Porous Scaffolds Support Extrahepatic Human Islet Transplantation, Engraftment and Function in Mice

Abstract

Islet transplantation as a therapy or cure for type 1 diabetes has significant promise but has been limited by islet mass requirements and long-term graft failure. The intrahepatic and intravascular site may be responsible for significant loss of transplanted islets. Nonencapsulating biomaterial scaffolds provide a strategy for architecturally defining and modulating extrahepatic sites beyond the endogenous milieu to enhance islet survival and function. We utilized scaffolds to transplant human islets into the intraperitoneal fat of immunodeficient mice. A smaller human islet mass than previously reported reversed murine diabetes and restored glycemic control at human blood glucose levels. Graft function was highly dependent on the islet number transplanted and directly correlated to islet viability, as determined by the ATP-to-DNA ratio. Islets engrafted and revascularized in host tissue, and glucose tolerance testing indicated performance equivalent to healthy mice. Addition of extracellular matrix, specifically collagen IV, to scaffold surfaces improved graft function compared to serum-supplemented media. Porous scaffolds can facilitate efficient human islet transplantation and provide a platform for modulating the islet microenvironment, in ways not possible with current clinical strategies, to enhance islet engraftment and function.

INTRODUCTION

Islet transplantation as a cell-based therapy for type 1 diabetes offers the possibility of a cure or long-term therapy for a chronic disease that is currently managed with daily injections (26,32). Even with intense regimens to control insulin and blood glucose levels, many type 1 diabetic patients still experience significant morbidity and early mortality from their disease (19,27). The clinical practice of transplanting donor islets by intraportal infusion into the liver vasculature has shown early promise. However, graft failure over the first few post-transplant years and the large donor islet mass required to achieve the initial success rates have limited islet transplantation to a small fraction of patients with type 1 diabetes, typically those who already have significant complications (20,24). The hepatic and intravascular site of islet transplantation has been identified as contributing to poor islet survival posttransplant due to the immediate damage caused by the instant blood-mediated inflammatory reaction (IBMIR) that occurs when islets encounter whole blood and a foreign hepatic environment that provides additional challenges to engraftment and long-term survival (3,23). Extrahepatic and extravascular sites have been challenging because of the need for creating space to accommodate the relatively large mass of donor cells required while facilitating survival and function (8,17,30). Biomaterial scaffolds provide an opportunity to define an extrahepatic site beyond a surgical pouch and enhance the site for maximal islet engraftment, survival, and function, significantly improving the efficiency of islet cell delivery as a therapy for type 1 diabetes.

The transplantation of islets on porous biodegradable materials, rather than encapsulated for immunoprotection, has emerged as a promising strategy for long-term islet function (4,13,25). Much of the biomaterial-focused research for islet transplantation has used a strategy of encapsulating islets to limit immune attack. Although this approach has benefits to the recipient by reducing or eliminating the immunosuppressive load, it has been challenged by long-term islet survival (2,9,15). Encapsulation inherently prevents transplanted islets from revascularizing, reenervating, and remodeling the surrounding tissue and extracellular matrix (ECM), which leaves islets chronically dependent on diffusion across large distances for oxygen, nutrition, and waste removal. While preventing direct immune attack, encapsulating materials must allow diffusion of insulin and glucose and thus may allow diffusion of islet antigens and host cytokines, preventing complete immunoisolation (16). In contrast, while requiring some form of immunosuppression or tolerance induction, porous biomaterial scaffolds integrate with the host, facilitating rapid tissue ingrowth and revascularization while providing the opportunity to architecturally define an extrahepatic site. Additionally, the polymeric material can be functionalized to present ECM proteins, locally deliver trophic factors or the encoding genes, or serve as a vehicle for cell cotransplantation to enhance transplant success (1,7,22,25).

