Single-Donor, Marginal-Dose Islet Transplantation in Patients With Type 1 Diabetes

Context Islet allografts from 2 to 4 donors can reverse type 1 diabetes. However, for islet transplants to become a widespread clinical reality, diabetes reversal must be achieved with a single donor to reduce risks and costs and increase the availability of transplantation.

Objective To assess the safety of a single-donor, marginal-dose islet transplant protocol using potent induction immunotherapy and less diabetogenic maintenance immunosuppression in recipients with type 1 diabetes. A secondary objective was to assess the proportion of islet transplant recipients who achieve insulin independence in the first year after single-donor islet transplantation.

Design, Setting, and Participants Prospective, 1-year follow-up trial conducted July 2001 to August 2003 at a single US center and enrolling 8 women with type 1 diabetes accompanied by recurrent hypoglycemia unawareness or advanced secondary complications.

Interventions Study participants underwent a primary islet allotransplant with 7271 (SD, 1035) islet equivalents/kg prepared from a single cadaver donor pancreas. Induction immunosuppression was with antithymocyte globulin, daclizumab, and etanercept. Maintenance immunosuppression consisted of mycophenolate mofetil, sirolimus, and no or low-dose tacrolimus.

Main Outcome Measures Safety (assessed by monitoring the severity and duration of adverse events) and efficacy (assessed by studying the recipients’ insulin requirements, C-peptide levels, oral and intravenous glucose tolerance results, intravenous arginine stimulation responses, glycosylated hemoglobin levels, and hypoglycemic episodes) associated with the study transplant protocol.

Results There were no serious, unexpected, or procedure- or immunosuppression-related adverse events. All 8 recipients achieved insulin independence and freedom from hypoglycemia. Five remained insulin-independent for longer than 1 year. Graft failure in 3 recipients was preceded by subtherapeutic sirolimus exposure in the absence of measurable tacrolimus trough levels.

Conclusions The tested transplant protocol restored insulin independence and protected against hypoglycemia after single-donor, marginal-dose islet transplantation in 8 of 8 recipients. These results may be related to improved islet engraftment secondary to peritransplant administration of antithymocyte globulin and etanercept. These findings may have implications for the ongoing transition of islet transplantation from clinical investigation to routine clinical care.

Type 1 diabetes remains a therapeutic challenge. The success rate of islet transplants has recently been increased markedly by transplanting a higher mean islet mass (11 547 [SD, 1604] islet equivalents [IEs]/kg) prepared from 2 to 4 donor pancreases and using glucocorticoid-free immunosuppression.^1-6 However, for islet transplants to become a widespread clinical reality, additional advances are still needed. In particular, restoration of insulin independence must be achieved with a single donor, as is also the case with pancreas transplants, to reduce the risks and costs and increase the availability of islet transplantation. We designed a protocol to limit ischemic injury of islets during pancreas storage, permit initiation of potent immunotherapy prior to transplantation, and minimize calcineurin inhibitor exposure.

Our study was a prospective, single-center, 1-year follow-up pilot trial conducted from July 2001 to August 2003. The primary efficacy end point was the proportion of recipients who achieve insulin independence in the first year after a single-donor islet transplant. We defined recipients as insulin-independent if they maintained fasting blood glucose levels below 126 mg/dL (7.0 mmol/L) and 2-hour postprandial levels below 180 mg/dL (10.0 mmol/L) after discontinuation of insulin.

A total of 8 patients (coincidentally all women) were enrolled. Eligibility criteria are detailed in the Box and detailed recipient characteristics are shown in Table 1. Our study protocol was approved by the local institutional review board, and written informed consent was obtained from all participants.

