Pancreatic Islet Transplantation in Humans: Recent Progress and Future Directions

Abstract

Pancreatic islet transplantation has become an established approach to β-cell replacement therapy for the treatment of insulin-deficient diabetes. Recent progress in techniques for islet isolation, islet culture, and peritransplant management of the islet transplant recipient has resulted in substantial improvements in metabolic and safety outcomes for patients. For patients requiring total or subtotal pancreatectomy for benign disease of the pancreas, isolation of islets from the diseased pancreas with intrahepatic transplantation of autologous islets can prevent or ameliorate postsurgical diabetes, and for patients previously experiencing painful recurrent acute or chronic pancreatitis, quality of life is substantially improved. For patients with type 1 diabetes or insulin-deficient forms of pancreatogenic (type 3c) diabetes, isolation of islets from a deceased donor pancreas with intrahepatic transplantation of allogeneic islets can ameliorate problematic hypoglycemia, stabilize glycemic lability, and maintain on-target glycemic control, consequently with improved quality of life, and often without the requirement for insulin therapy. Because the metabolic benefits are dependent on the numbers of islets transplanted that survive engraftment, recipients of autoislets are limited to receive the number of islets isolated from their own pancreas, whereas recipients of alloislets may receive islets isolated from more than one donor pancreas. The development of alternative sources of islet cells for transplantation, whether from autologous, allogeneic, or xenogeneic tissues, is an active area of investigation that promises to expand access and indications for islet transplantation in the future treatment of diabetes.

Essential Points

• Islet autotransplantation following total pancreatectomy involves complete resection of the pancreas, most commonly in the setting of recurrent acute or chronic pancreatitis, with islets isolated from the diseased pancreas returned to the patient via an intraportal infusion for intrahepatic engraftment

• Islet allotransplantation involves procurement of a deceased donor pancreas, most commonly for a recipient with type 1 diabetes, with islets isolated from the donated pancreas returned to the patient via an intraportal infusion for intrahepatic engraftment

• Care of the islet recipient requires attention to reduce inflammation and minimize thrombotic and bleeding risks while maintaining normoglycemia that is critical for islet survival until revascularization and engraftment are complete

• Autoislet recipients benefit from amelioration of pain, improved quality of life, on-target glycemic control, and in about a third of patients insulin independence; some patients may experience postprandial hypoglycemia

• Alloislet recipients benefit from amelioration of hypoglycemia, on-target glycemic control, improved quality of life, and in about half of patients insulin independence; islets from more than one donor pancreas are often required to achieve or maintain insulin independence

• Metabolic outcomes for both autoislet and alloislet recipients are dependent on the numbers of islets transplanted and the functional islet β-cell mass surviving engraftment

• Novel approaches are under evaluation to further reduce the detrimental consequences of peritransplant inflammation on intrahepatic islet engraftment efficiency and, for alloislets, calcineurin inhibitor–free immunosuppression that may have less potential for adverse effects on kidney function

Islet transplantation has been a long sought-after solution for reestablishing glucose homeostasis in diabetes caused by islet β-cell loss via the autoimmune attack of type 1 diabetes or from surgical resection or pancreatic fibrosis in pancreatogenic (type 3c) forms of diabetes. Initial work in rat models of diabetes published in the early 1970s largely from the laboratories of Paul Lacy at Washington University (1) and Clyde Barker at the University of Pennsylvania (2) established the efficacy of islet transplantation in normalizing glycemia in both syngeneic and allogeneic settings, and such work pointed toward the liver as the most efficient site of engraftment following portal vein infusion (3). Following validation of intrahepatic autologous islet transplantation for the prevention of surgical diabetes following pancreatectomy in dogs (4), John Najarian and David Sutherland at the University of Minnesota published in 1980 the first patient series of total pancreatectomy with islet autotransplantation for the treatment of chronic pancreatitis, demonstrating in some cases prevention or amelioration of postpancreatectomy diabetes (5). This proof of principle for the use of the liver for transplantation of isolated autoislets established the possibility for use of alloislets for treatment of humans with type 1 diabetes.

Efforts to enhance the purity and viability of isolated islets for transplantation by a number of groups (6, 7) led to the publication in 1988 of an automated method for isolation of human pancreatic islets by Camillo Ricordi (8) that was applied by David Scharp and Paul Lacy to the first case of allogeneic islet transplantation for type 1 diabetes resulting in a short period of insulin independence and published in 1990 (9). However, it was not until 2000 that the Edmonton group, led by James Shapiro, reported consistent achievement of insulin independence in seven out of seven consecutive patients with type 1 diabetes who received islets isolated from an average of two donor pancreases (10). This established the need for transplantation of a sufficient islet β-cell mass to attain insulin independence and near-normal glycemic control in humans with type 1 diabetes. It also required the avoidance of glucocorticoid use in the immunosuppression regimen to avoid islet toxicity and insulin resistance.

During the time since publication of the Edmonton protocol, many countries internationally have advanced allogeneic islet transplantation to the treatment of type 1 diabetes patients experiencing severe hypoglycemia and hypoglycemia unawareness or marked glycemic lability (11), or those already committed to chronic immunosuppression in support of a kidney transplant (12–14). Meanwhile, efforts in the United States have focused on improving the efficiency of both islet isolation and posttransplant islet engraftment to minimize the number of donor pancreases required to stabilize metabolic control and, when possible, eliminate the requirement for insulin therapy (15–17). In 2005 the Minnesota group led by Bernard Hering introduced several modifications to peritransplantation management and reported consistent achievement of insulin independence in eight out of eight consecutive patients with type 1 diabetes who received islets isolated from a single donor pancreas (18). This enhanced efficiency for allogeneic islet transplantation has been further evaluated by a phase 3 multicenter licensure trial conducted by the National Institutes of Health–sponsored Clinical Islet Transplantation Consortium and published in 2016 (19). Importantly, this trial selected patients with type 1 diabetes experiencing severe hypoglycemia accompanied by hypoglycemia unawareness or marked glycemic lability and achieved an HbA1c ≤6.5% in the absence of severe hypoglycemia, and thereby met the US Food and Drug Administration’s primary outcome of achieving an HbA1c <7.0% in the absence of severe hypoglycemia. A subsequent trial in Europe confirmed these benefits in comparison with intensive insulin therapy using a randomized controlled design (20).

This review is focused on updating our current knowledge about human islet transplantation, including the present clinical indications and outcomes for both islet autotransplantation and allotransplantation, as well as considering directions for future research. Less extensive information is included about successful pancreas transplantation, which was reviewed by Jennifer Larsen in Endocrine Reviews in 2004 (21), but nonetheless remains the most appropriate comparator for assessing the degree of success for islet transplantation. Lastly, novel sources of islet tissue for transplantation are considered, as well as the requirements they must meet to join pancreas and islet transplantation as approaches to β-cell replacement therapy in the treatment of diabetes.

Procedures for Islet Transplantation

Islet autotransplantation following total pancreatectomy

Islets transplanted in patients after undergoing total pancreatectomy and islet autotransplantation (TPIAT) are autologous, and as such, are not considered a manufactured biologic product by the US Food and Drug Administration. Thus, the practice of patients receiving their own islets requires processing under current good tissue practice (cGTP), but not the more stringent current good manufacturing practice (cGMP), regulations governing the processing of allogeneic islets for transplantation. Because a patient receives only those islets that can be isolated from their own pancreas, the approach to processing aims to maximize islet yield rather than islet purity.

The standard approach involves complete resection of the pancreas, most commonly in the setting of recurrent acute or chronic pancreatitis, together with the duodenum and the spleen, which both share their arterial blood supply with the pancreas (Fig. 1). When technically feasible, some centers will attempt to preserve the spleen in situ (22). The spleen may be preserved during total pancreatectomy either by separating the splenic artery and vein from the pancreas or by maintaining the short gastric vessels. The first approach affects in situ perfusion of the pancreas and is not recommended when islets will be isolated, whereas the second approach may lead to sinusoidal hypertension and the development of gastric varices. While gastrojejunostomy and choledochojejunostomy Roux-en-Y reconstruction is performed, the pancreas is brought to a cGTP, or where available cGMP, islet isolation facility where it is digested using a combination of chemical [collagenase and neutral protease blend (23, 24)] and mechanical (Ricordi chamber) methods, and it undergoes variable centrifuge purification that separates the acinar and ductal tissue from the islets (25, 26). Purification results in loss of some islets, so preparations remain rather impure to maximize islet recovery while minimizing by centrifugation the transplanted tissue volume. The final islet product is transplanted into the liver by infusing it intraportally via the splenic vein stump, a mesenteric vein, or recannulation of the umbilical vein (the latter directs islets only into the left portal vein). If excessive portal pressure is caused by a high tissue volume, a portion of the islet product may be placed in the peritoneal cavity or other nonhepatic sites. From the portal venous circulation the islets distribute throughout the liver wherein the islets lodge in the hepatic sinusoids that provide a natural bio-scaffold important to promote cell survival (27).

Figure 1.

Procedure for total pancreatectomy with islet autotransplantation. The diseased pancreas is resected during total pancreatectomy together with the duodenum and the spleen, which both share their arterial blood supply with the pancreas. While gastrojejunostomy and choledochojejunostomy Roux-en-Y reconstruction is performed, the pancreas is brought to a cGMP or cGTP islet isolation facility where it is digested by collagenases and undergoes variable centrifuge purification that separates the acinar from the endocrine cells with the goal of maximizing islet recovery while minimizing the transplanted tissue volume. The final islet product is transplanted intraportally via the splenic vein stump, a mesenteric vein, or recannulation of the umbilical vein. If excessive portal pressure is caused by a high tissue volume, a portion of the islet product may be placed in the peritoneal cavity or other nonhepatic site.

Because the source of islets is a diseased pancreas, several important challenges arise in optimizing islet yield and minimizing risk due to impure or contaminated preparations. Effective delivery of a collagenase enzyme blend to the pancreatic parenchyma is best achieved by transection of the midportion of the gland and cannulation of each half of the main pancreatic duct toward the pancreatic head and tail, respectively, for enzyme instillation. Dilated ducts require larger gauge catheters to prevent leakage of the collagenase solution during perfusion (26). This approach is markedly compromised by prior pancreatic surgery involving drainage procedures that open the main pancreatic duct along its length, and by extensive intraductal calcification that may develop in the course of chronic pancreatitis. Without effective delivery of collagenase enzymes for pancreas digestion, islets cannot be liberated from the exocrine compartment. Additionally, prior pancreatic surgery involving resection of pancreatic tissue, and especially distal pancreatectomy, affects the starting material available for digestion and may further compromise the potential islet yield following processing (28, 29). Finally, sterility of the islet product is often compromised by prior endoscopic procedures that breach the integrity of the sphincter of Oddi, including endoscopic retrograde cholangiopancreatography, sphincterotomy, and intraductal stenting, which allows access and colonization of small intestinal bacterial in the pancreatic ductal tree (26). Unlike with islet allotransplantaton, concern over bacterial contamination and a desire to complete the islet infusion during the same operation as the pancreatectomy has prevented more formal evaluation of islet culture prior to islet autotransplantation.

Given the complexities involved in isolating islets for autologous transplantation, and the relative rarity of an indication for TPIAT, there are far fewer clinical facilities capable of offering high-quality islet processing than pancreatic surgery. To provide islet autotransplantation to patients requiring total pancreatectomy closer to their homes, a few pancreatic surgery programs have partnered with regional or distant institutions offering islet isolation (30–32). The largest reported series (32) included 36 consecutive patients using a remote islet isolation laboratory with the time from pancreatectomy to receipt of islets for transplantation limited to ~8 to 9 hours, and it demonstrated similar clinical outcomes as prior reports using on-site islet processing (25, 33–36) (Table 1). This demonstrated feasibility of remote islet processing supports that lack of an on-site islet isolation facility should not discourage the use of islet autotransplantation for the prevention or amelioration of postsurgical diabetes in patients requiring pancreatectomy (37).

Table 1.

Comparison of Published High-Volume Center Experiences With TPIAT

Years	University of Minnesota (25)	University of Cincinnati (33, 34)	Leicester General Hospital (35, 36)	Cleveland Clinic(32)
	1997–2007	2000–2004	1994–2006	2008–2014
Number of patients	173	45	50	36
Time until islet infusion, h	2–7	3–4	2–4	8–9
Median islet yield, IE/kg	3054	4933a	2245	4308
HbA1c, 1-y median, %	NRb	NR	7.0	NR
HbA1c, 2-y median, %	NRb	NR	6.7	6.8
Insulin independence, 1-y median, %	32	NR	26	NR
Insulin independence, 2-y median, %	24	40	24	33

[Adapted with permission from Johnston PC, Lin YK, Walsh RM, et al. Factors associated with islet yield and insulin independence after total pancreatectomy and islet cell autotransplantation in patients with chronic pancreatitis utilizing off-site islet isolation: Cleveland Clinic experience. J Clin Endocrinol Metab 2015;100:1765-1770. Illustration presentation copyright by the Endocrine Society.]

Abbreviations: IE/kg, IE/kg recipient body weight, whereby an IE approximates a standard islet diameter of 150 μm; NR, not reported.

^a

Mean reported (34).

^b

A subgroup (n = 8) in whom insulin independence was tested all had HbA1c <7.0% (25).

Islet engraftment in the liver has been essential for the success of clinical islet transplantation to date. As a site, the liver provides transplanted islets with oxygenation via the portal circulation until revascularization by the hepatic arterial system occurs (38, 39). With revascularization there is also evidence for reinnervation of intrahepatic islets by the sympathetic nervous system in rodent models that is important for modulating islet activity and hormone secretion (40). Additionally, insulin and glucagon secreted from islets normally enter the portal circulation and are ~50% extracted by the liver where insulin suppresses and glucagon activates hepatic glucose production. The insulin delivered to the liver is secreted in coordinate pulses, the amplitude of which contributes to insulin action on the liver, and is directly dependent on the functional islet β-cell mass. When transplanted in the liver via portal vein infusion, human islets reestablish coordinated pulsatile secretion of insulin, and although the islets ultimately drain into the hepatic venous circulation, direct hepatic vein catheterization studies in two autoislet recipients indicated normal first-pass hepatic extraction, supporting that intrahepatic islets deliver insulin directly to the hepatic sinusoids (41). Indeed, the liver is the site deemed most efficacious in providing survival and function of transplanted islets to consistently reverse diabetes and achieve insulin independence in large animal models (42). Maximizing islet survival until engraftment is particularly crucial for autologous transplantation when there is no additional source of islet tissue.

As already mentioned, there is a trade-off with islet purification in that the quantity of islets recovered decreases with further purification. Thus, for the sake of maximizing islet recovery for transplantation, preparations of islets for autologous transplantation have lower purity than do preparations for allogenic transplantation, and the higher tissue volume associated with impure islet preparations increases the risk for portal hypertension and portal vein thrombosis, complicating intraportal islet delivery to the intrahepatic site. Purification should be used to reduce the final tissue volume for transplantation to <0.2 to 0.25 mL/kg recipient body weight up to a maximum of 15 to 20 mL (25, 43–45). For higher tissue volumes, or when excessive portal pressure develops during intraportal islet infusion (>25 cm H₂O) (45), a portion of the islet product may be placed in the peritoneal cavity or other nonhepatic site (29). Alternative sites for islet transplantation should only be used judiciously when it is unsafe to deliver the entire preparation to the liver given the limited experience and uncertain efficacy with nonhepatic islet transplantation. Attempts at placing the entire islet product in a nonhepatic site have rarely been attempted and to date have proved less successful. In three adult patients receiving autoislets in the brachioradialis muscle of the forearm, none maintained posttransplantation HbA1c <7.0% despite evidence of stimulated C-peptide production (46), whereas the same approach in a 7-year-old child did normalize glycemia and support growth with the provision of exogenous insulin during 2 years of reported follow-up (47). The intramuscular site has not been able to support the survival of alloislets (48). Four patients with contraindication to intraportal islet infusion underwent transplantation of autoislets to the bone marrow of the iliac crest that resulted in very marginal graft function (49). The bone marrow site is also expected to be more immunogenic than the liver for alloislets (50). The peritoneal cavity has been successful using the entire islet preparation in dogs (51) and, as mentioned above, this site is used in humans when excess islets remain after using the liver site (29, 52).