We have previously shown that the scaffolds are an effective platform for the extrahepatic transplantation of islets in syngeneic (4,25) and allogeneic mouse models (13) and now report the feasibility of transplanting human islets on polymer scaffolds using the non-obese diabetic-severe combined immunodeficient interleukin receptor γ deficient (NOD-scid IL-2Rγ^nu11; NSG) mouse model (14). Human islets have a long list of cytological and functional characteristics that make them distinct from murine islets (5,6,10,31). Despite their dramatically increased clinical relevance, human islets may not behave the same as murine islets when combined with biotechnologies due to differences in basement membrane structures, endocrine cell numbers and distribution, glucose responsiveness, and the reduced effectiveness of human insulin. These factors and the practical implications of transplanting a larger mass of islets combine to create challenges to reproducing the results seen with technologies developed for syngeneic murine islets when transitioning to human islets. These challenges highlight the importance of developing islet-enhancing technologies specifically for human islets to maximize the effectiveness of clinical translation. Thus, to investigate the potential for clinical relevance of scaffolds as vehicles for human islet transplantation, human islets were obtained from multiple clinical distribution centers and transplanted on scaffolds in the NSG mouse model, with a mass of islets sufficient to achieve diabetes reversal. We also investigated adding human ECM molecules as a means to present microenvironment signals to enhance islet function posttransplantation. The studies herein investigate microporous scaffolds as a method for human islet xenotransplantation in mice and as a platform technology for the development of interventions to enhance human islet transplantation.

MATERIALS AND METHODS

Scaffold Fabrication

Scaffolds were fabricated from poly(lactide-co-glycolide) (PLG) microspheres as described previously (4,25). Microspheres were formed by homogenizing a 6% solution of PLG (Lakeshore Biomaterials, Birmingham, AL, USA) in dichloromethane (Sigma-Aldrich, St. Louis, MO, USA) in a solution of 1% polyvinyl alcohol (Acros Organics, Pittsburgh, PA, USA). Washed and lyophilized microspheres were combined in a 1:30 ratio with NaCl particles (Mahnckrodt, St, Louis, MO, USA) 250–425 µm in diameter. The mixture was pressed in a steel die (International Crystal Labs (Garfield, NJ, USA) at 1,500 pounds per square inch and gas-foamed after equilibration to 800 pounds per square inch CO₂ gas. Salt particles were removed by immersion in water with repeated washing resulting in cylindrical scaffolds 5 mm in diameter and 1 mm in height. Scaffolds treated with ECM molecules and control scaffolds were exposed to 0.5 N NaOH (Mallinckrodt) for 20 s and thoroughly washed in water. Sixty microliters of purified human collagen IV (Sigma-Aldrich) or ECM mixture (BD Biosciences, Franklin Lakes, NJ, USA) was applied to scaffolds as a 1 mg/ml solution in 0.25% acetic acid (Sigma-Aldrich). After drying, scaffolds were rinsed in 70% ethanol and Connaught Medical Research Laboratories (CMRL) 1066 media (Mediatech, Manassas, VA, USA) prior to islet seeding. Islet aliquots were hand-seeded onto scaffolds immediately prior to transplantation with a custom-made micropipetting device immediately prior to transplantation made from a sharpened and thinned glass Pasteur pipette with an angled tip attached to a syringe to facilitate precise application of islets.

Human Islet Acquisition, Culture, and Assays

Human islets were obtained through the Integrated Islet Distribution Program (IIDP, http://iidp.coh.org) from multiple centers representing the IIDP and the Juvenile Diabetes Research Foundation (JDRF). Offers meeting criteria of >70% purity and >90% viability were accepted for transplantation studies. Upon receipt, human islets were washed in CMRL 1066 media (MM1 type) supplemented with 2.5% human AB serum (Mediatech). Islets were cultured at 27°C for up to 6 days prior to transplantation in the same medium, which was supplemented with 10% serum and changed every 48 h. On arrival and on the day of transplant, islets were assayed for viability as measured by ATP/DNA (nM/ng) by performing plate-based assays on triplicate aliquots for ATP with a CellTiter-Glo (Promega, Madison, WI, USA) kit and DNA with a Quant-iT picogreen (Invitrogen, Carlsbad, CA, USA) kit (28). Islet equivalents (IEQ) were calculated based on 10.5 ng DNA/IEQ, adjusted for batch purity. Islet purity and morphology were assessed by staining aliquots with dithizone (Sigma-Aldrich) and hand counting. Shipments received that contained islets with extremely poor or overdigested morphology, or contaminated islets were not transplanted.