Inclusion Criteria

1. Age 18 years or older

2. C-peptide–negative* type 1 diabetes for >5 years complicated by 1 of the following:
   
   Advanced secondary complications including proliferative retinopathy or clinically significant macular edema or photocoagulation, urinary albumin excretion >300 mg/d but proteinuria <3 g/d, or symptomatic autonomic neuropathy
   
   Metabolic lability/instability (≥2 episodes of severe hypoglycemia or ≥2 hospital admissions for ketoacidosis in the past year)
   
   Hypoglycemia unawareness (≥4 “reduced” responses in the Clark “hypoglycemia questionnaire”⁷)

Exclusion Criteria

1. Body weight >70 kg

2. Insulin requirements >40 U/d

3. Previous islet transplant

4. Abnormal renal function (creatinine clearance <60 mL/min [1.002 mL/s] per 1.73 m²)

5. Portal hypertension, abnormal liver enzyme test results, or history of significant liver disease

*Defined as C-peptide level <0.2 ng/mL after administration of 5 g of intravenous arginine.

Eighteen consecutive donor pancreases were procured from cadaver donors younger than 50 years with a body mass index (calculated as weight in kilograms divided by square of height in meters) of 27 or greater; the pancreases were preserved for 8 hours or less using the 2-layer method.⁸ ABO compatibility and a negative serum crossmatch for T cells were required, but HLA antigen matching was not. Islets were isolated as previously described.⁹ Briefly, preserved pancreases were perfused with cold Liberase (Roche Diagnostics Corp, Indianapolis, Ind).¹⁰ Islets were isolated by the automated method,¹¹ purified with continuous iodixanol density gradients in a Cobe 2991 cell separator (Gambro BCT, Lakewood, Colo), and cultured free-floating in supplemented CMRL 1066 for 1 day at 37°C and 1 day at 22°C.^9,12,13 Of the 18 consecutive cadaver donor pancreases processed for this study, preparations from 8 were transplanted (detailed donor and graft characteristics for those 8 are shown in Table 2). Islet preparations from the remaining 10 donor pancreases were not transplanted because of inadequate islet yield for single-donor islet transplantation. The mean islet yield of 7 of those 10 preparations was 301 428 (SD, 59 780) IEs.

After establishing access to the portal vein via minilaparotomy or percutaneous transhepatic portal venous catheterization, we infused 7271 (SD, 1035) IEs/kg of recipient body weight by gravity, along with heparin, 70 U/kg, on day 0 into 8 consecutive participants. Prophylactic anticoagulation was continued with intravenous heparin (target partial thromboplastin time, 50-60 seconds) for 48 hours, followed by enoxaparin (30 mg subcutaneously twice daily) through day 7.

Induction immunosuppression, initiated on day –2, consisted of rabbit antithymocyte globulin (RATG) (0.5 mg/kg of recipient body weight [day –2], 1.0 mg/kg [day –1], 1.5 mg/kg [days 0 through +2]),¹⁴ methylprednisolone (on day –2 only, 2 mg/kg), daclizumab (5 doses of 1 mg/kg every 2 weeks starting on day 0), and etanercept (50 mg intravenously 1 hour pretransplantation, followed by 25 mg subcutaneously on days 3, 7, and 10). Premedication for RATG included acetaminophen and diphenhydramine as well as pentoxifylline, which was extended through day 7 posttransplantation. Maintenance immunosuppression was initiated with sirolimus (0.2 mg/kg starting on day –2, followed by 0.1 mg/kg daily; target whole blood trough levels, 5-15 ng/mL, as tolerated) and reduced-dose tacrolimus (0.015 mg/kg twice daily, starting on day 1; target whole blood trough levels, 3-6 ng/mL). At 1 month posttransplantation, tacrolimus was gradually replaced with mycophenolate mofetil (750-1000 mg, twice daily); tacrolimus was either discontinued or dosed to a target trough level of less than 3 ng/mL. If target levels of sirolimus could not be achieved or maintained, however, tacrolimus (target level, 3-6 ng/mL) was continued.

Antimicrobial and antiviral prophylaxis with imipenem, vancomycin, trimethoprim/sulfamethoxazole, and valganciclovir was administered. Glycemic control was achieved with intravenous insulin from day –2 to day +2 relative to transplant and with subcutaneous insulin for at least 3 additional weeks.