Islet allotransplantation for the treatment of type 1 diabetes

Islets transplanted in patients with type 1 diabetes are allogeneic, and as such are regulated as a manufactured biologic product by the US Food and Drug Administration, with islet production requiring cGMP quality standards. In other countries such as Canada, Australia, and several in Europe with national health services, islet transplantation is available as standard of care for selected patients (see “Indications for Islet Transplantation” below). Importantly, the selection of a donor pancreas for islet isolation must adhere to ABO blood group antigen compatibility, avoid any unacceptable human leukocyte antigens (HLAs) that could react with preformed anti-HLA antibodies present in the recipient, and pass a T lymphocyte and B lymphocyte crossmatch between donor cells and recipient blood. Preexisting immunity against HLAs develops from previous exposure to nonself HLAs as occurs with pregnancy (the father’s HLA haplotype expressed by the fetus), red blood cell transfusions (that are ABO but not HLA matched), solid organ transplantation (e.g., prior kidney transplantation for diabetic nephropathy), human bone and tendon graphs commonly used in orthopedics, and human heart valves and ventricular assist devices sometimes required in cardiac surgery. Thus, close communication with both the organ procurement organization and institution tissue typing laboratory is essential for minimizing the risk of antibody–antigen recognition leading to acute rejection of the allogeneic product.

The standard approach involves procurement of a pancreas from a deceased donor without diabetes (Fig. 2). Procurement, preservation, and transport of the deceased donor pancreas for islet transplantation requires the same care and speed as that for a pancreas used for whole organ transplantation (53), with cold ischemia time kept to a minimum (<8 hours) (17). Although younger donors (<40 to 45 years of age) provide higher quality pancreases and islets (54, 55), and so may be optimal for both whole organ and isolated islet transplantation (56), fatty pancreases common in our increasingly overweight and obese population carry an increased risk for technical failure following whole pancreas transplantation. At the same time, organs procured from higher body mass index donors lend to higher islet yields following isolation (57), and so islet transplantation can make use of high-quality organs that would otherwise be discarded (58) and more often yield sufficient numbers of islets important for posttransplant outcomes (59). The successful use of nonheart-beating donors for clinical islet transplantation has been reported (60, 61), but such donation after circulatory death donors are infrequently used as a source of islets outside Japan where (heart-beating) brain death donation is extremely limited. Recent guidelines have been proposed on the procurement of pancreases from donation after circulatory death donors for use in transplantation, including a limit on warm ischemia time (<30 to 60 minutes) (62), but no comparative data on clinical outcomes with the standard use of brain death donors are available.

Figure 2.

Procedure for islet allotransplantation. A pancreas is procured from a deceased donor without diabetes and brought to a cGMP islet isolation facility where it is digested by collagenases and purified by centrifugation with the final islet product placed in culture for up to 72 h prior to intraportal delivery in a recipient with type 1 diabetes. Shown here is percutaneous, transhepatic access to the main portal vein via a right portal branch vein identified under ultrasound and fluoroscopic guidance. Alternatively, access to the main portal vein can be achieved via a mesenteric vein identified by direct visualization during minilaparotomy.

Once brought to the cGMP islet isolation facility, the deceased donor pancreas is digested by intraductal collagenase delivery (23, 24), mechanically separated using the Ricordi method (8), and purified by centrifugation with the final islet product placed in culture for 36 to 72 hours prior to intraportal delivery in a recipient with type 1 diabetes (63) (Fig. 2). Unlike the autoislet procedure, a period of islet culture may be particularly helpful for allogeneic transplantation because islet culture has been shown to decrease the number of passenger leukocytes and islet expression of class I HLA (64), as well as islet production of CCL2/MCP-1 and tissue factor (65–68), and so may result in a less immunogenic and inflammatory graft. Although direct comparisons of clinical outcomes from islet preparations with or without culture are not available, improved outcomes have been associated with the adoption of islet culture in the Collaborative Islet Transplant Registry (69), and a period of islet culture allows for the induction of immunosuppression in the recipient before exposure to allogeneic tissue. Access to the main portal vein for islet infusion can be achieved either via percutaneous, transhepatic access of a portal branch vein identified under ultrasound and fluoroscopic guidance (70, 71), or via a mesenteric vein identified by direct visualization during minilaparotomy (72).

Similar to the use of regional islet isolation facilities to facilitate access to autologous islet transplantation, several regional networks have developed in support of providing allogeneic islet transplantation for type 1 diabetes (58, 73–76). This approach not only brings the treatment closer to where a patient might live, but also takes advantage of the procurement of organs by teams outside the procurement area of the isolation facility who are committed to high-quality recovery and rapid transport of pancreata to the processing institution and appropriate preparation of the recipient while awaiting the return of the purified human pancreatic islet product for transplantation. Similar clinical outcomes have been reported for recipients of islets processed on-site or off-site following the same study protocol (58, 76).

In addition to the liver as a transplant site providing islets with oxygenation via the portal circulation until revascularization by the hepatic arterial systems occurs (38, 39), also important for allogeneic islet transplantation where islets from more than one donor pancreas may be transplanted is the capacity of the liver to accommodate sequential infusion of islets from three or even more donor pancreases (77). Nonetheless, exposure to more donor pancreases increases the risk of sensitization (development of alloantibody) to HLAs, and so an optimal approach for the patient with type 1 diabetes minimizes the number of donors required to achieve metabolic benefit. The major limitation with the liver as a site for allogeneic islet transplantation is the inability for biopsy to reliably include islet tissue necessary for pathologic evaluation (78), and so precludes accurate attribution of the cause for islet graft dysfunction or failure. The only report of successful islet engraftment outside of the liver is a case using an omental pocket created with a bioartificial thrombin scaffold (79). Although this approach may benefit patients with liver disease precluding intrahepatic transplantation, currently its use is limited to one time therapy and is not any more accessible to biopsy monitoring.

Indications for Islet Transplantation

Islet autotransplantation

Recurrent acute and chronic pancreatitis

Islet autotransplantation is indicated when total pancreatectomy is required for treatment of recurrent acute or chronic pancreatitis associated with frequent debilitating exacerbations and/or chronic, unrelentingly pain. The rationale for total pancreatectomy is removal of the source of pain and disease exacerbations to improve a very poor quality of life, reduce or eliminate chronic narcotic use, and facilitate the ability to maintain work and/or child care responsibilities. For most patients with chronic pancreatitis there is the eventuality of developing pancreatogenic (type 3c) diabetes due to progressive destruction of the pancreatic microarchitecture by fibrosis leading to islet damage and loss (80, 81). However, with total pancreatectomy the development of insulin-dependent diabetes is certain and immediate and must be understood as acceptable to patients considering surgery (29, 36). The rationale for transplanting the patients’ own islets into the liver is to prevent or ameliorate the diabetes caused by pancreatectomy. Islet autotransplantation does not decrease the risk of type 3c diabetes by relocating healthy islets from a diseased pancreas to the liver, but recovery of a high quantity of healthy islets is more likely to result in satisfactory control of the postpancreatectomy diabetes. For patients already experiencing diabetes, some institutions still offer islet autotransplantation as long as there is preoperative evidence of some preserved β-cell capacity for secretion assessed by stimulated C-peptide (82); however, evidence that this approach benefits control of the postoperative diabetes is lacking.

The causes of chronic pancreatitis in 409 adults and children 5 to 69 years of age treated by TPIAT at the University of Minnesota include idiopathic (41%), pancreas divisum (17%), hereditary [PRSS1 mutations (83)] and other genetic (14%), sphincter of Oddi dysfunction (9%), and alcohol abuse (7%) (29). It is now clear that interactions between obstructive, genetic [e.g., SPINK1 and CFTR mutations (84)], environmental, and metabolic factors contribute to the development of recurrent acute and chronic pancreatitis (85). Historically, patients contend with chronic pancreatitis and attendant interventional attempts, including cholecystectomy, sphincterotomy, pancreatic duct stenting, and pancreatic surgery, for an average of almost 10 years before pancreatectomy is performed as a last resort (29). Although pancreatic drainage procedures are indicated for decompression of a large, dilated main duct in the setting of pancreatitis (86), these approaches have proven ineffective in the treatment of patients with “small duct” disease. Currently, earlier consideration of pancreatectomy is being advocated to shorten the seemingly inevitable progressive decline in quality of life, with possible advantages of preventing opioid addiction and increasing the yield of islets recovered for transplantation (37). Resolution of pain following TPIAT is more much likely in the setting of recurrent acute than once chronic pancreatitis has become established (87).

Trauma and benign neoplasms of the pancreas

Even less common indications for islet autotransplantation include following pancreatic resection for trauma or other benign neoplasms of the pancreas (88–90). Typically, when trauma or benign tumors present in the neck of the pancreas, including insulinomas not amenable to enucleation (91), the ~70% to 80% pancreatectomy required for mass resection normally carries a high risk for postsurgical diabetes that can be prevented by islet autotransplantation. For glands that have not undergone previous instrumentation and are unlikely to have bacterial colonization, a period of islet culture can be applied while pathologic assessment is performed to ensure that any resected lesion is benign (89, 90). Remarkably, a case of successful islet autotransplantation following completion pancreatectomy for gunshot wound trauma sustained by a US airman with remote islet isolation and infusion within 24 hours of surgery has been reported (92, 93).

Successful islet autotransplantation cases have been reported following total pancreatectomy for indolent renal cell carcinoma metastases (94) and following salvage total pancreatectomy for pancreatic or ampullary adenocarcinoma (95–97); however, the risk of potentially disseminating otherwise resected malignant cells to the liver outweighs the potential benefit for glycemic control in patients with known malignant disease (98). In a series of 31 patients with known malignancy who underwent TPIAT, 3 developed de novo liver metastases during a median 2.5 years of follow-up (90).

Islet allotransplantation

Type 1 diabetes

Islet allotransplantation is indicated to treat type 1 diabetes complicated by hypoglycemia unawareness associated with severe hypoglycemia events and/or excessive glycemic lability, or by hyperglycemia associated with a stable functioning kidney graft. All patients should have completed a structured education program on basal-bolus insulin delivery with flexible dosing of modern insulin analogs using pump or multidose injection delivery based on frequent self-monitoring of blood glucose with or without continuous glucose monitoring.

The rationale for islet transplantation in hypoglycemia unawareness is that in long-standing type 1 diabetes with near-total destruction of islet β-cells and the accompanying loss of the α-cell response to hypoglycemia (99), replacement of islet function by transplantation addresses the underlying pathophysiology and allows for avoidance of hypoglycemia and stabilization of glycemic lability that otherwise contribute to the development of impaired awareness of hypoglycemia and risk for experiencing severe hypoglycemia events. Current evidence-informed guidelines recommend that patients with problematic hypoglycemia, defined by experiencing two or more episodes per year of severe hypoglycemia or by experiencing one episode in the context of impaired awareness of hypoglycemia, extreme glycemic lability, or major fear and maladaptive behavior, should be considered for either pancreas or islet transplantation (100). What effect recently available continuous glucose monitoring communicating pumps with automated insulin delivery (hybrid closed loop) may have on type 1 diabetes patients with problematic hypoglycemia is not yet known, as patients experiencing severe hypoglycemia have been excluded from registration trials of artificial pancreas technologies (101).

The rationale for islet transplantation when hyperglycemia is associated with a stable functioning kidney graft is that improvement in glycemic control may protect the transplanted kidney from recurrence of diabetic nephropathy with minimal added risk because immunosuppression is already required. Some institutions consider this simultaneously with kidney transplantation from a deceased donor, which has the advantage of minimizing exposure to additional HLAs because the islets and kidney come from the same donor, but the disadvantage of being limited to the number of islets isolated from the paired pancreas. Other patients with type 1 diabetes considered for islet transplantation are those with problematic hyperglycemia, defined by the presence of recurrent episodes of diabetic ketoacidosis or severe, rapidly progressing secondary complications of diabetes (102, 103). A rare but important indication is for patients with type 1 diabetes complicated by allergy or resistance to subcutaneously administered insulin (104). Finally, retransplantation of alloislets prepared from a viable allograft pancreatectomy has been performed successfully in a patient with type 1 diabetes experiencing recurrent episodes of graft pancreatitis who initially received the pancreas transplant for hypoglycemia unawareness (105), and in a patient with type 1 diabetes who underwent simultaneous pancreas/kidney transplantation that was complicated by bleeding at the pancreas graft arterial anastomosis (106).

Pancreatogenic (type 3c) diabetes

In addition to type 1 diabetes, any cause of insulin-deficient diabetes such as cystic fibrosis–related diabetes or other pancreatogenic forms of diabetes (e.g., chronic pancreatitis or following pancreatectomy) may be considered for islet transplantation when β-cell failure is associated with glycemic instability and either problematic hypoglycemia or hyperglycemia despite availability of, and adherence to, optimized medical care (102, 103). Although it is expected that C-peptide levels in such individuals would be low, undetectable levels of C-peptide are not required for determination of eligibility (102, 103). Islet allotransplantation for cystic fibrosis–related diabetes with unstable glycemic control and frequent hypoglycemia has been reported simultaneously with bilateral lung transplantation in four cases using islets isolated from the same donor and transplanted after as much as 7 days of islet culture to enable initial recovery from the lung transplantation (107), in a case simultaneously with and after bilateral lung and liver transplantation (108), and in a case after lung transplantation that resulted in a prolonged period of insulin independence (109). Whether the observed improvements in glycemic control and nutritional status noted following islet transplantation may impact lung function and mortality in cystic fibrosis requires further study (110).

Peritransplant Management of the Islet Recipient

Care of the autoislet recipient

At the time of pancreatectomy, insulin therapy is initiated IV per a modified hospital protocol targeting normoglycemia (32, 35) that serves to “rest” the islets at the time of transplantation when they have been stressed by the isolation procedure and are devascularized and so relatively underoxygenated (Fig. 3). Exposure of devascularized, underoxygenated islets to hyperglycemia has been shown to induce β-cell apoptosis (111) and should be strictly avoided (112). Furthermore, hyperglycemia has detrimental effects on both the function (113) and revascularization (114) of isolated human islets. Importantly, the degree of glycemic control achieved during the first 5 postoperative days is associated with the likelihood of achieving long-term insulin independence (32). At the time of islet infusion, anticoagulation is initiated with unfractionated heparin, usually at a dose of 50 to 70 U/kg recipient body weight (25, 28, 29, 32, 35, 43), which is either divided equally among the islet product bags [each composed of no more than 5 mL of tissue per 200 mL of infusion medium (26)] or between the islet product and a peripheral venous infusion. As soon as hemostasis is assured postoperatively, a continuous infusion of heparin is given either at a fixed rate (e.g., 500 U/h) or titrated to achieve and maintain the partial thromboplastin time (PTT) of ~50 seconds for the first 48 to 72 hours. The initial goal for anticoagulation is to prevent portal vein thrombosis, and because there is a balance between thrombosis risk and surgical bleeding risk, Doppler ultrasound of the liver on the first postoperative day assessing blood flow in the main as well as right and left branches of the portal vein is helpful to determine whether more aggressive anticoagulation is indicated, or whether less aggressive anticoagulation may be tolerated should there be an observed increase in surgical drain output or decrease in hematocrit suggestive of surgical site bleeding. In addition to serially monitoring of the hematocrit and PTT, hepatic transaminase levels are assessed as indicators of transient ischemic injury to the hepatic parenchyma following intraportal islet infusion, and in the absence of a resolving trend after the first few days may prompt repeat ultrasonography of the liver. Perioperative antibiotic prophylaxis should follow institutional practice for coverage against skin as well as upper gastrointestinal tract flora, and so have activity against both Gram-positive cocci and enteric Gram-negative bacilli until culture results are available from the transplanted islet product.