Mouse Transplantation

Male NOD-scid IL-2Rγ^null (NSG, Jackson Laboratories, Bar Harbor, ME, USA) mice, 8–5 weeks old, were used as recipients for transplant studies. Recipients were rendered diabetic 5–6 days pretransplantation with 130 mg/ kg streptozotocin (Sigma-Aldrich) in citric acid buffer injected intraperitoneally. Nonfasting blood glucose was measured using tail blood samples and an Accucheck Aviva glucometer (Roche Diagnostics, Indianapolis, IN, USA). Recipients were considered diabetic if blood glucose measurements greater than 300 mg/dl were maintained for the 2 days of pretransplantation. Recipients were anesthetized with an intraperitoneal injection of ketamine (Ketaset^®, Fort Dodge, IA, USA) and xylazine (Anased^®, Lloyd Labs, IA, USA). The abdomen was shaved and sterile-prepped. A midline incision was made, and the epididymal fat pads (EFPs) identified. An isletseeded scaffold was wrapped in each EFP and placed in the abdominal cavity. Scaffolds remained wrapped in EFP effectively and long-term without the need for suture (4). For kidney capsule implantation, the lumbar area was surgically prepared as above, the kidney visualized, and islets injected under the capsule using fine PE50 tubing (Becton Dickinson, Sparks, MD, USA) attached to a high-precision dispensing syringe for precise seeding of the islets. The wound was closed in two layers. Mice were given analgesics pre- and postoperatively as needed and monitored for signs of infection or duress. All murine studies were approved by the Northwestern University Animal Care and Use Committee.

Posttransplant Monitoring

Mice were followed by body weight and nonfasting blood glucose at least every 2 days for the first week posttransplant and at least weekly thereafter. No insulin was administered. In a portion of mice, an intraperitoneal glucose tolerance test (IPGTT) was performed. For the IPGTT, control mice and mice with islets transplanted on scaffolds or under the kidney capsule were fasted for 4 h. A baseline fasting blood glucose was then measured, and 2 g per kg body weight of 50% dextrose (Hospira, Lake Forest, IL, USA) was administered intraperitoneally. Blood glucose was measured at 15, 30, 60, and 90 min postinjection, and a baseline-adjusted area under the curve was calculated for each animal.

Histology

In a subset of mice, the implant-containing EFP(s) were removed, fixed, and paraffin-embedded. Five-micrometer sections were stained with hematoxylin and eosin (H&E) and Masson’s trichrome (Polyscience, Warrington, PA, USA) or used for immunohistochemistry with the following primary antibodies for islet cell hormones: (i) insulin, guinea pig anti-insulin (1:25, Dako, Carpinteria, CA, USA); (ii) somatostatin, rabbit anti-mouse somatostatin (1:500, Millipore, Billerica, MA, USA); and (iii) glucagon, mouse anti-mouse glucagon (1:500, R&D Systems, Minneapolis, MN, USA). To identify vascularization rat anti-mouse CD31 (Abcam, Cambridge, MA, USA) was used. Proliferating cells were identified with proliferating cell nuclear antigen (PCNA) Rabbit anti-PCNA (1:400, Abcam, Cambridge, MA, USA). Secondary antibodies used were Alexa Fluor 546-goat anti-rabbit or anti-mouse, Alexa Fluor 488 goat anti mouse (1:500, Invitrogen) or DyLight 488 donkey anti-guinea pig (1:500, Jackson, West Grove, PA, USA). Nuclear counterstaining was performed with Hoechst (1:2,000, Invitrogen). Stained slides were imaged and stitched together using PTGui Pro (New House Internet Services B.V., Rotterdam, The Netherlands). Analysis was performed in National Institutes of Health ImageJ (http://rsb.info. nih.gov/ij/).

