Development of Methods for Transplanting Human Islet within Macroencapsulation Device that Reverse Diabetes in Mice

Douglas Sobel ^* , Wanxing Cui

1. Georgetown University Medical Center, Washington D.C., USA

* Correspondence: Douglas Sobel 

Academic Editor: Chirag S. Desai

Received: October 09, 2025 | Accepted: February 01, 2026 | Published: March 20, 2026

OBM Transplantation 2026, Volume 10, Issue 1, doi:10.21926/obm.transplant.2601268

Recommended citation: Sobel D, Cui W. Development of Methods for Transplanting Human Islet within Macroencapsulation Device that Reverse Diabetes in Mice. OBM Transplantation 2026; 10(1): 268; doi:10.21926/obm.transplant.2601268.

© 2026 by the authors. This is an open access article distributed under the conditions of the Creative Commons by Attribution License, which permits unrestricted use, distribution, and reproduction in any medium or format, provided the original work is correctly cited.

1. Introduction

Type I diabetes mellitus is a chronic autoimmune disorder of islet destruction which causes a high rate of morbidity and mortality [1,2]. Islet transplantation is, however, a promising treatment [3,4].

Successful portal vein islet transplantation has led to islet transplantation in humans [3,4]. Although the efficacy of human islet transplantation has improved over time, a 10-year study found that, even after multiple transplants, only 4.8% of patients remain insulin-free [5]. Furthermore, the current procedure carries many potential complications with the present procedure, including surgical and post-surgical complications such as portal thrombosis, infection, and bleeding, as well as the serious requirement for generalized immunosuppressive therapy, all of which prevent this treatment from being the standard of care [6,7].

Identifying a safer transplant site would thus be helpful. Many alternative sites for transplantation have been studied in animals, including the peritoneum, omentum, muscle, testes, and spleen [8,9,10,11,12], but no successful islet transplants in humans have been reported.

Transplanting islets into the subcutaneous space has advantages, as this site allows easy surgical access, enables retrieval of transplanted cells, and avoids the complications of intraportal administration.

Studies have been performed in which islets are transplanted directly into the SC space and within planar macroencapsulation devices in animals [13,14,15]. Yet there are no reports of SC islet transplantation that normalize blood glucose in humans [16]. One of the main problems of SC islet transplantation is the very low islet viability due to the low oxygen torr in the SC space and fibrosis around the device along with the sensitivity of islets to oxygen [17,18]. Methods to increase oxygenation of SC-transplanted rodent islets [19,20] have not led to successful transplantation in humans. The present study will also explore methods to facilitate oxygenation of the transplant.

Because human islets differ substantially from mouse islets in both structure and function [21], the use of human islets in animal experiments may provide a better understanding of the issues that hinder successful human SC transplantation. However, there is a paucity of data on the subcutaneous transplantation of human islets, and no reports of transplanting human islets into planar macroencapsulation devices designed to inhibit allograft rejection by local immunotherapy, or into small-pore devices that inhibit immune cell penetration.

The objective of this report is to develop methodologies for subcutaneous transplantation of human islets in mice. In vitro studies examined various materials as islet matrices, the capacity of human islets to survive within macroencapsulation devices, and the potential islet-toxic effects of fibroblast growth factor 2 (FGF2), a growth factor to be used in in vivo studies.

FGF2 can increase oxygenation and possibly survival of the transplant by inducing angiogenesis [22,23]. This angiogenesis occurs through FGF-induced endothelial cell proteinases and plasminogen activators, which degrade blood vessel membranes and permit endothelial cells to differentiate into new blood vessels [24]. This, in turn, could augment oxygenation of the islet transplant and will also be assessed in vivo.

FGF was chosen to induce angiogenesis because many reports demonstrate that FGF induces angiogenesis in vivo, supporting transplanted cells, and there are instances where VEGF does not work as well [25,26,27]. Further, FGF may be clinically safer than VEGF-A, which causes unwanted edema [28].

In vivo studies herein examined various approaches to administering FGF2 to augment islet survival, the use of a SC-placed silicone scaffold to support human islets and increase their viability, and different methods for transplanting human islets SC into PTFE macroencapsulating devices.