Figure 3.

Peritransplant management for islet autotransplantation. At the time of pancreatectomy, insulin therapy is initiated IV according to a modified hospital protocol targeting normoglycemia, and at the time of islet infusion anticoagulation is initiated with unfractionated heparin. Once stable postoperatively and as oral intake is advanced, IV insulin and heparin are transitioned to subcutaneous regimens, antiplatelet therapy with aspirin is added, instruction in a carbohydrate-controlled diet is provided, and pancreatic enzyme replacement and acid-blocking therapy is initiated or resumed.

Once stable postoperatively and as oral intake is advanced, IV insulin and heparin are transitioned to subcutaneous regimens. Exogenous insulin injections are given for at least several weeks after the islet infusion to allow time for establishment of arterialization and adequate oxygenation of the intrahepatic islets (115, 116). Thus, education is required in the provision of intensive insulin therapy with basal-bolus insulin dosing based on at least four times daily glucose monitoring, and in adherence to a carbohydrate controlled diet that is also low in fat and accompanied with chronic treatment with oral preparations of pancreatic enzymes required to assist with digestion and absorption of nutrients and avoidance of diarrhea (117). Acid-blocking therapy is initiated or resumed both to prevent gastric inactivation of the administered pancreatic enzymes and to prevent the development of marginal ulceration around the gastrojejunal anastomosis. Protein supplements including high-protein shakes are often required to meet nutrient requirements and mitigate weight loss. Total enteral or parenteral nutrition is sometimes required in the postoperative period for protein-calorie malnutrition, and hyperalimentation can result in hyperglycemia and hepatic toxicity (118). Care should be taken to minimize the calories provided by simple sugars and adjust insulin doses ahead of planned changes in nutrition therapy to maintain normoglycemia during this critical period of intrahepatic islet engraftment.

After 48 to 72 hours of anticoagulation with heparin, in the absence of evidence for portal main or branch vein thrombosis, heparin may be transitioned to subcutaneous administration at high prophylactic dosing for at least the first 1 or 2 weeks (43) or until hospital discharge, which also serves to prevent microthrombi formation that can interfere with islet engraftment and revascularization. As a consequence of splenectomy, thrombocytosis is frequently observed, and antiplatelet therapy with full-dose aspirin is added as soon as surgically appropriate, or with low-dose aspirin when prolonged anticoagulation is planned. By discharge, those with splenectomy require immunization against encapsulated bacteria, including pneumococci, meningococci, and Haemophilus influenza type b, to reduce the risk of future postsplenectomy sepsis (119). Children should complete vaccination at least 2 weeks before splenectomy, and they require antibiotic prophylaxis for the first year postoperatively (120). Annual influenza vaccination should be performed using recommended inactivated and recombinant vaccines. Because the islets are autologous, no immunosuppressive drugs are required.

Care of the alloislet recipient

At the time a compatible islet preparation is available and enters culture (Fig. 4), the intended recipient receives induction and maintenance immunosuppression in the hospital where intensive insulin therapy is maintained targeting normoglycemia that again serves to “rest” the islets at the time of transplantation when they have been stressed by the isolation procedure and are devascularized and so relatively underoxygenated and susceptible to glucotoxicity (121). While practice varies in terms of islet culture time and choice and timing of induction agents, in the phase 3 study conducted by the Clinical Islet Transplantation Consortium (B7), the T lymphocyte–depleting polyclonal antibody thymoglobulin was initiated with the standard 1.5 mg/kg dose divided as 0.5 mg/kg on day −2 and 1.0 mg/kg on day −1, with additional 1.5 mg/kg doses on days 0, 1, and 2 when tolerated per prescribing guidelines. First piloted by Hering et al. (18), this approach allows for glucocorticoid premedication prior to the first dose on day −2 and dissipation of effects from the administered glucocorticoid and cytokines released from lysed T lymphocytes that are toxic to β-cells by allowing at least 36 hours to pass from the first dose of thymoglobulin to the islet infusion. Because the cytokine TNF-α is particularly toxic to β-cells and further may contribute to the induced expression of class II HLAs on alloislets (122), slow-release pentoxifylline is administered from day −2 to day 7 because it can inhibit TNF-α production (123), and immediately prior to the islet infusion the TNF-α inhibitor etanercept is administered IV followed by three additional doses subcutaneously on days 3, 7, and 10. Anticoagulation is again initiated with unfractionated heparin at a dose of 70 U/kg recipient body weight that is divided equally among the islet product bags. As soon as hemostasis is assured after the procedure, a continuous infusion of heparin is titrated to achieve and maintain the PTT at ~50 to 60 seconds for the first 48 hours, after which unfractionated heparin is transitioned to a subcutaneous regimen consistent with high prophylactic dosing for at least the first week (124), and it is discontinued only with Doppler ultrasound of the liver demonstrating normal blood flow in the main as well as right and left branches of the portal vein. In the absence of any procedurally related bleeding complication, antiplatelet therapy with low-dose aspirin is added by the second day (124). Again, serially monitoring of the hematocrit and PTT is required for the first few days together with hepatic transaminase levels as indicators of transient ischemic injury to the hepatic parenchyma following intraportal islet infusion (125, 126). Perioperative antibiotic prophylaxis should follow institutional practice for coverage against skin flora until culture results are available from the transplanted islet product.

Figure 4.

Peritransplant management for islet allotransplantation. At the time a compatible islet preparation is available and enters culture, induction and maintenance immunosuppression is initiated in the hospital where intensive insulin therapy is maintained targeting normoglycemia. At the time of islet infusion, anti-inflammatory therapy (e.g., etanercept) and anticoagulation are initiated, with unfractionated heparin transitioned to a subcutaneous regimen with antiplatelet therapy using low-dose aspirin added by the second day. [Shown is the B7 protocol from Hering BJ, Clarke WR, Bridges ND, et al. Phase 3 trial of transplantation of human islets in type 1 diabetes complicated by severe hypoglycemia. Diabetes Care 2016;39:1230–1240; other approaches may vary.]

Exogenous insulin is maintained for at least several weeks after the islet infusion to allow time for establishment of arterialization and adequate oxygenation of the intrahepatic islets. On the day of transplantation, it is helpful to transition from subcutaneous to IV insulin delivery that serves to both more effectively target normoglycemia prior to the islet infusion (124) and allow for temporary withdrawal of insulin should hypoglycemia result during the first 6 hours following infusion from islet degranulation (127). Any hypoglycemia can be corrected by continuous infusion of 10% dextrose with care taken not to overcorrect and expose the transplanted islets to hyperglycemia, with removal and reinstitution of IV insulin as soon as stable normoglycemia is restored. Transition back to subcutaneous insulin delivery should occur prior to the first posttransplant meal where bolus insulin should follow the individual’s pretransplant carbohydrate ratio. It remains important that intensive insulin therapy with basal-bolus insulin dosing based on at least four times daily glucose monitoring be continued together with adherence to a carbohydrate-controlled diet to maintain normoglycemia during this critical period of islet engraftment.

Immunosuppression for the alloislet recipient

Most current protocols implement induction immunosuppression with T lymphocyte–depleting agents such as thymoglobulin as discussed above (18, 19, 76) or alemtuzumab, a humanized monoclonal antibody against the cell surface glycoprotein CD52 (58, 128), or T lymphocyte inhibitory agents that block the IL-2 receptor (e.g., basiliximab). T lymphocyte depletion results in more effective immunosuppression than does IL-2 receptor blockade (129, 130), but results in cytokine release for which mitigation strategies such as that described above in the B7 protocol are likely important for optimizing transplanted islet survival and engraftment. In support of this concept, data from the Collaborative Islet Transplant Registry indicate that induction regimens based on a T lymphocyte–depleting agent result in superior rates of long-term insulin independence when compared with use of an IL-2 receptor antagonist, but only when combined with a TNF-α inhibitor (e.g., etanercept) (131). Consistent with this analysis, a single-arm trial adding the TNF-α infliximab to the IL-2 receptor antagonist daclizumab did not result in an improvement in the rate of insulin independence achieved over historical experience using daclizumab alone (132). Nevertheless, IL-2 receptor antagonists are preferentially used at the time of subsequent islet infusions to minimize adverse effects related to overimmunosuppression (19), and when the presence of allergy precludes use of a T lymphocyte–depleting agent (133).

Maintenance immunosuppression for islet transplant recipients remains calcineurin inhibitor based, either as low-dose tacrolimus used in combination with the mammalian target of rapamycin (mTOR) inhibitor sirolimus (also known as rapamycin) (19), or more standard dosing tacrolimus used in combination with mycophenolic acid as for kidney transplantation (58, 76). Substituting the calcineurin inhibitor with cyclosporine and the mTOR inhibitor with everolimus has also demonstrated efficacy in a phase 2 study (134). Glucocorticoids are avoided for recipients of islet transplantation alone; however, maintenance dosing of glucocorticoids (e.g., prednisone at 5 mg daily) that may be in place in support of a prior kidney transplant is generally continued, as this dosing remains physiologic and has not been associated with inducing insulin resistance (135). Antimicrobial prophylaxis against Pneumocystisjiroveci pneumonia is typically achieved with trimethoprim/sulfamethoxazole, and duration should consider recovery of the lymphocyte count when depleting agents are used for induction. In the case of sulfa allergy, atovaquone is preferred over dapsone as a second line agent because dapsone can affect interpretation of the HbA1c (136). Antiviral prophylaxis against cytomegalovirus (CMV) reactivation is required when either the donor or recipient evidences prior exposure by positive CMV antibodies and can be achieved with valganciclovir. When both donor and recipient are CMV negative, antiviral prophylaxis against herpes simplex and varicella zoster virus reactivation can be achieved with acylclovir. Standard immunizations for persons with diabetes should be up to date, including against pneumococcus, as well as diphtheria, tetanus, and pertussis, whereas those against live attenuated viruses such as varicella zoster virus are contraindicated. Annual influenza vaccination should be performed using recommended inactivated and recombinant vaccines, not those using live attenuated virus.

Clinical Benefits of Islet Transplantation

Outcomes following islet autotransplantation

Pain and quality of life

In a subgroup (n = 207) of the largest cohort of 409 autoislet recipients from the University of Minnesota treated between 1977 and 2011, 85% experienced resolution or significant improvement in pain and narcotic use (29). In a subgroup (n = 193) of the 250 patients treated after 2006 reporting on health-related quality of life during 2 years postoperatively, SF-36 measures improved significantly in all dimensions, irrespective of the ability to discontinue narcotic analgesia (29). Insulin independence was associated with greater improvement in dimensions of physical functioning, limitations, and vitality than for those who were insulin-dependent (29). Overall results of TPIAT were judged as good to excellent by 90% of recipients, fair by 8%, and poor by 2% (29). In another series of 85 consecutive patients undergoing total pancreatectomy (50 of whom received islet autotransplantation), >90% required chronic opiate analgesia before surgery, which reduced to ~40% at 1 year and further to ∼15% at 5 years (36), again supporting 85% effectiveness of pancreatectomy in addressing painful pancreatitis. Patient-reported outcomes were available from 50 of the 85 patients, with significant reductions noted in pain severity and frequency, and all viewing the operation as successful, although a few noted experiencing frequent problems (36). Assuming that missing data occurred at random in these larger studies, the results are consistent with earlier (33, 137) and smaller series (138) supporting that TPIAT effectively addresses the primary indication of chronic pain and reduced quality of life in ~80% of recipients.

Insulin independence and glycemic control

In most large series of islet autotransplantation at a median follow-up of 2 years approximately one-third of patients are insulin-independent (Table 1), approximately one-third are managed with daily long-acting insulin alone, and approximately one-third require intensive insulin therapy with a basal-bolus regimen administered by multidose injection or pump delivery (25, 32, 34, 36). Patients with minimal change disease on pancreatic imaging have higher isolated islet yields that are associated with higher likelihood of achieving insulin independence (32, 139). Similarly, preoperative imaging of pancreatic atrophy, calcifications, or ductal dilatation were each associated with lower isolated islet yields (140) and measures of posttransplant islet function (141). An early series from the University of Minnesota reported that about three-fourths of recipients receiving >300,000 autoislets [the number of islets normally found in a healthy pancreas is ~1 million (142, 143)] were insulin-independent at 2 years (137). Subsequently, insulin independence at 2 years was reported in about two-thirds of patients receiving >5000 islet equivalents [IE; standardized to an islet diameter of 150 µm (144)] per kg body weight, compared with about one-third of patients receiving <5000 but >2500 IE/kg, and very few patients receiving <2500 IE/kg (25). During 3 years of follow-up, HbA1c <7.0% was maintained in 94% who received >5000 IE/kg, with this rate declining to 86% and 71% when <5000 but >2500 IE/kg and <2500 IE/kg were used, respectively; overall, HbA1c <7.0% was maintained in 82% of recipients (29). Note, however, that despite these generally excellent outcomes in glycemic control when sufficient islets are used, some autoislet recipients have reported problems with postprandial hypoglycemic episodes (see below).

Pediatric outcomes

When considering children <18 years of age, outcomes for pain resolution and quality-of-life improvement are similar to those reported for adults, with >80% able to attend school or work (145). A study of 19 consecutive children aged 5 to 18 years undergoing TPIAT demonstrated significant improvement in all SF-36 measures (146). Insulin-independence rates are closer to 40% at 2 years in children undergoing TPIAT, with >50% of the youngest recipients between 5 and 12 years old insulin-independent at 1 year (120). In children during 3 years of follow-up, HbA1c <7.0% was maintained in 100% who received >5000 IE/kg, with this rate declining to 96% and 81% when <5000 but >2500 IE/kg and <2500 IE/kg were used, respectively (120).

Younger age and shorter duration of pancreatitis have been associated with higher islet yields, and similar to adults (137, 140, 147), prior pancreatic surgery and more extensive pancreatic fibrosis and atrophy were associated with lower islet yields (148), again arguing for earlier consideration of TPIAT in the course of recurrent acute or chronic pancreatitis (32). A subsequent study of 17 children between 3 and 8 years of age who underwent TPIAT demonstrated pain resolution in all, and two-thirds were insulin-independent with a median 2 years of follow-up (149). However, the durability of insulin independence throughout puberty and adolescence where both hormonal insulin resistance and accelerated growth significantly increase demand for insulin secretion remains uncertain (150), and longer-term follow-up in pediatric patients is warranted.

