejection when the diagnosis was based on either FNAB findings or a fall of UA. Sixteen dogs received a total pancreas allograft with an inadequate dose of oral cyclosporine (5 mg/kg/day). Fasting UA levels were measured daily and percutaneous FNAB with ultrasound guidance was performed three times weekly on all dogs. The diagnosis of rejection was made in alternate dogs with either a fall of UA to a level of less than 5000 IU/liter (median 7 days) or when the total corrected increment (TCI) of aspirated infiltrating cells was greater than 2.6 (median 5 days). Anti-rejection therapy consisted of methylprednisolone 10 mg/kg/day iv for 5 days and an increase of oral cyclosporine dosage to 25 mg/kg/day. Early vascular thrombosis (Day 2) occurred in three allografts. Diagnosis of rejection based on a low level of UA permitted the successful reversal of rejection in only one of six grafts, whereas five of seven grafts were successfully treated when rejection diagnosis was based on FNAB. Median allograft survivals were 9 days (range 8-19) and 32 days (range 11-63), respectively (P less than 0.01). The earlier diagnosis of allograft rejection made by FNAB improved the ability of conventional anti-rejection therapy to reverse pancreas allograft rejection and significantly improved allograft survival.

A canine model of whole pancreas transplantation with pancreaticocystostomy was studied for predictability of graft rejection using urinary amylase (UA) monitoring. Six pancreatectomized dogs were transplanted without immunosuppression and with acute rejection occurring at a median of 9.5 days (range 7–12 days). A differential loss of allograft exocrine and endocrine function was demonstrated. with a gradual decrease in UA after transplantation but maintenance of fasting blood glucose levels (FBGL) till the day before complete loss of graft structure. Another 13 dogs treated with cyclosporin (25 mg/kg per day) had prolonged graft survival ( P < 0.01) with an actuarial median survival of 91 days (range 8–159 days). Five allografts were lost because of rejection and eight dogs died with functioning grafts. Fasting spot levels of UA < 5000 iu/l or < 10000 iu/l had a positive predictivity of graft failure of 71% or 31%, respectively. Falls of UA levels of greater than 75% in 24 h and 48 h were seen equally in both rejecting and functioning allografts. This study confirmed the role of UA as an earlier marker of rejection than FBGL. The clinical role of UA will be important, but its use as a predictor of pancreas rejection may be dependent on a fall to a predetermined level rather than the rate of fall.

In clinical pancreas transplantation, duct-occluded segmental allografts are often used. There is concern that fibrosis following duct-occlusion may lead to progressive graft failure. In this study, sequential histology and endocrine function in long-term (up to 5 years) canine autografts were assessed. Segmental pancreatic autografts with residual pancreatectomy were performed, and the pancreatic duct was occluded with cyanoacrylate glue. Serial i.v. glucose tolerance tests (IVGTT) and percutaneous needle-core biopsies of the grafts were performed as long as grafts functioned. Ten dogs were long-term (greater than 18 months) survivors: 8 dogs had functioning grafts for a median of 48 months (range 18-60) after transplantation, and 3 dogs had graft failure at 21, 27, and 60 months. The mean 40-min blood glucose concentration (BGL-40') after i.v. glucose injection did not increase with time up to 5 years after grafting. Graft biopsies showed a universal picture of aggregated islet cells and fibrous replacement of acinar tissue. The total amount of fibrosis did not change with time, but the existing fibrosis became less cellular and more dense. This long-term study showed that in autografted animals, adequate endocrine function was maintained in the majority of cases, and progressive replacement of islet tissue by fibrosis could not be demonstrated in serial biopsies taken between 18 months and 5 years after autotransplantation. We therefore conclude, that while duct-occlusion results in extensive fibrosis, the process is not progressive, and although fibrosis may contribute to late graft failure this is not inevitable.