2. Materials and Methods

2.1 Materials

Provona alginate G and M was obtained from Millipore Sigma (Burlington, MA, USA). Geltrex, Corning Matrigel was obtained from Thermo Fisher Scientific (Waltham, MA, USA) and Cytiva Cytodex 1 and Cytodex 3 was obtained from VWR (Radnor PA, USA).

2.2 Animals

The animal protocols and procedures were approved by the Institutional Animal Care and Use Committee of Georgetown University. Athymic nude mice were obtained from Taconic Farms (Germantown, NY). All mice were female and 6 to 8 weeks old at the start of the experiments. Animals were housed in a specific pathogen-free environment at Georgetown University. The mice were fed a standard rodent chow diet.

2.3 Analytic Methods

Islet viability was assayed using Alamar Blue Viability Reagent (Invitrogen), obtained from Thermo Scientific (Waltham, MA, USA), following the manufacturer’s instructions. Absorbance was measured at 570 nm after a 48-hour incubation using a SPECTRAmax Microplate Reader (Molecular Devices, Sunnyvale, CA, USA). This assay assesses cellular reduction potential by measuring the reduction of resazurin to resorufin, a highly fluorescent product. Each experiment was performed in quadruplicate in a 96-well plate with 100 IEQ of islets in 0.15 ml of media, unless otherwise specified. Cell viability was reported as relative absorbance.

2.4 Device Fabrication

Fabrication of silicone scaffolds: Silicone DDU4351 (Nusil, Carpentaria, CA) was poured into a mold to form a cup-shaped scaffold with an internal volume of 200 µl. After curing for 18 hours at 25°C, a PTFE membrane (Sumitomo Electric Interconnect Products, Inc., 10 µm pore size) was glued onto the top of the silicone cup using DDU4351. The silicone devices were sterilized by gamma irradiation and loaded with cells in Matrigel prior to implantation. Device integrity and heat-sealed edges were assessed by observing leakage from water-filled devices.

Fabrication of PTFE Membrane Devices: PTFE membrane sheets (210 × 297 mm) from Sumitomo Electric Interconnect Products, Inc., with pore sizes of 0.4 µm and 10 µm and a thickness of 25 µm, were cut into strips. Three of the four sides were heat-sealed by swiping the edges with a soldering iron on medium heat to produce a device with final dimensions of 20 mm (L) × 8 mm (W). Prior to implantation, the devices were loaded with cells in Matrigel, and the fourth side was then heat-sealed to create an intact membrane device. In mice that were first preimplanted with devices not containing islets, silicone glue was used to seal the final edge after islets were loaded two weeks later. The thickness of the device was approximately 2-3 mm when filled with islet/matrix. Silicone placed along the edges of the device can serve as structural support and has been used in other prototypes, but not in the devices described in this manuscript.

2.5 Preparation of Human Islets

Human Islet Preparation: Human pancreatic islets were obtained from deceased donors through the Washington Regional Transplant Community (DCTC). Donor characteristics are summarized in Table 1. All donors had no history of diabetes. Islet purity exceeded 90%, viability was approximately 90%, and islet equivalent quantities (IEQ) averaged over 200,000 IEQ/pancreas, as determined by facility quality control procedures.

Table 1 Human Islet Donor Information.

Human islets were prepared as previously described [29] and were incubated in CMRL 1066 containing 10% FBS, 2 mM glutamine, 100 U/ml penicillin, and 100 µg/ml streptomycin prior to in vivo and in vitro studies.

2.6 Procedures

2.6.1 Diabetes Induction, Diagnosis and Maintenance

Streptozotocin-Induced Diabetes. Streptozotocin (Zanosar^®, NDC# 00703-4636-0) was obtained from Teva Parenteral Medicines, Inc., dissolved in sodium citrate buffer at pH 4.5, and administered intraperitoneally (IP) at 250 mg/kg [30].

Assessment and Diagnosis of Diabetes. Blood glucose was measured using home glucometer test strips. Animals were diagnosed with diabetes when blood glucose levels remained above 252 mg/dl for two consecutive days.

Diabetic Animal Maintenance. Diabetic mice were injected SC with 0.5 U of Lantus daily or with 0.5 U of Degludec every 48 hrs when blood glucose exceeded 350 mg/kg.