Safety of the transplant protocol was assessed by monitoring the severity and duration of procedural complications, serious infections, or islet- or immunosuppression-related adverse events.

Self-measured blood glucose concentrations (5 or more times daily), hypoglycemic episodes, basal C-peptide levels, and levels of glycosylated hemoglobin (by high-performance liquid chromatography with reference range 4.3%-6.0%; monthly through day 90 posttransplantation, then quarterly) were recorded throughout the study. Recipients also underwent oral glucose tolerance testing (75 g, 2 hours), intravenous glucose tolerance testing, and intravenous arginine stimulation. The acute insulin and C-peptide responses to intravenous arginine (AIR-arginine, ACR-arginine) and to intravenous glucose (AIR-glucose, ACR-glucose) were defined as the mean of the 3 highest poststimulus values between 2 and 5 minutes after the start of the stimulus administration minus the mean of 2 prestimulus values.¹⁵ Human C-peptide levels were measured by double antibody radioimmunoassay (Diagnostic Products Corp, Los Angeles, Calif) (interassay coefficient of variation [CV], 7.9% at a mean C-peptide level of 0.56 ng/mL and 9.1% at a mean of 1.6 ng/mL; intra-assay CV, 9.9% at a mean of 0.08 ng/mL and 2.1% at a mean of 0.73 ng/mL). Human insulin levels were measured by chemiluminescent immunoassay (Immulite; Diagnostic Products Corp) (interassay CV, 6.4% at a mean insulin level of 7.5 mU/L and 7.6% at a mean of 47.7 mU/L; intra-assay CV, 3.3% at a mean of 14.3 mU/L and 6.4% at a mean of 46.5 mU/L). Results were compared with those obtained in control individuals without diabetes, matched for age and body mass index. Islet graft loss was defined by the absence of basal and arginine-stimulated C-peptide levels.

Anti-GAD65 antibody, anti-ICA512 antibody, and anti-insulin antibody titers were measured with radiobinding assays.¹⁶

Data are presented as mean (SD) unless otherwise stated. Comparisons were performed using the 2-tailed t test. Analyses were performed using SAS version 9.1 (SAS Institute Inc, Cary, NC); P<.05 was used to determine statistical significance.

We did not observe procedural complications; serious infections; or serious, unexpected, and islet- or immunosuppression-related adverse events. Expected adverse events included lymphopenia and transient neutropenia requiring short-term (<1 week) administration of granulocyte colony-stimulating factor in 5 recipients. Intermittent oral aphthous ulcers were observed in all participants. No clinically significant changes in creatinine clearance or urinary albumin excretion were observed.

All 8 recipients became insulin-independent, with glycosylated hemoglobin levels within reference range and freedom from hypoglycemia. Of the 8 recipients, 5 have remained insulin-independent for longer than 1 year and 3 were insulin-independent for 121, 76, and 7 days (Table 3). Metabolic tests performed 180 or more days posttransplantation in the 5 recipients with sustained insulin independence showed a mean AIR-glucose level of 16.7 (SD, 5.5) μU/mL (30% [SD, 10%] of controls), an ACR-glucose of 1.23 (SD, 0.46) ng/mL (40% [SD, 15%] of controls), an AIR-arginine of 15.5 (SD, 3.7) μU/mL (53% [SD, 13%] of controls), and an ACR-arginine of 1.07 (SD, 0.15) ng/mL (59 [SD, 8%] of controls). For 4 of these 5 recipients, results of oral glucose tolerance testing at 180 or more days posttransplantation revealed 2-hour plasma glucose levels below 140 mg/dL (7.8 mmol/L). The 2-hour plasma glucose level in the fifth recipient was 208 mg/dL (11.5 mmol/L). These 5 recipients received daily mycophenolate mofetil doses of 1.5 to 2.0 g. They either achieved and maintained sirolimus trough levels greater than 9 ng/mL, with tacrolimus trough levels of 0 to less than 3 ng/mL, or achieved tacrolimus trough levels of 3 to 6 ng/mL in the absence of target sirolimus trough levels. The 3 recipients who resumed exogenous insulin therapy had received 1.5 g/d or more of mycophenolate mofetil but had subtherapeutic sirolimus trough levels (<9 ng/mL) in the absence of measurable tacrolimus trough levels (<3 ng/mL). For additional information, see Table 3 for metabolic monitoring and Table 4 for exposure to immunosuppressive drugs.