Statistics

Statistical analyses were performed using statistical package Graphpad Prism (Graphpad, La Jolla, CA, USA). Results are presented as mean ± standard error of the mean (SEM). Student’s t test or one-way analysis of variance with appropriate posthoc tests to determine statistical significance of groups were used. Differences in the number of days for diabetes reversal were compared using Kaplan–Meier survival analysis. A value of p < 0.05 was considered statistically significant. Error bars represent SEM in all figures.

RESULTS

Scaffolds Supported Extrahepatic Human Islet Transplantation in NSG Model

Initial studies examined the ability of microporous PLG scaffolds to serve as a platform for the transplantation of human islets into NSG mice. Human islets obtained from the IIDP via multiple islet centers were successfully seeded onto scaffolds (Fig. 1A) and transplanted in the EFPs of NSG mice with streptozotocin-induced diabetes. Two scaffolds were used per animal to enable islet seeding at a maximum density of ~50 IEQ/mm². As demonstrated by trichrome staining, at 145 days post-transplantation, the islets maintained a healthy morphology while the scaffolds were infiltrated with host tissue (Fig. 1B). Blood vessels and capillaries can be seen in and around the islets and throughout the scaffold.

Figure 1.

Scaffolds support extrahepatic human islet transplant in NSG mice. (A) Image of microporous scaffold seeded with human islets pretransplantation. (B) Trichrome staining of human islets on scaffolds 145 days postimplantation with magnified insets demonstrates healthy islet morphology (top right), host tissue infiltration of scaffold space (bottom right), and revascularization (black arrows) of islets and scaffolds. S, scaffold infiltrated with host tissue; E, epididymal fat pad (EFP). NSG, non-obese diabetic-severe combined immunodeficient interleukin receptor γ deficient.

Human Islet Variability Impacted Transplant Outcomes

Islet shipments demonstrated the variability of human islet preparations available for research and how the variability in viability and purity impacted graft function post-transplant in mice. Fourteen IIDP and JDRF shipments (age [mean ± SD]: 41.4 ± 12.2, sex: 4 female, 9 male, 1 unknown) were characterized for their viability, as measured by ATP/DNA ratio (28), and purity, as determined using dithizone staining. Although only islet shipments of high viability (>90%) and purity (>70%) were requested, human islet shipments demonstrated significant variability in viability and purity upon receipt (Fig. 2A). To determine how the measured viability and purity affected islet graft function, we analyzed the blood glucose for each of 21 mice that received islets from seven shipments at a density of 2,000 IEQ per mouse, using scaffolds (n = 16) or the kidney capsule (n = 5). Averaging blood glucose over the first 3 weeks posttransplant revealed a significant correlation between average blood glucose and the measured viability of the islets (ATP/DNA) on the day of transplant (p<0.0005) (Fig. 2B). A correlation between average blood glucose and islet purity was not observed (p = 0.112, data not shown).

Figure 2.

Human islet variability affected graft function. (A) Only islets with broadcast criteria of >90% viability and >70% purity were accepted for shipment. However, on arrival, significant variability in viability and purity was observed. (B) Using islets from seven different shipments, 2,000 islet equivalents (IEQs) were transplanted into mice on scaffolds (n = 16) or under the kidney capsule (n = 5). Analysis of the recipients’ blood glucose during the first 3 weeks posttransplantation revealed a strong correlation with islet viability as measured by ATP/DNA on the transplant day (p < 0.0005).

Human Islet Graft Function Was Dose Dependent and Identified a Minimal Mass of Islets

We subsequently investigated the minimum islet mass capable of reversing diabetes in NSG mice with streptozotocin-induced diabetes. Using islets from seven IIDP offers, representing four different islet distribution centers, 58 mice were transplanted with 500, 1,000, or 2,000 human IEQ each. Islets from these shipments had a measured ATP/ DNA ratio of 6.2 ± 1.4 and purity of 70% ± 4.5%. Average blood glucose during the first week following transplantation demonstrated a significant difference between doses (Fig. 3). Only mice receiving 2,000 IEQ had average blood glucose under 200 mg/dl, suggesting that this was the minimal engrafting mass of human islets needed to reverse diabetes in NSG mice.