IP GTT Test. IP GTT was performed in all experiments to assess the effects of the silicone scaffold and devices, with and without FGF2. Baseline blood glucose was measured after a 4 hr fast, and animals were then injected IP with 2 g/kg of glucose. Blood glucose was measured at 30 min, 60 min, 90 min, 120 min, and 240 min.

2.6.2 Subcutaneous Islet Injections

Islet suspensions in Matrigel were administered subcutaneously above the right thoracic mammary gland.

2.6.3 Renal Capsule Injection of Islets

After the mouse is anesthetized, the skin is shaved and disinfected with Betadine, and the kidney is localized by palpation. A 0.5-1 cm incision is made in the peritoneum to expose the kidney. Slight pressure is applied to both sides of the incision to allow the kidney to slide out of the abdominal cavity. A small scratch is made on the right flank of the kidney capsule using a 25-gauge syringe needle. A PE50 tube filled with islets is inserted from the posterior end of the capsule to create space by moving the PE50 tube and slowly injecting the islets. Once the islets are injected, the PE50 tube is removed, and the scratch is cauterized with a low-temperature cautery pen (Surgicare, Dayville, CT). The kidney is then placed back into the cavity, and residual blood is cleaned with physiological solution. Both peritoneal and skin incisions are sutured, and the mice are placed on a heating pad until they are active.

2.6.4 Surgical Implantation of Scaffold and Macroencapsulation Device

Sterile technique will be used throughout the surgery, including sterilized surgical equipment, either by autoclaving or cold-sterilization with Amerse Germicide or a bead sterilizer (Convatec, St. Louis, MO), and a sterile surgical field during the procedure.

Mice were anesthetized in a chamber with 3% isofluorane, and the skin was shaved and disinfected with alcohol and Betadine, swabbing from the center outward over the surgical area. A skin incision was made on the ventral side of the mouse near the right thoracic mammary gland. The skin was retracted, and a subcutaneous pocket was created by dissection on the dorsal side of the right thoracic mammary gland. The device or scaffold containing the islets was then inserted, and the skin was folded back and secured with clips. The incision was then stapled or sutured. Mice were placed on a heating pad until they were active. Animals were monitored every 30 min for 3-4 hours after the procedure, followed by daily monitoring. If signs of pain were observed, buprenorphine SR was administered. Additionally, 3 doses of meloxicam (1/day) may be given post-op in conjunction with buprenorphine SR as a multimodal analgesic combination. Topical antibiotics may be applied in cases of bite marks and ruptured sutures.

2.7 Experimental Procedure

2.7.1 Effect of FGF2 in 0.4 µm Macroencapsulation Devices

Protocols to assess FGF2 in 0.4 µm macroencapsulation devices. The effect of FGF2 applied inside a device and applied inside plus outside a device was assessed in an experiment with and without a 2-week preimplantation period prior to islet transplantation. Mice were placed in the following groups (n = 4).

Group 1, No islets (+); Group 2, Islets placed in the renal capsule (◊); Group 3, FGF2 placed inside a 0.4 µm Device without preincubation (■). Mice were SC transplanted with a device containing 1,000 human islets, 120 µl of Matrigel, 16.66 µl heparin (540 units/ml), and 9 µl of FGF2 (10 mcg/ml, final concentration 600 ng/ml); Group 4, FGF2 inside and outside a 0.4 µm Device (×) without implantation. Mice received 0.15 ml of the following solution outside the transplant site: 120 µl of Matrigel, 16.7 µl (540 units/ml) heparin in PBS, and 9 µl of FGF2 (50 mcg/ml in DMEM) to obtain a final concentration of 3,000 ng/ml. Then, a device containing 1,000 human islets, 120 µl of Matrigel, 16.66 µl heparin (540 units/ml), and 9 µl of FGF2 (10 mcg/ml, final concentration 600 ng/ml) was implanted; Group 5: Preimplantation of device with FGF2 inside device (Δ). Devices containing FGF2 as described above (Group 3), without islets, were SC transplanted. After 2 weeks, 1,000 human islets were introduced into the silicone scaffold, and the devices were then sealed with silicone glue; Group 6, Preimplant of FGF2 and device inside and outside a 0.4 µm Device (□). FGF2 in Matrigel, as described above in Group 4, was administered at the transplant site, followed by insertion of a device containing FGF2 in Matrigel, as described above in Group 3. After 2 weeks, 1,000 human islets were introduced into the devices, which were then sealed with silicone glue.