Of the 3 participants with graft failure, 2 tested positive for anti-GAD65 and anti-ICA512 in the pretransplantation period. In contrast, none of the 5 who remained insulin-independent tested positive for anti-GAD65 and anti-ICA512.

Graft failure was followed by allosensitization in 2 recipients.

Our results mark a distinct advance in islet transplant efficacy. We not only achieved insulin independence using islets from only 1 donor pancreas (as compared with 2 to 4 in the Edmonton trial¹), we also achieved superior glycemic control (as evidenced by normal results of oral glucose tolerance testing in 4 of 5 recipients with sustained insulin independence) using significantly fewer islets (7271 [SD, 1035] IEs/kg vs 11 547 [SD, 1604] IEs/kg; P<.001). We had previously achieved insulin independence in 4 of 6 participants with type 1 diabetes who received an islet mass of 10 302 (SD, 2594) IEs/kg from 1 donor pancreas.⁹ However, transplantation of such an islet mass is only available from a limited number of donor pancreases and obscures assessment of the ability of a given protocol to permit reversal of diabetes with a lower islet mass retrievable from a larger subgroup of donor pancreases.

Determining the reasons for our high success rates with a lower islet mass from a single donor pancreas will have important ramifications for the advancement of the field. Previous studies by us and others have suggested that excluding pancreases from donors older than 50 years, limiting cold storage to less than 8 hours and using the 2-layer preservation method, avoiding use of Ficoll during islet purification, and culturing islets pretransplantation could conceivably preserve the potency of transplanted islets.^13,17-22 Since pancreas procurement, preservation, islet processing, and culture protocols in the 2 studies were all identical, we assume that the potency was the same and therefore interpret the high efficacy of single-donor, marginal-dose islet transplants in our current trial as preliminary evidence of improved engraftment. In this study, induction therapy was with RATG, combined with daclizumab and etanercept. The resistance of islet-directed autoimmune responses to conventional immunosuppressive drugs^23-27 and the immediate exposure of intraportally transplanted islets to primed autoreactive, islet beta cell–directed T cells have also provided a strong rationale for pretransplant initiation of RATG, which is known to cause selective depletion of activated T cells and dose-dependent depletion of resting T cells.²⁸ Many of these effects are shared with the anti-CD3 monoclonal antibody, hOKT3γ1 (Ala-Ala), used in our previous trial.⁹ Thus, they may not sufficiently explain the ability of the protocol used in our current trial to facilitate reversal of diabetes after single-donor, marginal-dose islet transplants. Therefore, the results are possibly related to the peritransplant administration of the soluble tumor necrosis factor receptor etanercept. Tumor necrosis factor α is cytotoxic to human islet beta cells.²⁹ In murine models, selective inhibition of tumor necrosis factor α in the peritransplant period has promoted reversal of diabetes after marginal-mass islet transplants.³⁰ Etanercept administration is a new addition to our protocol and distinguishes this trial from our previous trial,⁹ so it could have been a major factor allowing consistent diabetes reversal with a low islet dose.