Figure 3.

Human islet graft function is dose dependent. After transplantation, islet graft function, as measured by average blood glucose, is highly dependent on the number of islets transplanted. Mice were transplanted with 500 (n = 6 under kidney capsule, n = 12 on scaffolds in EFP), 1,000 (n = 5 under kidney capsule, n = 14 on scaffolds in EFP), or 2,000 (n = 5 under kidney capsule, n = 16 on scaffolds in EFP) human IEQ each. For 500 vs. 1,000 and 1,000 vs. 2,000: p < 0.05, 500 vs. 2,000: p < 0.001.

Human Islets Survived Long-Term on Scaffolds in NSG Model and Were Engrafted in Host Tissue

Using 2,000 IEQ from a high-quality (ATP/DNA = 7.6, purity = 95%) islet shipment, glycemic control at human levels was restored in NSG mice and maintained past 140 days (n = 10) (Fig. 4A). In healthy control NSG mice, the average nonfasting blood glucose level was ~150 mg/dl, whereas transplantation of the human islets in diabetic NSG mice resulted in average nonfasting blood glucose levels under 100 mg/dl. At 90 days, an IPGTT revealed that engrafted human islets transplanted on scaffolds or under the kidney capsule normalized blood glucose with glucose excursions equivalent to those observed in healthy control mice, demonstrating that engrafted human islets were functionally robust and able to control a large glycemic load (Fig. 4B). At 169 days post-islet transplantation, the implants in select mice were removed to confirm that euglycemia was due to the human islet transplants. Once the implants were removed, hyperglycemia immediately returned (Fig. 4C).

Figure 4.

Engrafted human islets maintained long-term glycemic control and accommodated glucose challenges. (A) In 10 mice receiving 2,000 IEQ of high-quality islets, diabetes was reversed beyond 140 days when transplanted under the kidney capsule (n = 5) or on scaffolds (n = 5). Diabetes was reversed and glycemic control at human blood glucose levels (<100 mg/dl) was achieved in all mice. (B) At 90 days posttransplant, an intraperitoneal glucose tolerance test (IPGTT) was performed. The baseline-corrected area under the curve (AUC) for healthy control NSG mice (n = 3), mice with kidney capsule implants (KC; n = 3), and mice with scaffold implants (n = 5) was the same. (C) After 169 days of stable blood glucose, hyperglycemia returns immediately in mice whose islet-containing implants are removed (n = 2). The data for both animals are plotted.

Histology of islets on scaffolds 145 days posttransplant demonstrated the presence of the other two dominant islet cell populations via staining for glucagon (α-cells) and somatostatin (δ-cells) (Fig. 5A). The islets, which are avascular at transplant, were revascularized as visualized by insulin and CD31 staining (Fig. 5B). Additionally, staining with anti-PCNA antibodies demonstrated cells positive for insulin and PCNA, consistent with β-cell proliferation and indicating that islet turnover and remodeling was occurring (Fig. 5C).

Figure 5.

Histology demonstrates engrafted islet characteristics. (A) Immunoflourescence for glucagon (left image) and somatostatin (right image) demonstrates distribution of islet endocrine cell populations. Dashed line represents islet margin. Nuclei are stained with Hoechst. (B) Islets were revascularized as visualized by CD31 immunofluorescence. (C) Islets contained cells positive for insulin (cytoplasmic) and proliferating cell nuclear antigen (PCNA; nuclear), indicating that β-cell proliferation was occurring. White arrows and magnified inset demonstrate PCNA⁺ nuclei in cells positive for cytoplasmic insulin.