2.7.2 Effect of FGF2 with 10 µm Macroencapsulation Devices

Mice were SC-transplanted with devices containing FGF2 (600 ng/ml in Matrigel) or devices not containing FGF. After 2 weeks, 1,000 human Islets were introduced into the silicone scaffold and devices, which were then sealed with silicone glue. Mean blood glucose levels were compared with those of islet-transplanted mice in devices not preimplanted with FGF2.

2.7.3 Effect of FGF2 with Silicone Scaffold

Mice were SC transplanted onto scaffolds containing FGF2 (600 ng/ml in Matrigel), as described above. Group 3 did not include islets. After 2 weeks, 1,000 human islets were added to the silicone scaffold, and the scaffolds were sealed with silicone glue. Blood glucose levels were compared with those of islet-transplanted mice in scaffolds that were not preincubated with FGF2.

2.8 Statistical Analysis

All statistical analyses were performed using Microsoft Excel and SPSS 25.0 (SPSS, Inc., Chicago, IL). Differences were evaluated using Student’s t-test or One-way ANOVA. When One-way ANOVA was used to compare group means, Tukey’s HSD was used to assess statistically significant differences.

Previous studies reported estimates of 80% (treated) and 5% (control) for islet transplant-induced normoglycemia. With a sample size of 4, power tables with 0.75 power supported an alpha = 0.05. Data are presented as mean (SD). No outliers were excluded.

2.9 Ethics Statement

The animal protocols and procedures were approved by the Institutional Animal Care and Use Committee of Georgetown University.

3. Results

3.1 In Vitro Viability Studies

3.1.1 Assessment of Cell Matrix

The viability of human islets cultured in Pronova alginate 60% guluronate monomers (60% g), Pronova alginate 60% mannuronate monomers (60% m), Geltrex, and Matrigel were compared. The mean absorbance of islets cultured in each matrix increased over time from day 3 to day 17. The mean absorbance of islets cultured in each matrix at all time points were greater (p < 0.01) than the mean absorbance of islets in a standard culture plate (Figure 1) at days 3, 10, and 17. Islets incubated in Pronova alginate (60% m), Pronova alginate (60% g), Matrigel, and Geltrex had similar mean absorbance scores at all time points. Thus, Matrigel, Geltrex, and Pronova alginates provide similar matrix support for islets in vitro.

Figure 1 Effect of different cell matrices on viability of human islets in vitro by Alamar Blue testing. Mean absorbance scores of islets (n = 4 technical replicates each) incubated in Provona alginate (60% m). Provona alginate (60% g), Matrigel, and Geltrex at 3, 10, and 17 days are depicted. * p < 0.01 vs ‘only cells’ (plate with no matrix).

The ability of Cytodex 1 and Cytodex 3, with and without Geltrex and Matrigel, to support human islet viability in vitro was then compared (Figure 2). The mean alamar absorbance of islets incubated in Matrigel and Geltrex was similar on Day 2, Day 16, and Day 24, while mean absorbance was greater in Matrigel only on Day 8.

Figure 2 The effect of Cytodex 1 and Cytodex 3 with and without Geltrex and Matrigel at 2:1 and 1:2 proportions to support human islets viability in vitro. Mean net absorbances (n = 4 technical replicates each).

At many time points, the mean alamar absorbance was lower when islets were incubated in C1 Cytodex with Geltrex at 1:2 and 2:1 ratios (Cytodex: Geltrex) and when islets were incubated in C1 Cytodex with Matrigel at 1:2 and 2:1 ratios (Cytodex: Matrigel). Similarly, at many time points, the mean Alamar absorbance was lower when islets were incubated in C3 Cytodex with Geltrex at 1:2 and 2:1 ratios (Cytodex: Geltrex) and when islets were incubated in C3 Cytodex with Matrigel at 1:2 and 2:1 ratios (Cytodex: Matrigel).

Furthermore, islets incubated with each dose of C1 Cytodex or C3 Cytodex alone, without Matrigel or Geltrex, had lower mean absorbance than those incubated in Matrigel or Geltrex. Thus, C1 Cytodex and C3 Cytodex provided no additional support for human islet viability in vitro compared with Matrigel and Geltrex.