Long-term outcomes

There is attrition in insulin independence with time; available series with 5-year follow-up have reported rates between 10% and 15% (25, 36), or approximately half the rates initially achieved that continue to depend on the transplanted islet mass (29). Nevertheless, many patients have been reported with insulin independence persisting for >10 years following TPIAT (25, 29, 120, 151). For previously nondiabetic patients undergoing ~70% to 80% pancreatic resection for trauma or benign neoplasms located at the neck of the pancreas and subsequent islet autotransplantation, all achieved long-term normoglycemia, with ~95% insulin-independent at 2 years (90, 152, 153) and >90% remaining so with up to 7.5 years median follow-up (153). Thus, strategies that preserve or enhance the surviving functional islet β-cell mass are critical to ensuring long-term durability of the metabolic benefit afforded by islet autotransplantation.

Outcomes following islet allotransplantation

Early clinical results

The Edmonton protocol established that glucocorticoid-free immunosuppression together with a subsequent islet infusion from a second donor pancreas could reproducibly render a recipient with type 1 diabetes insulin-independent provided >9000 IE/kg were transplanted (11). However, by a median 2 years of follow-up most patients returned to requiring some insulin therapy despite ongoing evidence of islet graft function indicated by restored C-peptide production (154), suggesting establishment of only a marginal mass of surviving islets after transplantation. These disappointing long-term outcomes for insulin independence were confirmed by an international multicenter clinical trial of the Edmonton protocol, which also confirmed a more positive outcome of improved glycemic control without hypoglycemia experienced by >70% of recipients selected on the basis of experiencing refractory, disabling hypoglycemia (155). This multicenter clinical trial also demonstrated a center effect for achieving the primary outcome of insulin independence, where this occurred most frequently at sites with more extensive experience with pancreatic processing for islet isolation and immunosuppression management in transplant recipients. At the same time it became apparent that the withdrawal of immunosuppression for clinically failed islet grafts resulted in a high rate of sensitization to HLAs (156, 157), raising theoretical concern that finding suitable donor organs for possible future transplant indications may be more difficult. Thus, the emphasis in outcomes shifted away from insulin independence using islets procured from multiple donors to the stabilization of metabolic control using the least number of donor pancreases required for durable clinical benefit (158). This would require a focus on the peritransplant period to ensure that a larger number of high-quality islets survived the engraftment process.

Improving outcomes and quality of life

The biggest change came with the introduction of immunosuppressive protocols at the University of Minnesota that incorporated more potent induction agents with a period of islet culture and peritransplant anti-inflammatory, antithrombotic, and intensive insulin therapy, with improved rates of insulin independence occurring more frequently with islets isolated from a single donor (18) and most remaining insulin-independent for >2 years after transplant (134), suggesting that a greater proportion of the transplanted islets survived engraftment. The effect of peritransplant intensive antithrombotic and insulin therapy on increasing the rate of insulin independence achieved with islets transplanted from a single donor pancreas was independently confirmed (124). Importantly, these results were obtained using similar low-dose calcineurin inhibitor and mTOR inhibitor maintenance therapy as in the Edmonton protocol, providing evidence that appropriate dosing of currently available immunotherapy is not toxic to islets nor rate limiting for improving transplanted islet graft survival and functional durability. Although 1-year single donor insulin independence occurred less frequently (13 of 48 subjects) in the multicenter B7 trial (19) than in the pilot study from the University of Minnesota (5 of 8 subjects) (18), a similarly high rate of sustained single donor insulin independence (7 of 11 subjects) was reported by at least one other center participating in the B7 study (159), showing that these results are reproducible but may be center-dependent owing to careful attention paid to less emphasized variables such as maintenance of strict glycemic control during the period of islet engraftment. Examining the transplanted islet mass in these two cohorts further indicates that although not always sufficient, transplantation of >6000 IE/kg recipient body weight is necessary for a chance of obtaining insulin independence from a single donor (18, 159).

In the largest cohort of 677 alloislet recipients reported by the international Collaborative Islet Transplant Registry, when analyzed by year of transplantation as early (1999 to 2002, n = 214), middle (2003 to 2006, n = 255), or recent (2007 to 2010, n = 208), 55% of recipients from the most recent era remained insulin-independent at 2 years (69). Rates of insulin independence at 3 years after transplant improved from 27% to 37% to 44%, with the proportion of recipients receiving sequential islet infusions decreasing from 60% to 65% to <50% in the most recent era (69). Despite higher rates of sustained insulin independence observed with the transplantation of islets from fewer donor pancreases in the most recent era, receiving islets from more than one donor remained associated with a higher likelihood of achieving insulin independence across all eras by multivariate analysis (69). Other factors associated with a higher likelihood of achieving insulin independence in the multivariate analysis included increasing recipient age, islet culture ≥6 hours, islet stimulation index ≥1.5, and induction therapy with T lymphocyte depletion combined with a TNF-α inhibitor (69). Additionally, ever achieving insulin independence was associated with significantly prolonged islet graft function and accompanying metabolic stability regardless of a return to some insulin therapy (69). During 3 to 5 years of follow-up, near-normal glycemic control indicated by HbA1c <6.5% was maintained by ~60% of recipients across all eras. Although >90% of patients were experiencing severe hypoglycemia prior to transplant, >90% remained free of severe hypoglycemia events in all eras through 5 years of follow-up (69). These results support the benefit of islet culture (18, 132), islet quality, and more potent induction regimens (131) on achieving insulin independence, as well as the practice of considering a subsequent islet infusion when necessary to achieve insulin independence to maximize the long-term benefit of metabolic control imparted by a functioning islet graft (160, 161).

In a cohort of 23 alloislet recipients followed for 3 years, diabetes quality of life improved significantly and was mainly influenced by glycemic control and not insulin independence, although the reintroduction of insulin led to some modification of effect (162). Importantly, procedure- or immunosuppression-related adverse events did not negatively impact quality of life (162). A subsequent report of health-related quality of life that included 28 alloislet recipients followed for 3 years did not show significant improvement, although fear of hypoglycemia was markedly reduced during the period of observation to levels below that of disease controls with type 1 diabetes not being considered for islet transplantation (163). In a cohort of 27 alloislet recipients followed for 5 years, improvement was seen for depression and some but not all general health-related quality-of-life measures, with a robust and durable reduction in fear of hypoglycemia for both behavior and anxiety assessments (164). These results show a clear benefit in islet allotransplantation for improving quality of life that has been reduced by problematic hypoglycemia.

Standardizing outcomes

With the emphasis of islet allotransplantation shifting to establishing metabolic control in the absence of hypoglycemia, programs for islet alone transplantation established selection criteria based on patients experiencing severe hypoglycemia accompanied by hypoglycemia unawareness or marked glycemic lability and a primary outcome of achieving an HbA1c <7.0% in the absence of severe hypoglycemia (Table 2). The results of prospective single-arm clinical trials of islet allotransplantation indicate >80% of recipients achieve this primary outcome at 1 year (76, 165), and that this effect can be further sustained at 2 years in >80% of recipients (19). Importantly, in the phase 3 B7 trial only those recipients who lost islet graft function entirely experienced a recurrence of severe hypoglycemia, and so >90% of recipients were protected from experiencing a severe hypoglycemic event during 2 years of follow-up (19). Moreover, in the B7 study that involved the largest consecutive cohort of 48 transplant recipients, there were significant improvements in diabetes distress, fear of hypoglycemia, as well as patient self-assessments of personal well-being (166). Health-related quality-of-life measures improved significantly in five of the eight SF-36 domains, and results were not significantly different between those who achieved or did not achieve insulin independence (166). Thus, although the burden of chronic immunosuppressive drug treatment is often argued as a reason against consideration of islet transplantation as a therapy for type 1 diabetes, these phase 3 study results clearly demonstrate that appropriately selected candidates experience clinically significant and personally meaningful improvement in glycemic control and quality of life.

“…these phase 3 study results clearly demonstrate that appropriately selected candidates experience clinically significant and personally meaningful improvement in glycemic control and quality of life.”

Table 2.

Comparison of Published Consortium Trials With Standardized Outcomes for Islet Allotransplantation

Years	Australian (76)	United Kingdom (58)	GRAGIL-2 (165)	National Institutes of Health (B7) (19)
	2013a	2008–2011	2003–2010	2008–2012
Number of patients	17	20	24	48
Duration of islet culture, h	12–24	<24	<4	36–72
Median islet yield, IE/kg	15,366	8770	9716	11,972
HbA1c, 1-y median, %	6.5b	6.3	6.2	5.6
HbA1c, 2-y median, %	NR	6.2	NR	5.6
Insulin independence, 1-y median, %	NR	45	37.5	52
Insulin independence, 2-y median, %	47	15	46	42

Abbreviations: IE/kg, IE/kg recipient body weight, whereby an IE approximates a standard islet diameter of 150 µm (144); NR, not reported.

^a

Publication year.

^b

Mean reported (76).

These outcomes have recently been further validated against intensive insulin therapy. The multicenter Australian consortium assessed glycemic control and variability in patients with type 1 diabetes and severe hypoglycemia while on multiple daily injections of insulin, again after ~1 year of continuous subcutaneous insulin infusion (pump), and again 1 year following islet allotransplantation (167); continuous subcutaneous insulin infusion did not affect HbA1c but did modestly reduce glucose variability and lowered the rate of severe hypoglycemia events, whereas islet transplantation reduced HbA1c from a median 8.2% to 6.4% that was associated with a further marked reduction in glucose variability and elimination of hypoglycemia. The multicenter TRIMECO study enrolled patients with type 1 diabetes experiencing either severe hypoglycemia or poor glycemic control after kidney transplantation in the first randomized clinical trial to assess a transplant intervention for the treatment of diabetes (20); after 6 months the HbA1c was significantly lower with islet allotransplantation compared with intensive insulin therapy (5.6% vs 8.2%; P < 0.0001), and 23 of 25 subjects were protected from severe hypoglycemia events with islet transplantation compared with only 8 of 22 subjects receiving optimized medical management (P < 0.0001). Importantly, the outcome of achieving an HbA1c <7.0% in the absence of severe hypoglycemia was achieved in 21 of 25 subjects by islet transplantation and 0 of 22 with intensive insulin therapy (P < 0.0001), and there was significantly improved health-related quality of life in the transplant group compared with insulin group (20). Although the duration of comparative assessment in the TRIMECO study was limited to 6 months to allow subjects to cross over and receive the alloislet intervention, these results further substantiate by randomized clinical trial design the 2-year outcomes reported from the single cohort phase 3 B7 trial (19, 166).

Recent criteria have been developed for defining outcomes for β-cell replacement therapy in the treatment of diabetes regardless of the source being an islet or pancreas transplant (102, 103). These Igls criteria are based on the achievement of HbA1c targets, absence of severe hypoglycemia events, reduction in insulin requirements, and restoration of clinically significant C-peptide production, and they provide consistent classification of treatment success as that reported by the phase 3 B7 trial (19). Use of consistent definitions of graft functional and clinical outcomes is critical to informing the comparative effectiveness of existing and future forms of β-cell replacement therapy.

Long-term outcomes and comparison with pancreas transplantation

Longer-term outcomes have also significantly improved for graft function following islet allotransplantation. Although the initial 5-year outcomes data reported that with the Edmonton protocol 10% of alloislet recipients remained insulin-independent (154), other groups have reported 5-year insulin independence using the Edmonton protocol in >20% to 45% of both islet-alone and islet-after-kidney recipients (165, 168). In more recent protocols that depend on induction therapy with T cell depletion and TNF-α inhibition, 5-year insulin independence may be as high as 50% and significantly prolonged when compared with induction therapy based on IL-2 receptor blockade as used in the Edmonton protocol (131). As with TPIAT, patients are now being reported with insulin independence persisting >10 years following islet allotransplantation (169, 170).

As the rate of long-term insulin independence following islet transplantation is approaching the 5-year insulin independence rates of ~50% reported for pancreas transplantation alone and pancreas-after-kidney transplantation (171), consideration should be made for islet transplantation as an alternative to solitary pancreas transplantation because it carries considerably less risk for morbidity by avoiding major abdominal surgery. However, with simultaneous pancreas/kidney transplantation 5-year insulin independence is >70% (171), supporting the primacy of transplanting a pancreas simultaneously with a deceased donor kidney (when a living kidney donor is not available) for long-term durable graft function and resultant glycemic control. The superior long-term functional outcomes associated with simultaneous pancreas/kidney transplantation has been attributed to less risk for immunologic graft loss because transplanting organs from the same donor (and consequently HLA type) enables earlier detection of rejection episodes using biomarkers from both the kidney (serum creatinine) and pancreas (serum amylase and lipase) with treatment directed at one organ resolving subclinical rejection present in the biochemically stable organ. Insulin independence rates are lower with simultaneous islet/kidney transplantation, even when a second islet infusion is included, although reported series have relied on the Edmonton protocol (172). Nevertheless, this approach resulted in significant improvement in HbA1c without hypoglycemia when compared with a control group of type 1 diabetic individuals who had undergone kidney transplant alone, and significantly less morbidity than that seen in a control group of type 1 diabetic individuals who had undergone simultaneous pancreas/kidney transplantation where 40% required relaparotomy (172). Ideally, pancreas and islet transplantation should be offered as complementary, not competing, approaches to providing β-cell replacement to provide the most optimal clinical benefit to risk balance for each potential candidate while maximizing use of the largest number of quality donor pancreases offered for transplantation (173).

Metabolic Benefits and Consequences of Islet Transplantation

Islet autotransplantation following total pancreatectomy

β-Cell secretory capacity

The prevention or amelioration of postpancreatectomy diabetes by islet autotransplantation is dependent on the mass of islet β-cells that survives engraftment (151). As already discussed, most recipients receiving >300,000 autoislets (of the ~1 million islets normally found in a healthy pancreas) (137), or receiving >5000 IE/kg body weight achieve insulin independence, with very few receiving <2500 IE/kg meeting this objective (25). Because the number of islets that survive engraftment are always likely to be less than the number transplanted, the functional β-cell mass of an islet graft can be most accurately determined in vivo from the β-cell secretory capacity. The β-cell secretory capacity is measured by glucose potentiation of insulin or C-peptide release in response to injection of a nonglucose insulin secretagogue such as arginine or glucagon, and it correlates with calculated β-cell mass in animal models of β-cell reduction (174), and with resection (175) and transplantation (176–178) of a hemipancreas in humans. In insulin-independent, normoglycemic autoislet recipients, the β-cell secretory capacity correlates with the number of transplanted islets, an observation now replicated in independent cohorts of 8 and 10 patients (179, 180). In these cohorts of autoislet recipients selected on the basis of achieving optimal functional outcomes and studied between 1 and 8 years after TPIAT, the β-cell secretory capacity was ~35% to 40% of normal, very close to that predicted on the basis of the numbers of islets transplanted (179, 180) (Fig. 5). Whether the presumably lower engrafted β-cell mass for a comparable number of islets transplanted in less successful cases may be caused by nonspecific inflammatory and thrombotic mechanisms remains a focus of investigation (43). In four patients studied before and almost 1 year after TPIAT, the decrease in β-cell secretory capacity was inversely proportional to the number of islets transplanted, meaning that those who received more islets maintained closer to their presurgical insulin secretory reserve (179). Further longitudinal assessment of five patients who each received >300,000 autoislets and achieved insulin independence and normoglycemia demonstrated stable β-cell secretory capacity over 2 to 4 years, including in a patient between 12 and 13 years after TPIAT (151). These results support that islet survival and long-term functional durability may require establishing a functional islet β-cell mass >30% of normal.

Figure 5.