Engrafted Islets Placed Under Glycemic Stress Failed Over Time

To further investigate graft function, three mice were transplanted with 2,000 IEQ (1,000 IEQ on each of two scaffolds), and euglycemia (i.e., blood glucose under 100 mg/dl) was achieved and maintained. At 145 days posttransplant, one of the two scaffolds (i.e., half of the transplanted islet mass) was removed (Fig. 6). All three animals remained euglycemic for 4 days after scaffold removal, suggesting that sufficient insulin secretion was maintained during that time period. However, after 4 days, glucose levels increased consistent with failure of the islet graft. One animal became hyperglycemic rapidly, reaching blood glucose levels of over 300 mg/dl within 10 days. The second animal had slower graft failure, with euglycemia lasting more than 14 days after removing the scaffold. The third animal maintained normoglycemia, albeit with a higher average blood glucose level (>100 mg/dl). After 25 days, the remaining implant was removed from each mouse, and blood glucose in all three mice increased to > 300 mg/dl.

Figure 6.

Engrafted islets failed when placed under long-term glycemic stress. In three mice with 145 days of stable graft function, half of the transplanted mass was removed by removing one fat pad containing 1,000 IEQ. This resulted in gradual loss of glycemic control. Removal of the remaining mass resulted in return to pretransplant levels of hyperglycemia (blood glucose > 300 mg/dl). The data for the three mice are plotted.

ECM Coatings Improved Human Islet Graft Function

The scaffolds were also used to assess the impact of specific ECM components on the islet transplant microenvironment (25). Scaffolds were coated with purified human collagen IV, a human ECM mixture containing laminin, collagen IV, and heparan sulfate proteoglycans, or control serum-supplemented media, seeded with human islets and then transplanted into NSG mice. A subminimal mass (less than 2,000 IEQ) of islets from one of two different shipments (1,500 IEQ with an ATP/DNA of 3.9 or 1,200 IEQ with and ATP/DNA of 4.9) were implanted into diabetic mice on treated scaffolds (n = 8 for each condition). A subminimal mass of islets was transplanted to increase the sensitivity of the transplants to differences between groups. This resulted in relatively low rates of murine diabetes reversal (sustained nonfasting blood glucose <200 mg/dl). The end goal of these studies was the improvement of human islet graft function and not simply the reversal of murine diabetes. Thus, the presence of any graft function, characterized as levels of blood glucose below 300 mg/dl, was used to identify differences between conditions. A Kaplan–Meier analysis indicated that addition of collagen IV to scaffolds significantly improved the degree of graft function in mice receiving a subminimal mass of islets (Fig. 7, p < 0.02 vs. serum control). The ECM mixture appeared to affect graft function as well, but to a lesser degree (p = 0.08 vs. serum control). We additionally performed analysis of the average blood glucose values of each group, which indicated a significant improvement in average blood glucose with the addition of ECM or collagen IV to scaffolds (ECM p < 0.05 and collagen IV p < 0.001 vs. serum control; ECM vs. collagen IV was not significant), which is consistent with the conclusions from the Kaplan–Meier analysis.

Figure 7.

Addition of ECM molecules enhanced transplant outcomes. Kaplan–Meier analysis of animals receiving transplants of subminimal mass of islets on scaffolds treated with control media (serum), a human extracellular matrix (ECM) mixture, or human collagen IV. Graft function was determined by the percentage of animals with blood glucose of <300 mg/dl (n = 8 per condition).