The viability of human islet incubated in a PTFE macroencapsulating device containing Matrigel was compared with that of islets incubated on a standard culture plate in vitro, using 10, 100, and 1,000 IEQ per well (Figure 3).

Figure 3 Comparison of the viability by Alamar Blue testing of human islet incubated in a PTFE macroencapsulating device containing Matrigel versus incubated on a standard culture plate in vitro using 10, 100, and 1,000 IEQ islets per well (n = 4 technical replicates). * p < 0.01 vs Islets on plate with similar number of islets.

In culture plates and in devices, the mean absorbance was significantly higher (p < 0.01) with a greater number of incubated islets. Furthermore, the mean absorbance was higher (p < 0.01) when islets were incubated in a device compared with standard plate culture, using an equal number of islets. Thus, islets are more viable in the macroencapsulating device than in culture plates.

3.1.2 Assessment of FGF on Human Islets

The potential toxicity of FGF2 on human islet viability was assessed by measuring Alamar Blue absorbance in islets incubated with FGF2 at concentrations of 0 mcg/ml, 1 mcg/ml, and 10 mcg/ml (n = 4) (Figure 4). There were no significant differences in mean absorbance across FGF2 concentrations.

Figure 4 The effect of FGF2 on human islet viability in vitro by Alamar Blue testing. The mean absorbance of human islet was assessed with 0, and 1 mcg/ml and 10 mcg/ml FGF2 and were not significantly different. (n = 4 technical replicates).

3.2 In Vivo Transplant Studies

3.2.1 Direct Subcutaneous Transplant of Human Islets with and without FGF-2

We examined whether human islets directly transplanted into SC could lower blood glucose in diabetic nude mice and whether pre-implanting the SC site with FGF2 two weeks before islet idtransplantation could improve islet viability and produce a greater reduction in blood glucose in diabetic nude mice. The mean blood glucose levels of diabetic mice transplanted with human islets were not significantly lower than those of mice not transplanted with islets (Figure 5). The positive dcontrol mice that received human islets transplanted into the renal capsule had normal mean blood glucose levels from 4 days until the end of the study.

Figure 5 Mean blood glucose of mice transplanted with no islets (+) (n = 4), islets transplanted into the renal capsule (▲) (n = 3), islets transplanted SC space without FGF2 (●) (n = 4), and islets transplanted SC with FGF2 (X) (n = 4). * p < 0.05 vs islets transplanted SC directly.

Save for Day 26, the mean blood glucose of diabetic mice transplanted with human islets at an SC site and preincubated with FGF2 was significantly (p < 0.01) lower from Day 6 until the end of the experiment than in mice SC-transplanted with human islets without FGF2 and in control mice (Figure 5). The lowest mean blood glucose in mice transplanted with islets preincubated with FGF2 was 310 mg/dl, reached at 50 days. Even with FGF2 preincubation, direct SC transplantation of islets was minimally effective in lowering blood glucose.

3.2.2 Subcutaneous Transplant of Islets within a Scaffold

Because SC-transplanted human islets did not function well even with FGF2 administration, alternative methods for SC transplantation of human islets were explored. We hypothesized that SC transplantation of islets onto a surface or scaffold would support islet cell viability and function. We chose to test a silicone scaffold because islets have previously been shown to grow on silicone [31].

We thus determined whether SC placement of a silicone scaffold at the transplantation site would augment islet function and lower blood glucose in transplanted diabetic mice. Furthermore, we examined whether preimplanting a silicone scaffold with FGF2 two weeks before the SC transplant of islets could enhance the effectiveness of the islet transplant. Mean basal blood glucose levels in mice SC-transplanted with islets on the scaffold were significantly lower (p < 0.01) than in control mice without islet transplantation from 6 days post-transplant until the end of the experiment (Figure 6). These scaffold-transplanted mice achieved blood glucose levels below 200 mg/dl by 10 days and below 150 mg/dl by 14 days. All these mice remained euglycemic until the end of the experiment.

Figure 6 The effect of a scaffold alone and a scaffold preincubated with FGF2 on the ability of SC transplanted human islets to lower blood glucose in nude mice. Groups (n = 4): No islet (■); Islet on scaffold (●); and islets on preincubated scaffold and FGF2 (▲). * p < 0.01 vs mice administered no islets. # p < 0.05 vs mice transplanted with islets on scaffold without FGF.