Moreover, replacing or minimizing tacrolimus at 1 month posttransplantation, as we did in our current trial, may have enhanced the function of engrafted islets. Low-dose calcineurin inhibitor therapy, with no or minimal doses of steroids, has previously been associated with significantly reduced insulin sensitivity and beta cell secretory reserve,³¹ suggesting that even low-dose tacrolimus therapy may limit the ability of a reduced islet mass to reverse diabetes. Maintenance immunosuppression with mycophenolate mofetil and sirolimus, shown to be synergistic in experimental studies,³² is without diabetogenic or nephrotoxic adverse effects and is sufficiently potent, provided induction immunosuppression is administered and adequate sirolimus levels are achieved and maintained. In light of the results on exposure to immunosuppressive drugs (Table 4), it seems likely that the islet graft failure experienced by 3 recipients was caused by alloimmunity and/or recurrent autoimmunity. The observation that 2 of 3 recipients with graft failure, but none of the 5 who remained insulin-independent, tested positive for both anti-GAD65 and anti-ICA512 in the pretransplant period suggests a possible involvement of autoimmunity in graft failure. More detailed studies in a larger series of recipients will be needed to accurately ascribe islet graft loss to metabolic or immunologic reasons.

In conclusion, potent induction immunotherapy as used in this study may increase the ability of low-dose islet allografts to reverse diabetes and may minimize nephrotoxicity and cardiovascular toxicity by sparing calcineurin inhibitor dosing. While these findings may suggest a distinct advance in islet transplantation, further study in a larger population with longer follow-up will be critical to assess the risk-benefit ratio of this emerging therapeutic option.

Corresponding Author: Bernhard J. Hering, MD, Department of Surgery, University of Minnesota, Mayo Mail Code 195, 420 Delaware St SE, Minneapolis, MN 55455 (bhering@umn.edu).

Author Contributions: Dr Hering had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analyses.

Study concept and design: Hering, Kandaswamy, Eckman, Matsumoto, Hunter, Sutherland.

Acquisition of data: Hering, Kandaswamy, Ansite, Nakano, Sawada, Matsumoto, Ihm, Zhang, Parkey, Hunter.

Analysis and interpretation of data: Hering, Eckman, Sutherland.

Drafting of the manuscript: Hering, Ansite, Eckman, Nakano, Matsumoto, Ihm, Zhang.

Critical revision of the manuscript for important intellectual content: Hering, Kandaswamy, Eckman, Sawada, Parkey, Hunter, Sutherland.

Statistical analysis: Sawada, Ihm.

Obtained funding: Hering.

Administrative, technical, or material support: Hering, Kandaswamy, Ansite, Eckman, Nakano, Matsumoto, Zhang.

Study supervision: Hering, Sutherland.

All authors contributed to the preparation of the report.

Funding/Support: This study was supported by grants from Roche Laboratories Inc (RO49272), the National Center for Research Resources, National Institutes of Health (MO1-RR00400 and U42 RR016598-01), and the Juvenile Diabetes Research Foundation (JDRF #4-1999-841). SangStat provided rabbit anti-thymocyte globulin. Wyeth-Ayerst supplied sirolimus.

Role of the Sponsors: The sponsors had no role in the design or conduct of the study; the collection, management, analysis, and interpretation of the data; or the preparation, review, or approval of the manuscript.

Acknowledgment: We are indebted to Kathy Duderstadt, Kathy Hodges, and Carrie Gibson for their invaluable contributions as study nurse coordinators; to the staff members of the General Clinical Research Center for excellent patient care; to Jeremy Oberbroeckling for his technical expertise in the islet isolation laboratory; and to Robin Jevne, PhD, and Dylan Zylla for data analysis and presentation. We appreciate the critical manuscript review by Klearchos Papas, PhD. We are grateful to Richard Bergenstal, MD, David Kendall, MD, and Lisa Fish, MD, for verifying participant eligibility; to George Eisenbarth, MD, PhD, for measuring autoantibodies; and to Edmond Ryan, MD, for assisting with metabolic studies. We thank LifeSource, other organ procurement organizations, and Julianne Zabloski for their efforts in pancreas procurement and Deborah Butterfield of Diabetes Portal for assistance in participant recruitment. We thank Mary Knatterud, PhD, for editing the manuscript.