Functional islet β-cell mass estimated from the β-cell secretory capacity. β-Cell secretory capacity is measured as the acute insulin response to glucose-potentiated arginine and shown for insulin-independent autoislet and alloislet recipients as the percentage of nondiabetic control values. Although in both groups a functional islet β-cell mass of ~40% to 50% of normal is achieved, when corrected for the number of islet equivalents transplanted (and assuming 1 million islets are present in the controls), there remains an ~25% loss of islet β-cell mass in the alloislet recipients.

Metabolic testing prior to TPIAT is important to evaluate for pancreatogenic (type 3c) diabetes and assess some measure of functional β-cell capacity for secretion. This is typically done by assessment of HbA1c, fasting glucose, and oral glucose tolerance in response to either a standardized oral glucose tolerance test (OGTT; 75 g of glucose) or mixed-meal tolerance test (6 mL/kg up to 360 mL of Boost^® high protein or equivalent containing ~50 g of carbohydrate) (181). Both 1-hour and 2-hour measures of glucose should be assessed postingestion because an elevated 1-hour OGTT glucose may already indicate a reduced β-cell secretory capacity in patients with pancreatic disease (182). Indeed, not only do standard criteria for prediabetes (HbA1c ≥5.8%, fasting glucose ≥104 mg/dL, 2-hour glucose ≥141 mg/dL) predict the need for insulin therapy after TPIAT, but also a 1-hour glucose ≥189 mg/dL (183). Corresponding C-peptide levels, with or without insulin, should be measured during the OGTT (183) or mixed-meal tolerance test (184), with the combination of fasting glucose <100 mg/dL and stimulated C-peptide >4 ng/mL best predicting the likelihood of an islet isolation yield ≥2500 IE/kg (184). These same measures also appear predictive for estimating islet yield from pediatric patients (185). Nevertheless, these oral challenge techniques and even IV glucose tolerance testing that measures first-phase insulin secretion as a surrogate measure of β-cell secretory capacity (186) lack precision to predict islet isolation outcome for any individual (184). Using arginine stimulation testing in deceased organ donors, an acute insulin response ≥55 µU/mL best predicted the likelihood of an islet isolation yield ≥250,000 IE and/or ≥4000 IE/kg (187). Further study is warranted to determine whether glucose potentiation of arginine induced insulin secretion to provide an estimate of functional β-cell mass prior to TPIAT may better predict islet isolation outcomes and postsurgical diabetes risk.

Insulin sensitivity

Insulin sensitivity has variably been measured as modestly impaired or normal in otherwise normoglycemic patients with chronic pancreatitis (188, 189), a finding likely explained by the presence of a chronic disease studied in various phases of evolution. Importantly, insulin sensitivity as measured by the frequently sampled IV glucose tolerance test increased from before to 12 and 18 months following TPIAT during a randomized clinical trial of the dipeptidyl pepdidase-4 inhibitor sitagliptin vs placebo in both study groups (190). Although this well-conducted study did not find any differences in metabolic outcomes with sitaglitpin use in TPIAT recipients, the normal estimates of insulin sensitivity measured during follow-up suggest that any potential decrease in insulin sensitivity due to chronic pancreatitis resolves postoperatively, assuming at least that near-normal glycemic control is maintained as demonstrated in the referenced study (190).

Glucose counterregulation

Hypoglycemia is reported by some insulin-independent recipients of autoislets that in some cases may be severe (191, 192), and suggests the presence of an inherent defect in glucose counterregulation, that is, the physiologic defense against the development of low blood glucose. Normally in response to decreasing plasma glucose, endogenous insulin secretion is turned off and glucagon secretion is activated with the resulting decreased ratio of insulin to glucagon exposed to the liver increasing endogenous glucose production to prevent or correct the development of hypoglycemia. Secondary to this primary islet response against hypoglycemia, brain recognition of hypoglycemia activates sympathoadrenal responses to release epinephrine that becomes critical for increasing endogenous glucose production when the glucagon response is impaired (193) and also generates autonomic symptoms that alert the individual to ingest food. Thus, defects in both islet and sympathoadrenal responses must interact to contribute to the development of hypoglycemia (Fig. 6).

Figure 6.

Potential mechanisms explaining the risk for hypoglycemia experienced by recipients of intrahepatic autoislets following total pancreatectomy, as well as the amelioration of hypoglycemia experienced by type 1 diabetic recipients of intrahepatic alloislets. [Includes data from Rickels MR, Liu C, Shlansky-Goldberg RD, et al. Improvement in beta-cell secretory capacity after human islet transplantation according to the B7 protocol. Diabetes 2013; 62:2890-2897; and from Robertson RP, Bogachus LD, Oseid E, et al. Assessment of beta-cell mass and alpha- and beta-cell survival and function by arginine stimulation in human autologous islet recipients. Diabetes 2015;64:565–572.]

In 1992 it was first reported that TPIAT recipients had impaired glucagon responses to insulin-induced hypoglycemia, even though they had normal glucagon responses to IV arginine and a liver biopsy demonstrated α-cells containing glucagon-positive granules (194). This led to a study in total pancreatectomized dogs that compared glucagon secretion during insulin-induced hypoglycemia in groups of animals receiving either intrahepatic autoislets or intraperitoneal autoislets (51). The dogs receiving intraperitoneal islets had greater glucagon responses than did the dogs receiving intrahepatic autoislets, although both groups had similarly reduced glucagon responses to IV arginine (51). In another study of total pancreatectomized dogs receiving either intraomental autoislets or intrasplenic autoislets, mild non–insulin-medicated hypoglycemia resulted in normal inhibition of insulin secretion but impaired glucagon responses in both groups of animals, which again displayed similarly reduced glucagon responses to IV arginine (195). More recently, in a cohort of intrahepatic TPIAT recipients routinely studied by the investigators but not selected for having reported hypoglycemia, it was discovered that upon close questioning 11 out of 14 recipients acknowledged episodes of hypoglycemia during daily living that occurred either after meals and/or during exercise (52). Self-monitoring of blood glucose levels in these 11 documented glucose levels that reached nadirs of ~37 mg/dL and continuous glucose monitoring in five of these recipients demonstrated ~15% of time spent with hypoglycemia (<70 mg/dL) (52). In this study, hyperinsulinemic hypoglycemic clamps demonstrated markedly impaired glucagon responses in recipients of intrahepatic islets only compared with normal glucagon responses present in five recipients who had received autoislets in nonhepatic as well as hepatic sites; again, glucagon responses to IV arginine were similarly reduced in all transplant recipients (52).

One hypothesis to explain post-TPIAT hypoglycemia is that intrahepatic islets fail to detect hypoglycemia owing to the locally higher glucose levels present in the liver. In support of this hypothesis, in a rodent model of intrahepatic islet transplantation only after glycogen depletion did an increase in glucagon occur in response to hypoglycemia, suggesting that an increase in endogenous glucose production and consequent intrahepatic levels may inhibit activation of intrahepatic α-cells, as local glucose levels may be higher than the peripherally measured hypoglycemia (196). However, the intrahepatic islet β-cells exposed to the same local glucose levels do appropriately respond to hypoglycemia with complete suppression of insulin secretion (52), evidencing appropriate sensing of peripheral glycemia. Although the evidence supports that nonhepatic islets secrete more glucagon in response to hypoglycemia than do their intrahepatic counterparts, recipients of nonhepatic islets received a larger total number of islets, and those placed in the pelvic portion of the peritoneal cavity revascularize in the distribution of the systemic rather than portal circulation that would bypass first-past hepatic extraction of glucagon leading to higher peripherally measured levels (197). A difference in venous drainage by transplant site would also explain why impaired glucagon responses were seen in dogs with intraomental or intrasplenic autoislets that are both nonhepatic but still portally drained (195), whereas dogs with intaperitoneal autoislets exhibited normal responses (51). Future prospective studies are needed to compare the glucagon response to hypoglycemia in recipients of comparable numbers of islets transplanted in hepatic and nonhepatic sites, as well as the endogenous glucose production response because the peripherally measured glucagon levels are markers of the α-cell response and not direct measures of the counterregulatory response.

Another hypothesis to explain post-TPIAT hypoglycemia is that alimentary hypoglycemia develops as a consequence of the Roux-en-Y gastrointestinal reconstruction that follows pancreaticoduodenectomy and is similar to other well-recognized forms of “late dumping syndrome” where altered nutrient transit with rapid delivery to the distal small intestine leads to more rapid glucose absorption and exaggerated glucagon-like peptide-1 (GLP-1) and insulin responses (198, 199) with dysregulated islet function that in some individuals leads to late postprandial hypoglycemia (198, 200–202). In support of this hypothesis, the hypoglycemia that is associated with TPIAT occurs postprandially and can be managed with widely varying degrees of success by the ingestion of small frequent meals and avoidance of simple sugars (52, 191).

In a mixed-meal study involving 10 TPIAT patients experiencing postprandial hypoglycemia, glucose levels reached a higher early peak than in controls, and then fell to levels below baseline between 4 and 6 hours postingestion that in five patients included a nadir <70 mg/dL not seen in controls (203). When corrected for the numbers of islets transplanted, the insulin and C-peptide secretory responses were greater than normal, whereas the glucagon response was greater early after the meal and lower later postingestion compared with controls (203). These data were interpreted to suggest that a mismatch between insulin levels and glucose levels may occur in the later hours after TPIAT meals such that endogenous insulin levels induce hypoglycemia. In this study glucagon levels did not increase above normal in the recipients who developed hypoglycemia postprandially. Such a defect in glucagon secretion despite lower late postprandial glucose concentrations is also exhibited by patients with other forms of alimentary hypoglycemia who have intact native islets and may be explained by a combination of the paracrine effect of the increased β-cell response on inhibiting α-cell function (204), with an additional contribution from the glucagonostatic effect of exaggerated GLP-1 release (205, 206), and desensitization of central mechanisms for responding to low blood glucose. In fact, nondiabetic individuals studied before and after Roux-en-Y gastrointestinal reconstruction as part of gastric bypass surgery for obesity develop reduced neurohormonal responses to insulin-induced hypoglycemia, including a marked reduction in glucagon and epinephrine secretion (207). Because exposure to even mild hypoglycemia leads to blunting of subsequent sympathoadrenal responses to hypoglycemia, including sympathetic nervous system augmentation of glucagon release as well as activation of epinephrine secretion (208), it is possible that induction of hypoglycemia-associated autonomic failure is operative after TPIAT and contributes to the impaired glucagon responses to hypoglycemia. Additionally, subtle differences in the endogenous glucose production response during exercise in autoislet recipients might contribute to the development of hypoglycemia (209), particularly if conducted late postprandially that would further contribute to the risk for experiencing hypoglycemia.

Islet allotransplantation for type 1 diabetes

β-Cell secretory capacity

The reversal of diabetes by islet allotransplantation is also dependent on the mass of islet β-cells that survives engraftment (159). As already discussed, most recipients with type 1 diabetes receiving a total of >9000 IE/kg from at least two donor pancreases under the Edmonton protocol achieve insulin independence (11); however, by a median 2-year follow-up, most patients return to requiring some insulin therapy despite ongoing evidence of islet graft function indicated by restored C-peptide production (154), suggesting establishment of only a marginal mass of surviving islets posttransplantation. Indeed, in insulin-independent, normoglycemic alloislet recipients treated with the Edmonton protocol at two institutions, the β-cell secretory capacity was only ~25% of normal despite receiving >600,000 to 900,000 IE (210, 211). Using another glucocorticoid-free regimen with islets transplanted from pooled donors to achieve insulin independence, Keymeulen et al. (212) also demonstrated a β-cell secretory capacity of ~25% of normal in alloislet recipients. Collectively, these results indicate a markedly reduced engrafted islet β-cell mass in alloislet recipients that is just at the margin of what is required to avoid hyperglycemia, and so help to explain the eventual return to insulin therapy experienced by most recipients treated under the Edmonton protocol.

The lower functional islet β-cell mass for the numbers transplanted under the Edmonton protocol supports early loss of transplanted islets before engraftment due to nonspecific inflammatory and thrombotic mechanisms (66, 213) that has been estimated by positron emission tomography/CT study to affect ~25% of the islet product (214). Further early islet loss may be attributed to excessive metabolic demand inducing endoplasmic reticulum stress (215). Importantly, new induction immunosuppression protocols introduced by the University of Minnesota and adopted elsewhere incorporated peritransplant anti-inflammatory, antithrombotic, and intensive insulin therapy, with improved rates of insulin independence occurring more frequently with islets isolated from a single donor (18, 124) and being sustained for a longer duration posttransplant (131, 134), suggesting that a greater proportion of the transplanted islets survived engraftment. The B7 protocol was designed to test this approach using the T lymphocyte–depleting polyclonal antibody thymoglobulin together with a period of islet culture and the TNF-α inhibitor etanercept and pentoxifylline to address inflammation, anticoagulation to address thrombosis, and intensive insulin therapy for at least 2 months to minimize metabolic demand during engraftment.

Following the B7 protocol, the resulting β-cell secretory capacity was ~40% to 50% of normal in 11 consecutively treated patients who received >600,000 IE, with all becoming insulin-independent (159) (Fig. 5). Seven of the patients received islets from a single donor pancreas and four received islets from a second donor pancreas, such that when considering the lower numbers of islets transplanted than with the Edmonton protocol represents an approximately threefold improvement in engraftment efficiency with the B7 protocol with repeat measures of the β-cell secretory capacity stable at 1 year (159). These findings suggest that a reserve capacity for insulin secretion more than necessary to achieve initial insulin independence is required to maintain long-term islet graft function. Indeed, during the second year of follow-up only 1 patient of the 11 with the lowest β-cell secretory capacity returned to low-dose insulin therapy to maintain normoglycemia (216). Why more islets appear to survive in successful cases of islet autotransplantation when compared with allotransplantation (Fig. 5) is unknown, but it may be due to the source being a living rather than deceased donor or some tropic factor derived from the nonislet tissue present in higher amounts in less pure preparations. Similar to autoislets, however, these results from alloislets also support that islet survival and long-term functional durability may require establishing a functional islet β-cell mass >30% of normal.

Importantly, the improved results for engrafted islet β-cell mass with islet allotransplantation were obtained using the same low-dose tacrolimus and sirolimus maintenance immunosuppression as in the Edmonton protocol. Although both tacrolimus and sirolimus have known human islet toxicity at levels supratherapeutic for their desired immunosuppressive effects (217, 218), β-cell secretory capacity is fully normal in uncomplicated whole-pancreas transplant recipients who have received 100% of an islet β-cell mass despite tacrolimus-based immunosuppression (219), and studies using sirolimus at plasma drug concentrations targeted in clinical practice have no deleterious effect on human islets (220). Curiously, in a canine autoislet transplant model, treatment with sirolimus was associated with a 13% reduction in insulin clearance without any change in insulin sensitivity (221), and which might exert a modest benefit in terms of extending the metabolic effect of secreted insulin. These findings support that modern dosing of currently available immunotherapy is not toxic to islets nor rate limiting for improving transplanted islet graft survival and functional durability.