Effect of FGF-2 Administration on Transplanting Islets within a Scaffold.

When examining whether preimplanting FGF2 into a silicone scaffold 2 weeks before the SC transplant could augment the effectiveness of the islet transplant in lowering blood glucose in transplanted mice, the mean blood glucose in mice transplanted with preimplanted FGF2 and scaffolds was significantly lower (p < 0.01) than in mice transplanted with scaffolds without FGF2 from 2 to 14 days post-transplant. After 14 days, mean blood glucose remained similar in each group for the rest of the experiment (Figure 6).

The mean blood glucose in mice transplanted with scaffolds without FGF2 decreased to under 150 mg/dl by day 14 and remained at that level until the end of the experiment. In contrast, in mice transplanted with scaffolds preincubated with FGF2, the mean blood glucose decreased to under 150 mg/dl by day 8. The serum glucose responses to IP glucose infusion at 20 and 35 days post-transplant were similar in mice transplanted with FGF2 and non-FGF2 scaffolds (Figure 7). Thus, the use of a silicone scaffold enables successful SC transplantation of human islets, and FGF2 accelerates the development of euglycemia.

Figure 7 a and b: IPGTT of mice transplanted with scaffold alone or with preincubated scaffold containing FGF2 at 20 and 35 days post transplant. Groups: No Islets (■), Scaffold (●), Scaffold + FGF2 (▲).

3.2.3 Subcutaneous Transplant of PTFE Macroencapsulation Device

The effects of different methods of FGF2 administration and preimplantation of 0.4 µm pore devices on the mean blood glucoses in mice transplanted with human islets was studied (Figure 8).

Figure 8 Effects of different methods of FGF2 administration and preimplantation of 0.4 µm pore devices on the mean blood glucoses in mice transplanted with human islets Groups (n = 4): Group 1 No islets (+); Group 2 Renal capsule (◊); Group 3 FGF2 placed in 0.4 µm Device (■); Group 4 FGF2 in and out of 0.4 µm Device (×); Group 5 Preimplant FGF2 in 0.4 µm Device (Δ); Group 6 Preimplant FGF2 in and out of 0.4 µm Device (□). * p < 0.05 vs control-no islet. ^ p < 0.05 vs FGF inside device (no preimplantation). # p < 0.05 vs FGF inside and outside (no Preimplantation).

Effect of Transplanting a 0.4 µm Macroencapsulating Device Containing FGF2.

Mice transplanted SC with human islets within the 0.4 µm pore device containing FGF2 had significantly lower mean blood glucose levels than control mice without transplanted islets from Day 5 through Day 37 (p < 0.05) (Figure 8). The lowest mean blood glucose in transplanted mice was 304 mg/dl, occurring at 21 days. After that, blood glucose remained in a similar range until the end of the experiment. Mean blood glucose response to IP glucose infusion was significantly lower in mice transplanted with the 0.4 µm pore device than in control mice without transplanted islets at 30, 60, 90, 120, and 240 minutes at both 15 and 28 days post-transplant.

Effect of Adding FGF-2 Outside in Addition to Inside the 0.4 µm PTFE Device.

When assessing the effect of adding FGF2 both outside and inside a 0.4 µm device, compared with FGF2 only inside the device, we found that the mean blood glucose of mice transplanted with FGF2 placed inside and outside the device was significantly lower than that of mice transplanted with FGF2 only inside the device from day 10 post-transplant until the end of the study (Figure 8). Mean blood glucose declined to under 150 mg/dl by day 28 in mice receiving FGF2 inside and outside the device, but declined to only 304 in mice administered FGF2 only inside the device. Mean blood glucose response to IP glucose infusion was significantly lower in mice transplanted with FGF2 outside and inside the device vs mice administered FGF2 only inside at 0, 30, 60, 90, 120, and 240 minutes at both 15 and 28 days post-transplant (Figures 9a, b).

Figure 9 IPGTT in mice at Day 15 (a), 28 (b) following SC transplantation of human islet within a 0.4 µm device. Groups: Group 1 No islets (+); Group 2 Renal capsule (◊); Group 3 FGF2 placed in 0.4 µm Device (■); Group 4 FGF2 in and out of 0.4 µm Device (×); Group 5 Preimplant FGF2 in 0.4 µm Device (Δ); Group 6 Preimplant FGF2 in and out of 0.4 µm Device (□). * p < 0.05 vs control-no islet. ^ p < 0.05 vs FGF inside device (no preimplantation). # p < 0.05 vs FGF inside and outside (no Preimplantation).