DISCUSSION

Adopting scaffolds for islet transplantation has the potential to significantly advance the effectiveness of islet transplantation for type 1 diabetes, which has been limited in its clinical application to a small fraction of patients. Multiple reports in the literature have called for the development of alternative approaches to improve the clinical approach, but strong evidence for an extrahepatic site has yet to suggest a clearly superior alternative (15,17,30). Biomaterial scaffolds can be leveraged to provide greater benefits than may be achieved with only the adoption of a new surgical site; they have the ability to architecturally define and subsequently introduce modifications to the transplant environment to enhance islet survival and function in ways that cannot be achieved in the currently used approach of portal infusion. Microporous PLG scaffolds have previously been demonstrated as highly effective in syngeneic and allogeneic tolerance-inducing murine models of islet transplantation (13). However, there exist significant differences in the cytoarchitecture and physiology of human and mouse islets that make direct translation of murine results challenging. In human islets, the cytological arrangement and relative numbers of β-, α-, and δ-cells within each islet differ from murine islets (5,6). There is a unique human peri-islet double basement membrane, distinct from the basement membrane found in mice (18,31). Many key drug-metabolizing and stress-related enzymes such as cytochrome P450 (CyP450), heat shock protein70 (HSP70), catalase, superoxide dismutase, heme oxidase, and protein kinase C (PKC) exhibit different levels of function or expression in human and mouse islets (29,33,34). Additionally, pathways for glucose sensitivity and responsiveness also vary between islets from different species (6,10,12). Finally, molecules that can induce β-cell damage such as nitric oxide, hydrogen peroxide and inflammatory cytokines have significantly different effects between species (10,33). These and other differences result in challenges to translating success seen with technologies developed with syngeneic islets directly to human islets. Thus, acknowledging the scope of these differences and to improve the clinical relevance and applicability of scaffolds for islet transplantation, we investigated the transplantation of human islets on scaffolds into immunodeficient NSG mice.

Scaffolds supported the survival and function of human islets at the EFP site in NSG mice. The human islet-NSG model enabled investigation of the variables relevant to human islet function from clinically relevant sources. This model demonstrated long-term islet function and maintenance of glycemic control within a human range using fewer islets than previously reported (21). Serial transplant studies demonstrated the minimal mass of islets required for diabetes reversal and how variability in pretransplant islet quality directly affected transplant outcomes. These results additionally support the continued use of ATP/DNA as a relatively simple and cost-effective measure of islet viability with a strong correlation to the resulting graft function (28). As was demonstrated, increasing the glycemic load per engrafted islet by reducing the mass of islets available resulted in graft failure in most animals. The result of high insulin demand on islets may have additionally contributed to low diabetes reversal numbers observed when transplanting a mass of islets below a certain threshold. The mechanism for this outcome will require further study but may reflect the induction of islet loss by the consequences of increased insulin production beyond a certain level. These results indicate that a high degree of demand can induce graft failure in an otherwise stable implant.

We were able to demonstrate that, in the context of a subminimal mass of islets, the addition of human ECM components to scaffolds, specifically collagen IV, can improve human islet graft function in NSG mice compared to serum-supplemented media. In their most basic form, microporous scaffolds define the transplant site architecturally, thereby controlling the islet distribution. Scaffolds can additionally be coated with and present ECM proteins to islets or deliver trophic factors or the genes encoding these factors (11,22,25). These capabilities enable the investigation and local implementation of numerous interventions that have previously been shown to facilitate islet function and survival in vitro but lacked effective in vivo delivery systems. The results herein have demonstrated the addition of purified human ECM to the scaffold surface and the positive impact on islet graft function.

As demonstrated, microporous scaffolds can serve as an important element in an extrahepatic approach that may significantly improve the efficiency and effectiveness of islet cell transplantation as a therapy for type 1 diabetes. The NSG mouse as a model for human islet transplantation offers the potential to identify factors that enhance human islet survival and function in vivo. As a platform technology, advancements to scaffold design and the addition of scaffold-mediated interventions that improve islet transplant, such as the demonstrated ECM coatings, can be easily incorporated into future designs. Continued development of this model may lead to conditions where scaffolds for extrahepatic islet transplantation will demonstrate significantly enhanced effectiveness relative to the hepatic site, providing the basis for clinical adoption. An improved efficiency could enable the current islet supply to treat a significantly larger number of type 1 diabetes patients. In summary, these results establish microporous scaffolds as an effective method for human islet xenotransplantation in mice and support the use of scaffolds for future translational studies as a platform technology for developing interventions to enhance human islet transplantation.