Insulin sensitivity

Insulin sensitivity is impaired in type 1 diabetes (222), a consequence of absent endogenous insulin secretion, exposure to frequent periods of sustained hyperglycemia (223), and impaired sensitivity of lipolysis to inhibition by insulin with accompanying increases in free fatty acids (224, 225) that affects insulin action at both the liver and skeletal muscle (226–228). Intraheptic islet transplantation restores endogenous insulin secretion that not only corrects the hyperglycemia of type 1 diabetes, but also normalizes free fatty acid metabolism that is associated with improvement in insulin sensitivity (229–231). Interestingly, current glucocorticoid-free immunosuppression with tacrolimus and sirolimus has been implicated in the development of insulin resistance in rodents (232, 233), whereas in both a canine islet autotransplant model with or without the calcineurin inhibitor with cyclosporine (221) and a nonhuman primate islet allotransplant model (234) treatment with sirolimus did not impair insulin sensitivity. Moreover, in humans, appropriate dosing of sirolimus for its immunosuppressive effects in nondiabetic subjects demonstrated no effect on hepatic insulin sensitivity and actually enhanced peripheral insulin sensitivity during nutrient stimulation with amino acids (235). In longitudinal assessment of the same 12 individuals with type 1 diabetes from before to after islet allotransplantation, measures of both hepatic and peripheral insulin sensitivity improved to estimates not different than normal (236, 237). Thus, intrahepatic islet transplantation corrects the defects in insulin action present at both the liver and skeletal muscle in type 1 diabetes, and modern dosing of glucocorticoid-free immunosuppression with low-dose tacrolimus and sirolimus does not induce insulin resistance in this population.

Glucose counterregulation

In contrast to reports of hypoglycemia following islet autotransplantation, type 1 diabetic recipients of islet allotransplantation are entirely protected from hypoglycemia for the duration of islet graft function, even when requiring low-dose insulin therapy to maintain normoglycemia (Fig. 6). With the autoimmune destruction of islet β-cells in type 1 diabetes, the near-total loss of endogenous insulin secretion not only eliminates any autoregulatory capability to reduce excessive insulin, but also abolishes the islet α-cell glucagon response to hypoglycemia that requires paracrine activation from neighboring β-cells through an intraislet decrement in insulin secretion (238–240). In type 1 diabetic recipients of intrahepatic islet transplants, there is recovery of the physiologic islet cell responses to insulin-induced hypoglycemia whereby endogenous insulin secretion is appropriately suppressed and glucagon secretion is partially restored (241–243). In earlier cross-sectional studies involving subjects treated according to the Edmonton protocol, the glucagon response to insulin-induced hypoglycemia was either reported as being absent in one study (244) or, when compared with euglycemic control experiments [that control for the inhibitory effect of experimental hyperinsulinemia on glucagon secretion (245)], was reported as partial albeit markedly less than normal (241, 243). Subsequent longitudinal study of the same 12 patients investigated before and 6 months after islet allotransplantation (whereby 11 were treated according to the B7 protocol) demonstrated a significant improvement in the glucagon response to ~40% to 50% of normal that was associated with complete restoration of the endogenous glucose production response to insulin-induced hypoglycemia (242), an effect that remained upon repeat assessment at 18 months (216). The magnitude of the glucagon response was similar to the ~40% to 50% of a normal β-cell secretory capacity reported from this cohort of B7 treated patients (159). These results are consistent with the partial glucagon responses to insulin-induced hypoglycemia observed in type 1 diabetic recipients of pancreatic segments (246), and they support the dependence of the glucagon response on the number of transplanted islets surviving engraftment. It remains possible that the release of glucagon from islets within the liver may create an early increase in endogenous glucose production with subsequent attenuation of the stimulated glucagon levels measured peripherally by locally increased glucose. However, the normal suppression of C-peptide levels during hypoglycemia indicates that intrahepatic islet β-cells do appropriately sense and respond to the degree of peripheral hypoglycemia (241, 242), arguing against islet exposure to locally increased glucose.

“…type 1 diabetic recipients of islet allotransplantation are entirely protected from hypoglycemia for the duration of islet graft function.”

In the absence of intact islet cell responses to hypoglycemia in type 1 diabetes, secondary sympathoadrenal (epinephrine secretion and autonomic symptom generation) responses become critical to increase endogenous glucose production from the liver and alert the individual to ingest food to correct low blood glucose, whereas pituitary–adrenal (GH and cortisol secretion) responses generally operate later in an attempt to maintain increased hepatic glucose production (247). However, both the glycemic threshold (i.e., the glucose level that elicits the response) and magnitude of these hormonal and symptom responses are impaired by recurrent episodes of hypoglycemia that induce brain adaptation to low blood glucose and lead to a syndrome of hypoglycemia-associated autonomic failure, also known as hypoglycemia unawareness (248), that increases the risk of life-threatening hypoglycemia 20-fold in type 1 diabetes (249). By 6 months following islet allotransplantation, there is normalization of the glycemic thresholds for activation of epinephrine, GH, and autonomic symptom responses to insulin-induced hypoglycemia (250), evidencing recovery from hypoglycemia-associated autonomic failure. Although the magnitude of the epinephrine response may be only partially improved after 6 months (242, 250), additional longitudinal follow-up of the same alloislet recipients for 2 years demonstrated essentially no time with hypoglycemia (<70 mg/dL) by continuous glucose monitoring, and after 18 months complete normalization of the epinephrine response to insulin-induced hypoglycemia (216). These findings indicate that >6 months of strict avoidance of hypoglycemia may be required for complete autonomic nervous system function recovery in patients with hypoglycemia unawareness.

Importantly, even in alloislet recipients requiring insulin to support partial graft function there is modest recovery of glucagon and epinephrine secretion with partial restoration of the endogenous glucose production response to insulin-induced hypoglycemia that is associated with clinical protection from hypoglycemia (243). Continuous glucose monitoring studies have demonstrated similar reductions in mean glucose, glucose variability, and serious hypoglycemia (<54 to 60 mg/dL) relative to that in type 1 diabetes for both insulin-independent and insulin-requiring alloislet recipients (251–253). That the islet graft is responsible for these improvements in continuous glucose monitoring metrics of glycemic control is supported by their significant correlation with measures of islet β-cell graft function (254, 255). Thus, even partial improvements in glucose counterregulation and hypoglycemia symptom recognition likely work in concert with reduced glucose variability to provide the robust protection from hypoglycemia afforded by islet transplantation. Such avoidance of hypoglycemia is distinct from the results described above following TPIAT, and from outcomes seen with nontransplant intervention for type 1 diabetes complicated by hypoglycemia unawareness (256).

Secondary complications of diabetes

The effect of islet allotransplantation on secondary complications of diabetes has been assessed by a few nonrandomized studies. Thompson and colleagues (257) conducted a prospective cohort study of islet allotransplantation in 26 nonuremic patients with type 1 diabetes vs prior or ongoing intensive medical therapy focused on improving glycemic, blood pressure (including use of angiotensin system blockade), and serum lipid control in another 18 waitlisted patients. Baseline HbA1c was ~8.0% in both groups, and after a median ~36-month follow-up was significantly lower at 6.6% following islet transplantation vs 7.4% with intensive medical therapy (258) whereas blood pressure and lipids were comparable (257). The primary outcome of change in measured glomerular filtration rate (GFR) was less after islet transplantation, being −1.4 mL/min/y/1.73 m² vs −3.6 mL/min/y/1.73 m² following institution of intensive medical therapy; there were no between group differences in urinary albumin excretion or progression over time (259). Retinopathy graded as greater than mild nonproliferative progressed in 0 of 51 eyes following islet transplantation and more often in 10 of 82 eyes during intensive medical therapy; nerve conduction studies remained stable in both groups (258). In a single cohort followed for 5 years following islet allotransplantation, sensory, but not motor, nerve conduction studies improved over time but differences were not evident until after 3 years (260). These findings support a possible effect of islet allotransplantation on slowing the progression of diabetes microvascular complications, and they suggest that future randomized studies will require >3 years of follow-up to confirm differences from intensive medical therapy.

Mechanisms of Islet Graft Dysfunction and Failure

Metabolic exhaustion: endoplasmic reticulum stress, oxidative stress, islet amyloid

Since the first experiments of islet transplantation using animal models, a nonimmunologic mechanism of islet “stress” has been speculated to account for cases of recurrent hyperglycemia in the absence of evidence of immune system activation (2). As discussed for both autoislet and alloislet recipients, a reserve capacity for insulin secretion more than necessary to achieve initial insulin independence (>30% of normal β-cell secretory capacity) is required to maintain long-term islet graft function. This reserve capacity likely protects islets from excessive metabolic demand inducing endoplasmic reticulum stress (215), oxidative stress (261), or amyloid deposition (262), as has been described in preclinical islet transplant models and that may be toxic to islets. Excessive β-cell demand for secretion may be indicated by increased secretion of proinsulin relative to insulin, resulting in an elevated molar ratio of proinsulin to insulin, which occurs with recruitment of immature secretory granules containing an abundance of incompletely processed proinsulin (263). Hyperglycemia per se decreases β-cell insulin content and increases β-cell secretion of proinsulin from isolated human islets (264). Autoislet recipients with hyperglycemia similarly demonstrate decreased fasting insulin and C-peptide with elevations of proinsulin and the proinsulin-to-insulin ratio (265). Importantly, in studies involving insulin-independent and insulin-dependent alloislet recipients where the mean HbA1c was ~6.0%, fasting proinsulin-to-insulin and proinsulin secretory ratios in response to glucose-potentiated arginine were normal (219, 266). In contrast, in a study involving autoislet and alloislet recipients where the mean HbA1c was ~7.1%, fasting proinsulin-to-insulin ratios were elevated (267), consistent with an effect of exposure to chronic hyperglycemia. Thus, the maintenance of near-normal glycemia (HbA1c ≤6.5%) with insulin therapy when appropriate may alleviate excessive metabolic demand on the islet graft as indicated by protection against the secretion of incompletely processed proinsulin, and prevent stress-induced exhaustion of islet graft function.

Additional evidence for the importance of avoiding hyperglycemia in islet recipients comes from the finding of amyloid deposition in intrahepatic islets of three out of four cases with postmortem examination ~2 to 5 years after transplant (268, 269). Islet amyloid is composed of islet amyloid polypeptide (IAPP or amylin) fibrils deposited within and surrounding β-cells where they exhibit direct toxicity (270). IAPP is cosecreted from the β-cell with insulin (271), but normally is inhibited from forming amyloid by appropriate proportions of insulin and other factors in the β-cell (272). During glucose-potentiated arginine testing in insulin-independent alloislet recipients there is disproportionately increased IAPP-to-insulin secretion (273), suggesting that hyperglycemia may disturb regulation of insulin and IAPP within transplanted islets and facilitate the development of islet amyloid. Moreover, incompletely processed pro-IAPP may be more amyloidogenic than IAPP (274), and hyperglycemic islet transplant recipients with disproportionately increased fasting proinsulin also exhibit disproportionately increased fasting levels of the pro-IAPP_1–48 intermediary (275). In a nonhuman primate transplant model, intrahepatic islets developed amyloid deposits coincident with animal growth, increasing β-cell secretory demand with the presence of amyloid predating a decline in β-cell secretory capacity and subsequent recurrence of hyperglycemia (262). Thus, avoiding hyperglycemia and reducing metabolic demand through the provision of exogenous insulin when necessary should prevent nonimmunologic islet graft loss and improve long-term functional outcomes for islet recipients (276).

Insulin therapy should also be provided to islet recipients during periods of intercurrent illness that may increase metabolic demand for insulin, as well as in advance of the anticipated development of insulin resistance as occurs with glucocorticoid therapy and pregnancy. Indeed, the administration of glucocorticoids for indications not related to the islet transplant has been shown to result in severe hyperglycemia and consequently islet graft failure in previously insulin-independent autoislet recipients (277). In addition to the marked induction of insulin resistance by glucocorticoids increasing demand for insulin secretion from a reduced functional β-cell mass, the hyperglycemia resulting from an inadequate hyperinsulinemic response occurs together with glucocorticoid-induced elevation of free fatty acids that accelerate β-cell damage through glucolipotoxicity (278). In a canine model of intrahepatic islet autotransplantation, a short course of the glucocorticoid prednisone induced hyperglycemia and accelerated the time to islet graft failure, but importantly these effects were prevented by the administration of insulin for the duration of glucocorticoid exposure (279). Pregnancy increases metabolic demands for insulin through both the development of maternal insulin resistance and the requirement for increased nutritional intake to support maternal and fetal growth. Few reports are available but they support a benefit to preemptive initiation of insulin at the time pregnancy is detected to maintain gestational targets for fasting and postprandial glucose that can result in optimal birth outcomes and prevent deterioration of islet graft function with a return to postgestational insulin independence (280–284). These observations support that appropriately timed administration of exogenous insulin to alleviate transplanted β-cells from increased requirements for secretion during periods of excessive metabolic demand while avoiding glucotoxicity can prevent stress-induced islet graft dysfunction and loss, and it is important to sustaining long-term islet graft function for transplant recipients.

“…reducing metabolic demand through the provision of exogenous insulin when necessary should prevent nonimmunologic islet graft loss…”

Another strategy for alloislet recipients in addition to the provision of insulin therapy to protect marginal islet grafts from further functional deterioration is administering a supplemental islet infusion. Similar to infusing islets from a second or sometimes a third donor pancreas to achieve insulin independence that is associated with more durable long-term islet graft function (69), patients who later return to a requirement for insulin therapy may be considered for a supplemental islet infusion to reestablish a reserve capacity for insulin secretion with the goal of restoring insulin independence and further promoting long-term islet graft function (161, 258). For this strategy to succeed it is important that immunologic mechanisms are not responsible for the islet graft dysfunction. The reported studies to date have administered supplemental islet infusions with adjuvant treatment including the GLP-1 agonist exenatide based on evidence that GLP-1 has antiapoptotic effects on cultured human islets (285, 286). Another study administered exenatide at the time of initial islet infusions that resulted in early insulin independence (287), but subsequently supplemental islet infusions were required to maintain more durable freedom from exogenous insulin (288). Future randomized controlled trials of adjuvant treatment with a GLP-1 agonist at the time of an initial or supplemental islet infusion on surviving functional islet graft mass are needed.

In other studies patients with islet graft dysfunction were treated with exenatide that was associated with a reduction in insulin requirements without any improvement in glycemic control during a 3- to 6-month period (289, 290); however, by 12 months most patients required an increase in exogenous insulin or a supplemental islet infusion to maintain glycemic control (291). More islet recipients experienced nausea and vomiting with exenatide than that reported in studies of patients with type 2 diabetes, and the tolerated doses were lower. In islet recipients GLP-1 acutely increases the proinsulin secretory ratio in response to glucose-potentiated arginine (292), and chronic exenatide leads to disproportionate increases in proinsulin and IAPP secretion relative to insulin following mixed-meal stimulation (291), effects that suggest GLP-1 agonists increase metabolic demand on a marginal islet β-cell mass that may be detrimental compared with use of exogenous insulin on maintaining long-term islet graft function. Similar ineffectiveness has been reported with the dipeptidyl peptidase-4 inhibitor sitagliptin that slows degradation of the endogenous incretin hormones GLP-1 and glucose-dependent insulinotropic polypeptide, where a 6-month treatment in alloislet recipients was associated with a reduction in insulin requirements without any improvement in glycemic control (293). Consistent with these results, a randomized controlled trial of a 12-month treatment with sitagliptin in autoislet recipients did not affect metabolic outcomes (190).

Alloimmune rejection and sensitization

As for all allogeneic forms of transplantation, alloislets are susceptible to both acute and chronic forms of rejection that may be associated with early and late loss of islet graft function. Alloislets are typically not matched to the recipient for HLAs, and with the possibility for the transplantation of islets from more than one donor there is potential for both a larger number of mismatched HLAs as well as repeat HLA mismatches. Pretransplant sensitization to islet donor HLAs from previous exposure to nonself HLAs through pregnancy, red blood cell transfusions, solid organ transplantation, or human tissue grafts can result in acute rejection of an islet transplant (294, 295). Compared with solid organ transplants, isolated islets are more susceptible to complete loss following acute rejection, likely due to their presentation to the recipient’s immune system as a dispersed cell suspension (296). Because there is no established treatment available for acute rejection of an islet transplant, the transplantation of alloislets bearing HLAs reactive against even very low levels of preformed alloantibody in the recipient should be avoided even when T and B lymphocyte crossmatches are negative. In fact, even alloislet recipients previously sensitized to HLAs not present on an islet graft also experience a higher rate of graft loss that may be explained by cross-reactive or epitope-spreading alloantibody (297, 298). Acute cellular rejection may also develop in the absence of alloantibody, and in one case diagnosed clinically 1 month after transplantation, a 6-day course of high-dose glucocorticoid administered under islet protection with a continuous IV insulin infusion rescued islet graft function (299).

Chronic rejection may also occur with the development of de novo alloantibodies specific for donor HLA that has been temporally associated with loss of islet graft function and β-cell secretory capacity (300, 301). Islet graft loss appears more often with the development of de novo alloantibodies against the islet graft occurring in the context of a reduction in immunosuppression (302) rather than when alloantibodies are detected under adequate immunosuppression (157, 303). Although the identification of anti-islet donor HLA antibodies is a negative prognostic marker for future islet functional outcomes (304), whether early detection and subsequent intensification of immunosuppression may prolong islet graft function and survival is not known. In one case where the development of anti-islet donor HLA antibodies coincided with a deterioration in islet graft function 2 years after transplantation, treatment of clinically diagnosed humoral rejection with rituximab and IV immunoglobulins led to improvement in glycemic control and reestablishment of insulin independence (305).

Although sensitization to HLAs with the development of anti-islet donor alloantibodies has been reported in up to 20% of islet recipients receiving ongoing immunosuppressive drug treatment, the risk of sensitization increases with withdrawal of immunosuppression such that most patients discontinuing immunosuppression develop anti-HLA antibodies when assessed by sensitive assays based on flow cytometry methods (156, 157, 306). In islet recipients undergoing immunosuppression withdrawal, increased cytotoxic T lymphocyte avidity for mismatched HLAs may be seen even in the absence of alloantibody (307). In single-center analyses, the risk of sensitization has not been associated with the number of islet donors or number of mismatched HLAs between the islet recipient and donors (156, 157, 306); however, in an analysis conducted by the Collaborative Islet Transplant Registry there was a trend toward lower risk for sensitization when subsequent islet infusions contained repeat HLA class I mismatches (308). Until the consequences of sensitization on outcomes of potentially needed future organ transplants are known, it reasons that both minimizing the number of islet donors used per recipient and, in the absence of donor-specific anti-HLA antibodies (and ideally cytotoxic T lymphocyte responses), repeating HLA mismatches with subsequent islet infusions may decrease the risk of sensitization and requires further investigation.

Autoimmune recurrence in type 1 diabetes

Unique to islet allotransplantation for type 1 diabetes is the possibility for a recurrence of autoimmunity affecting the transplanted islet β-cells that can also result in early or late loss of islet graft function (309). Recurrent autoimmune diabetes was first demonstrated following initially successful transplantation of hemipancreases donated by identical twins that following reversal of diabetes underwent selective β-cell destruction by the twin recipients with type 1 diabetes, thereby recapitulating the development of the autoimmune disease (310). Although standard immunosuppressive therapy for the most part prevents autoimmune recurrence, there are well-documented cases of recurrent autoimmune diabetes affecting standard deceased donor whole-pancreas transplant recipients (311, 312). Patients with type 1 diabetes exhibiting elevated titers of autoantibodies and cellular autoimmunity prior to islet transplantation appear at increased risk for failure to achieve insulin independence and subsequent islet graft loss (313–316). That recurrent autoimmune diabetes can manifest in an islet recipient has been documented in biopsy studies where insulitis containing lymphocytic islet infiltration is seen with the absence of insulin-positive β-cells and presence of glucagon-positive α-cells (317, 318). Such islet graft loss may be heralded by either an increase in titer of an existing autoantibody or the development of a new autoantibody or cellular autoimmune response after transplant (304, 314, 319, 320). Unfortunately, no strategy for intensification of immunosuppression has been tested to determine whether islet graft function and survival may be prolonged after the identification of new autoimmune markers. Nevertheless, current immunosuppression for islet transplantation, as for pancreas transplantation, does abrogate autoimmunity for the vast majority of recipients with type 1 diabetes. Minimization of long-term immunosuppression that may be acceptable without increasing the risk for alloimmune rejection may nonetheless allow for recurrent autoimmunity resulting in chronic islet or pancreas graft loss. Interestingly, increases in both autoantibody and cellular autoimmune responses against type 1 diabetes–associated antigens glutamic acid decarboxylase-65 and islet antigen-2 have been observed during the course of developing de novo alloimmune responses in islet transplant recipients (298, 307), which might accelerate immune-mediated destruction of islet β-cells in the context of alloimmune recognition in patients with type 1 diabetes.

Another possible potentiating factor for the development of recurrent autoimmune diabetes in islet transplant recipients is active CMV infection. Although occurring rarely following islet transplantation as discussed below, CMV infection has been associated with a trend toward shorter islet graft survival (321). Although CMV may directly infect allograft tissues and increase the risk for allograft rejection (322), CMV infection has also been associated with the development of type 1 diabetes (323) and with the recurrence of autoimmune diabetes following pancreas transplantation (324). Autoreactive T cell clones from individuals with type 1 diabetes have been shown to recognize CMV peptide, supporting molecular mimicry as a potential mechanism for the induction of β-cell autoimmunity by CMV (325). Future investigation is required to assess this potential mechanism for islet graft loss and to gain insight into the mechanisms leading to the initial development of type 1 diabetes that may be recapitulated in the context of islet allotransplantation.

Procedural Risks and Complications of Islet Transplantation

Islet autotransplantation

Procedural risks

As already discussed, the main procedural risks of islet autotransplantation at the time of total pancreatectomy are portal vein thrombosis, surgical site bleeding that may be exacerbated by the heparinization required to minimize the risk for portal vein thrombosis, and the introduction of pathogens from a contaminated islet preparation. Minimization of tissue volume through purification of the islet preparation, monitoring of portal venous pressures during islet infusion, and placement of a portion of the islet product in the peritoneal cavity when elevated portal venous pressure is encountered can all reduce the risk of portal vein thrombosis (44). However, because autoislet preparations are kept relatively impure to optimize islet recovery, the tissue volume is higher for the number of islets transplanted and results in greater increases in portal venous pressure with autoislet compared with alloislet transplantation (326), and the risk for portal branch vein thrombosis may be estimated at ~10% (44). When a portal branch vein clot remains present by hospital discharge, long-term anticoagulation is required for 3 to 6 months to reestablish patency of the portal tree and prevent clot propagation. Independent of the potential effects on the portal vasculature, islet engraftment in the liver can result in periportal steatosis due to locally elevated insulin concentrations surrounding the intrahepatic islets that appears as heterogeneous echogenicity when assessed by ultrasound (327); such changes in radiologic liver appearance have not been associated with any adverse consequences for liver or islet graft function.

“Unique to islet allotransplantation for type 1 diabetes is the possibility for a recurrence of autoimmunity…”

Most patients considering TPIAT have undergone endoscopic or surgical interventions with the potential for introducing enteric bacteria into the pancreatic ductal system, which can lead to colonization of the pancreas with microorganisms. Risk for experiencing an infectious complication is minimized by avoiding islet culture and ensuring that perioperative antibiotics cover upper gastrointestinal tract flora while awaiting the results of microbial cultures taken from the transplanted islet product. Although positive microbial cultures may be seen in most cases, the risk of a related infectious complication is reduced to <15% with perioperative antibiotic prophylaxis covering both Gram-positive cocci and enteric Gram-negative bacilli (328–330). Additional infectious complications with pathogens not related to the islet product also occur as expected during recovery from pancreatectomy surgery.

Islet allotransplantation

Procedural risks

The procedural risks for islet allotransplantation are similar to those for islet autotransplantation, although they generally occur at lower frequency owing to the higher purity of the islet product, less invasive procedure for islet delivery, and use of pancreata that have not been previously instrumented. As with intraportal delivery of autoislets, portal pressure changes are related to the transplant tissue volume, but with the higher level of purification the complete islet preparation can invariably be accommodated within the portal circulation under heparinization with the risk for portal branch vein thrombosis <5% (77, 331); maintenance of therapeutic heparinization for the first 48 hours after infusion reduces this risk further (332). In addition to heparinization, use of the bag method for islet infusion instead of a syringe also helps to decrease the incidence of portal branch vein thrombosis (333). With the closed gravity-fed bag system the rate of islet infusion is controlled with natural feedback decreasing the rate with any increase in portal pressure, and so prevents the precipitous development of portal hypertension (333). While baseline portal pressures remain normal after two subsequent islet infusions, a greater increase in portal pressure during subsequent transplants has been reported by some (77) but not all (126) series. Thus, although up to three islet preparations may be safely accommodated by the liver over time (77, 126), the risk for developing portal hypertension or branch vein thrombosis may also increase over time, and this suggests that consideration should be made that even greater purity and lower tissue volume preparations be used with subsequent islet infusions.

Bleeding complications are primarily associated with the percutaneous transhepatic approach to access the portal venous circulation for islet infusion with a risk <10% (331). The primary types of bleeding complications include perihepatic hematomas, which are usually self-limited, and intraabdominal hemorrhage that usually requires transfusion of packed red blood cells and laparoscopy or laparotomy to ensure achievement of hemostasis and evacuate intraabdominal blood. Rarely, injury to a hepatic artery may result in formation of an arteriovenous fistula or arterial pseudoaneurysm that may rupture and necessitate embolization of the feeding artery for bleeding control (334, 335). Other rare complications can include injury to a subcostal blood vessel or puncture of the gallbladder. The use of ultrasound guidance may reduce the number of needle passes required to gain access to the portal vein, helping to minimize the risk for procedural complications. The use of a Gelfoam gelatin sponge as a hemostatic plug instilled in the hepatic catheter track helps to reduce (70), but does not eliminate (71), the risk of experiencing a bleeding complication. Other hemostatic agents may be more effective provided that ablation of the entire intraparenchymal catheter tract is achieved (336, 337). Initial results using Avitene microfibrillar collagen paste suggest that the bleeding risk can be eliminated entirely (338). A very low risk for bleeding may also be achieved using a minilaparotomy approach with direct visualization of the mesenteric vein used for islet infusion to the portal circulation (72); however, this transmesenteric approach introduces risk for other common surgical complications and may require a longer hospitalization (339).

Elevated liver enzymes are commonly encountered during the first several days following intrahepatic islet transplantation with a pattern of transaminases that may reach twofold to fivefold the upper limit of normal, is related to the peri-infusion change in portal venous pressure, and typically peaks after a few days and declines by the second week after infusion (125, 126). This transient response to presumed hepatic ischemia or nonspecific inflammatory injury is not accentuated by subsequent islet infusions, in fact appears greater with initial transplants, and is not associated with any long-term sequelae. Similar to that reported for intrahepatic autoislets, alloislet engraftment in the liver can result in periportal steatosis due to locally elevated insulin concentrations surrounding the intrahepatic islets (340) that appears as heterogeneous echogenicity when assessed by ultrasound (341); such changes in radiologic liver appearance have not been associated with any adverse consequences for liver or islet graft function, although they could indicate islets that are under increased demand for insulin secretion that warrants further investigation.

Immunosuppressive drug risk and toxicity

The requirement for immunosuppressive drug treatment to protect alloislets from alloimmune rejection and recurrent autoimmunity introduces additional risks that may be specific to the individual immunosuppressive drug agents as well as related more generally to immunosuppression. As discussed earlier, allergy or intolerance to one immunosuppressive drug agent can most often be addressed by substitution with an alternative agent from the same or similar drug class. Important to all immunosuppressive drug approaches is the risk for certain immunosuppression-related infections and malignancies (342). Antimicrobial prophylaxis is required against P.jiroveci pneumonia and against reactivation of prior infection from members of the herpes virus family (herpes simplex virus, varicella zoster virus, and CMV, the latter when either the donor or recipient evidences prior exposure by positive CMV antibodies). Transmission of CMV from seropositive donors to seronegative recipients is observed less frequently with alloislets than that reported with solid organ grafts due to the low tissue volume, absence of lymphoid tissue, and low number of passenger leukocytes seen in islet preparations. Nevertheless, CMV viremia has been detected in ∼5% of alloislet recipients despite receiving initial prophylaxis with valganciclovir and can occur late after transplantation with more risk seen in individuals treated with T lymphocyte–depleting compared with IL-2 receptor–blocking antibodies (321, 343). Although symptomatic disease remains rare, monitoring for viremia should extend beyond the period of prophylaxis and any time symptoms such as unexplained fever, diarrhea, myelosuppression, or abnormal liver function tests present that might be compatible with CMV infection. Posttransplantation lymphoproliferative disorders are associated with Epstein-Barr virus (EBV) infection developing under immunosuppression; risk is markedly reduced (<1%) for individuals with evidence of immunity against EBV prior to transplantation (344). The first two cases of posttransplantation lymphoproliferative disorders in islet transplant recipients were recently reported, having presented at 24 and 80 months after transplant (345). Risk for the development of skin cancers, predominantly squamous-cell carcinomas, is increased during exposure to chronic immunosuppression and necessitates sun protection and regular dermatologic monitoring for early detection and ablation of premalignant lesions (346).

An important concern related to immunosuppressive drug toxicity is the risk for impairment in kidney function that has been associated with both calcineurin inhibitors and mTOR inhibitors. When alloislets are transplanted in patients with type 1 diabetes, the use of low-dose tacrolimus in combination with sirolimus is associated with decline in estimated GFR of ~5 mL/min/y/1.73 m² (155, 347). Importantly, however, the change in kidney function is very heterogeneous with some patients experiencing steeper declines and others improvement in measures of GFR (347). Early reductions in GFR may be related to normalization of glycemia with associated alleviation of glomerular hyperfiltration and the renovascular constricting effects of calcineurin inhibitors that is dose-dependent (348). Some patients taking tacrolimus in combination with sirolimus develop albuminuria (~5%) that may be associated with more accelerated reduction in kidney function, but fortunately is reversible upon discontinuation of the sirolimus and replacement with mycophenolic acid (349). Use of tacrolimus in combination with mycophenolic acid following alloislet transplantation has been associated with a lower decline in measured GFR of 1.4 mL/min/y/1.73 m² that is also less than the 3.6 mL/min/y/1.73 m² observed in a control group receiving intensive medical therapy with no between-group differences seen in urinary albumin excretion or progression (259). Thus, careful selection of patients for normal kidney function prior to initiation of immunosuppression and careful monitoring of both kidney function and albuminuria while receiving immunosuppression are critical to appropriately tailor immunosuppressive drug choices and targets to minimize the potential for adverse renal effects (350). Similar consideration is also important for islet-after-kidney transplantation, where there is a similar need to balance immunosuppressive protection of both the islet and kidney grafts with the potential nephrotoxicity of the required immunosuppression that is nevertheless directed by that required for the kidney rather than the islet transplant. More research is required to determine the relative benefit of long-term improvement in glycemic control vs harm of immunosuppressive drug therapy on kidney function in patients with type 1 diabetes.

“The requirement for immunosuppressive drug treatment to protect alloislets from alloimmune rejection and recurrent autoimmunity introduces additional risks…”

Future Developments in Clinical Islet Transplantation

Alternative sources of islet tissue

Porcine islets

The number of islets available for transplantation is a major limitation for both autoislet and alloislet approaches to β-cell replacement therapy. Therefore, the establishment of an unlimited source of islet tissue for transplantation has been a long-sought-after goal. One approach has been the development of xenoislets for transplantation that is most advanced for porcine islets. Porcine islets have the advantages of sourcing from established pathogen-free herds, targeting similar normoglycemia as present in humans, and physiologic capability to handle large demands for insulin secretion. Importantly, porcine IAPP contains amino acid substitutions in the region corresponding to residues 20 to 29 that prevent the formation of fibrils (351, 352). Intrahepatically transplanted porcine islets in an autograft model maintained normoglycemia in young pigs that experienced a 50% increase in body weight during an 18-month period without the development islet amyloidosis (353), and so porcine xenoislets may be less susceptible to metabolic exhaustion of an established transplanted islet β-cell mass over time. Disadvantages include the larger immunologic barrier of xenogeneic than allogeneic tissue that presents an additional risk for hyperacute rejection and requires more intensive immunosuppression (354, 355) and theoretical concern over the potential for transmission of zoonotic infections such as porcine endogenous retroviruses. Novel genome editing approaches may enable the breeding of pigs with significantly lower immunogenicity (356) and retroviral burden (357) that ideally would enable β-cell replacement therapy with less immunosuppression than is presently required for alloislets. Should this promise be realized, there is additional hope that porcine islets would also escape recognition by memory T lymphocytes and evade autoimmune recurrence.

As with humans, intraportal delivery for intrahepatic engraftment has had the most success with diabetes reversal by porcine islets in streptozotocin-induced diabetic nonhuman primates (358). Although genetically engineered porcine islets continue to undergo preclinical evaluation using this model, clinical phase 1/2 studies of alginate microencapsulated porcine islets (Diabecell) transplanted into the peritoneal cavity of patients with type 1 diabetes have been completed in New Zealand (NCT00940173) and Argentina (NCT01739829) (ClinicalTrials.gov; accessed 7 May 2018). Safety data using these nongenetically engineered porcine islets have not detected transmission of either porcine endogenous retroviruses or other porcine microorganisms (359, 360), although efficacy outcomes have not been reported. Whether microencapsulated islets transplanted in sites other than the liver can physiologically regulate glucose homeostasis as effectively as revascularized intrahepatic islets remains to be determined (361).

Stem cell–derived islets

Recent progress in the generation of functional islet β-cells from human stem cell sources has raised hope for the establishment of another unlimited source of islet tissue for transplantation. One approach involves the differentiation of human embryonic stem cells (hESCs) to a pancreatic progenitor stage in vitro that has the potential to further differentiate into functional pancreatic islets in vivo capable of reversing streptozotocin-induced diabetes in immunodeficient mouse models (362, 363). However, this process was also accompanied by the sporadic growth of mesodermal cells reminiscent of the formation of teratomas, and it highlighted the potential risk of transplanting immature stem cell–derived tissue. Such progenitor tissue has also been shown to reverse streptozotocin-induced diabetes in immunodeficient mice when transplanted under the skin using a macroencapsulation device (TheraCyte) (364). Use of a macroencapsulation device allows for removal of the stem cell–derived islet graft for safety monitoring or for cause assessment, and it may allow for immunoisolation of what is essentially an allogenic graft when embryonic stem cells are used (365). Although untested in a large animal model, a combination hESC-derived pancreatic progenitor in a macroencapsulation devise (VC-02) is currently under clinical phase 1/2 evaluation in patients with type 1 diabetes in the United States (NCT03163511, ClinicalTrials.gov; accessed 7 May 2018). Because mature human alloislets survive and function poorly in encapsulation devices (366), these early phase clinical studies should inform whether stem cell–derived islets may be more resistant to hypoxia than isolated human islets or may require alternative encapsulation approaches incorporating oxygen delivery (367).

Another approach to stem cell–derived islets involves the differentiation of hESCs to a pancreatic islet stage in vitro that has the capacity for glucose-dependent insulin secretion before transplantation and can reverse (368) or ameliorate (369) diabetes in immunodeficient mouse models. These stem cell–derived islets are again allogeneic and so will require the development of compatible immunosuppressive or immunoisolation (370) approaches before translation to early phase clinical trials. Stem cell–derived islets have also been generated from patients with type 1 diabetes using inducible pluripotent stem cells (371) and somatic cell nuclear transfer (372). Although the generation of cell therapy using a patient’s own genetic material will avoid alloimmunity upon transplantation, and so may be most attractive for patients with pancreatogenic forms of type 3c diabetes, recurrent autoimmunity will remain a barrier for patients with type 1 diabetes. As these cell products are considered for testing in early phase clinical trials, additional work is also required for understanding their physiology in comparison with isolated human islets as both an in vitro (373) and in vivo reference standard.

Novel approaches to combat engraftment-limiting inflammation and coagulation

As discussed, nonspecific inflammatory and thrombotic mechanisms are activated upon transplanted islet exposure to blood in the portal circulation that effect early islet survival and the efficiency of islet engraftment in the liver. Although the adoption of a number of approaches, including islet culture, peritransplant anti-inflammatory therapy with pentoxifylline and TNF-α inhibitors (e.g., etanercept), and peritransplant anticoagulation therapy with heparin products, has been associated with an improvement in intrahepatic islet engraftment efficiency (159), there remains a loss of transplanted islet mass in the early posttransplant period (Fig. 5). Additional targets for anti-inflammatory therapy based on preclinical studies using mouse models of intrahepatic islet transplantation include the protease inhibitor α-1-antitrypsin (374) and the chemokine receptor for CXCL1 and CXCL8 (CXCR1/2) (375). α-1-Antitrypsin may promote transplanted islet survival in part through induction of IL-1 receptor antagonist production (376), and administration of the IL-1 receptor antagonist anakinra also promotes survival of transplanted human islets, including an additive effect to that also seen with etanercept (377). Inhibition of IL-1β signaling with anakinra protects islet β-cells directly against apoptosis and indirectly against the potentiation of amyloid formation (378).

Preliminary results from early phase clinical trials in humans have reported favorable effects of anakinra (379) and of reparixin, an allosteric inhibitor of CXCR1/2 (375), on functional outcomes following intrahepatic alloislet transplantation in a small number of subjects. CXCL1 is released by islets in response to IL-1β and operates as a chemokine for polymorphonuclear and natural killer cells expressing CXCR1/2 that are attracted to the engraftment site within the liver (375). Reparixin is currently under clinical phase 2/3 evaluation in a North American trial of intrahepatic islet autotransplantation following total pancreatectomy (NCT01967888), as well as phase 3 evaluation in an international trial of intrahepatic islet allotransplantation for type 1 diabetes (NCT01817959) (ClinicalTrials.gov; accessed 7 May 2018). Future comparative studies will be required to determine which approach, or combination of approaches, can most safely and effectively protect transplanted islets from inflammation-mediated destruction during the critical posttransplantation period until engraftment.

Novel approaches to immunosuppression

The dependence of islet allotransplantation on calcineurin inhibitor–based regimens for immunosuppression limits current application for patients with type 1 diabetes because of the risk for renal toxicity and consequent impairment in kidney function. Two novel approaches employing inhibition of T lymphocyte adhesion and costimulation have undergone evaluation as part of calcuneurin inhibitor–free immunosuppression regimens in islet transplantation. The anti–leukocyte functional antigen-1 antibody efalizumab inhibits T lymphocyte activation and trafficking by blocking the costimulatory attachment of the CD11a subunit of leukocyte functional antigen-1 to the intercellular adhesion molecule-1. Belatacept is a fusion protein composed of the Fc fragment of human IgG1 immunoglobulin linked to the extracellular domain of cytotoxic T lymphocyte-associated protein-4, which blocks the costimulatory attachment of CD28 to CD80 and CD86 on antigen-presenting cells that is required for T-cell activation. Efalizumab-based, calcineurin inhibitor–free immunosuppression has been associated with the achievement of insulin-independence in 12 intrahepatic alloislet recipients with eight patients receiving islets from a single donor pancreas and four receiving islets from a second donor pancreas (380, 381), and belatacept-based, calcineurin inhibitor–free immunosuppression has been associated with the achievement of insulin independence in five intrahepatic alloislet recipients with four patients receiving islets from a single donor pancreas and one receiving islets from a second donor pancreas (382). All patients have exhibited remarkably stable kidney function through at least a median 1-year follow-up.

The islet graft functional outcomes for efalizumab- and belatacept-based immunosuppression are similar to what has been reported using the B7 protocol that depends on the calcineurin inhibitor tacrolimus (159), and they support that inhibition of T lymphocyte adhesion and costimulation can provide effective protection against alloimmune and autoimmune responses, as well as likely anti-inflammatory effects that may promote islet engraftment. There have been concerns, however, regarding immunologic safety. Both efalizumab and belatacept have been associated with increased risk of posttransplant lymphoproliferative disorders and progressive multifocal leukoencephalopathy caused by reactivation of JC virus infection. Efalizumab was withdrawn from the market by its sponsor after four cases of progressive multifocal leukoencephalopathy were reported out of >40,000 patients receiving the drug for the treatment of psoriasis, and belatacept has had its prescriptions related to kidney transplantation restricted to patients with evidence of immunity against EBV and so at lower risk for posttransplant lymphoproliferative disorders. The lower doses of these agents studied in islet allotransplantation than used for their labeled indications may allow for an acceptable risk/benefit profile, and a completed phase 2 study of belatacept in intrahepatic alloislet transplantation for type 1 diabetes will provide important safety and efficacy data regarding this approach (NCT00468403; ClinicalTrials.gov; accessed 7 May 2018).

“…elevated markers of β-cell injury at 24 hours after islet transplantation have been associated with worse islet graft functional outcomes within the first 3 months posttransplant…”

Novel assays to monitor β-cell death

The identification of islet β-cell loss following transplantation depends on functional assessment of glycemic control, requirements for exogenous insulin, and measures of insulin secretion (often using C-peptide) that is most accurate when using dynamic evaluation of the β-cell secretory capacity as an estimate of functional islet β-cell mass (102, 103). Implementation of these metabolic methods to replace less sophisticated traditional measures of β-cell dysfunction (such as fasting levels of glucose, C-peptide, and HbA1c) are essential to better detect cellular injury prior to the development of islet graft dysfunction. These more sensitive indicators could lead to interventions that might prevent clinically significant islet graft loss. This may be particularly helpful to either guide immune monitoring of humoral and cellular alloimmune and autoimmune markers, or to interpret the potential significance of a newly detected alloantigen or autoantigen reactivity for a transplanted islet β-cell graft before functional deterioration is evident. Because as much as 25% of the transplanted islet mass may be lost early following intraportal infusion (214), several approaches have been used to identify damaged islets in the circulation during this period, including by insulin mRNA (383, 384), glutamic acid decarboxylase-65 (385), miRNA375 (386), and unmethylated insulin DNA (387, 388). These studies have also been used to detect β-cell death in cultured islets prior to transplantation (385, 387), and so may be useful for investigations aimed at optimizing both islet viability following isolation and survival after infusion.

Importantly, elevated markers of β-cell injury at 24 hours after islet transplantation have been associated with worse islet graft functional outcomes within the first 3 months posttransplant (385, 388), and they can be modulated by anti-inflammatory therapy during the first week posttransplant (386), suggesting possible utility for both predicting early engraftment and assessing interventions aimed at improving islet survival during the engraftment period. Additionally, early persistent biomarker elevation has been associated with worse islet graft function at 3 months (387), and secondary “peaks” occurring later during follow-up may predict subsequent islet graft functional decline (384). In contrast to an increase in biomarker signifying islet β-cell injury and death, another approach monitors transplanted islet-specific exosomes isolated from peripheral blood samples using anti-HLA antibodies specific to the islet donor and where preliminary data suggest a decline in islet-specific exosome signal may be detectable prior to autoimmune recurrence (389). Further validation of these novel assays for monitoring posttransplant islet β-cell loss is required in both autoislet and alloislet settings, as well as for alloislets in conjunction with assays for detecting humoral and cellular alloimmunity and autoimmunity.

Conclusions

In conclusion, procedures for islet transplantation are established for patients requiring total or subtotal pancreatectomy for benign diseases of the pancreas where intraportal delivery of autologous islets may prevent or ameliorate postsurgical diabetes, and for patients with type 1 or pancreatogenic (type 3c) diabetes where intraportal delivery of allogeneic islets may establish on-target glycemic control often without the requirement for insulin therapy. The indication for total pancreatectomy with islet autotransplantation is most often to alleviate pain and improve quality of life for patients with recurrent acute or chronic pancreatitis, whereas for islet allotransplantation the indication is most often to alleviate severe hypoglycemia and improve quality of life for patients with excessive glycemic lability or hypoglycemia unawareness, as well as to improve glycemic control for patients already requiring immunosuppression in support of another transplanted organ (e.g., a kidney transplant). Insulin independence is experienced by about one-third of autoislet recipients, who are dependent on the number and health of the islets isolated from their own pancreas, and by about half of alloislet recipients, with the latter sometimes receiving islets from a second (or third) donor pancreas, and is dependent on the functional islet β-cell mass surviving engraftment. The improvement in outcomes for alloislets has been sufficiently robust to consider islet transplantation as a complementary alternative to solitary whole-pancreas transplantation, whereas insulin independence occurs more frequently when pancreas transplantation is performed simultaneous with kidney transplantation from the same organ donor. Novel approaches are under evaluation to further reduce the detrimental consequences of peritransplant inflammation on the intrahepatic engraftment efficiency of both autoislets and alloislets that may result in insulin independence more often and require less donor pancreases, as well as for calcineurin inhibitor–free immunosuppression that may have less potential for adverse effects on kidney function. Xenogeneic and stem cell–derived sources of islet tissue have entered early phase clinical trials, although much remains to be learned regarding the in vivo physiology and immunogenicity of the various products, as well as their potential compatibility with microencapsulation and macroencapsulation devices being tested for immunoisolation and cell retrieval and monitoring purposes. These advances provide more options for cellular therapy in the treatment of diabetes available now as autologous and allogeneic sourced tissue, and they promise potentially unlimited sources of islet tissue available for future application.

Abbreviations

Abbreviations
 
• cGMP

  current good manufacturing practice

 
• cGTP

  current good tissue practice

 
• B7

  Clinical Islet Transplantation Consortium

 
• CMV

  cytomegalovirus

 
• EBV

  Epstein-Barr virus

 
• GFR

  glomerular filtration rate

 
• GLP-1

  glucagon-like peptide-1

 
• hESC

  human embryonic stem cell

 
• HLA

  human leukocyte antigen

 
• IAPP

  islet amyloid polypeptide

 
• IE

  islet equivalent

 
• mTOR

  mammalian target of rapamycin

 
• OGTT

  oral glucose tolerance test

 
• PTT

  partial thromboplastin time

 
• TPIAT

  total pancreatectomy and islet autotransplantation

Acknowledgments

We thank Drs. Chengyang Liu and Ali Naji of the University of Pennsylvania Perelman School of Medicine (Philadelphia, PA) for performing a critical review of the manuscript.

Financial Support: This work was supported by Public Health Service Research Grants R01 DK 091331-7 (to M.R.R.), R01 DK 97830-5 (to M.R.R.), and R01 DK 39994-26 (to R.P.R.).

Disclosure Summary: The authors have nothing to disclose.