Effect of Preimplanting 0.4 µm PTFE Device Containing FGF-2.

We determined whether preincubating a 0.4 µm pore device containing only FGF2 inside the device for two weeks prior to islet cell transplantation would improve islet function and further decrease blood glucose in diabetic animals (Figure 8). The mean blood glucose of mice transplanted with devices containing FGF2, preimplanted 2 weeks prior to the addition of islets (Group 5), was significantly (p < 0.05) lower than that of mice transplanted with devices containing islets and FGF2 without preimplantation (Group 3) from Day 8 until the end of the experiment (Figure 8). Blood glucose declined to under 200 mg/dl at 15 days and to under 150 mg/dl by 24 days. Mean blood glucose remained in the euglycemic range until the end of the experiment at 37 days. The mean blood glucose response to IP glucose infusion was significantly lower in mice transplanted with devices containing FGF2 inside the device and preimplanted with FGF2, compared with mice administered FGF2 inside the device without preimplantation, at all-time intervals at 15 and 28 days post-transplant (Figures 9a, b).

Effect of Preimplanting PTFE Device with FGF-2 Outside and Inside the PTFE Device.

When assessing the effect of preincubating the device with FGF2 placed inside and outside a 0.4 µm device prior to transplant, we found that the mean blood glucoses in mice transplanted with devices preincubated with FGF2 inside and outside the device (Group 6) were lower than in mice transplanted with FGF2 inside and outside the device without preimplantation (Group 4) from day 5 until day 28. Thereafter, the mean blood glucoses were similar until the end of the experiment at 37 days (Figure 8).

From day 18 through day 28, the mean blood glucose of mice receiving preimplanted devices with FGF2 both inside and outside the device (Group 6) was lower than that of mice receiving preimplanted devices with FGF2 only inside the device (Group 5).

At Day 15, the mean blood glucose response to IP glucose infusion was significantly lower (p < 0.01) in mice transplanted with FGF2 inside and outside the device with preimplantation, compared with mice administered FGF2 inside and outside the device without preimplantation, at 30, 60, 90, 120, and 240 minutes (Figure 9a). There were no differences in mean blood glucose response to IP glucose between these groups at any time point on Day 28 post-transplant (Figures 9a, b).

Thus, the SC transplant of a 0.4 µm -pore macroencapsulating device containing FGF2 lowers blood glucose but not to euglycemic levels, whereas placing FGF2 within and around the device and preimplanting the device with FGF2 effectively normalizes blood glucose.

3.2.4 Subcutaneous Transplant of a 10 µm Pore Macroencapsulation PTFE Device

The ability of a larger-pore PTFE (10 µm) macroencapsulation device to support human islets in the subcutaneous space and lower blood glucose in diabetic nude mice was studied.

The mean blood glucose of mice transplanted with islets within the 10 µm pore device was lower than that of the control mice not administered islets at every time point after 3 days (n = 4 each) (Figure 10). By 8 days, all mice were euglycemic and remained so until the end of the experiment.

Figure 10 The effect of a 10 µm pore PTFE device containing human islets to lower the blood glucose in diabetic nude mice. Groups: No Islets (□); 10 µm Device (×); renal capsule (○). (n = 4 each group). * p < 0.05 vs ‘no islet’.

From 10 days post-transplant until the end of the experiment, the mean blood glucose of mice transplanted with the 10 µm pore device were similar to that of the positive control mice transplanted with islets in the renal capsule.

Mean glucose following IPGTT was lower in 10 µm pore device-transplanted mice than in no-islet control mice at all time points on days 15 and 28. Glucose response to IP glucose infusion was lower in mice transplanted with islets into the renal capsule than in 10 µm pore device-transplanted mice at 60 and 90 minutes on day 15. IPGTT at 28 days showed similar mean blood glucose levels at each time point (Figure 11).

Figure 11 a and b: IPGTT of diabetic nude mice transplanted human islets in a 10 µm pore PTFE device at Day 15 and Day 28. Groups: No Islets (□); 10 µm Device (×); renal capsule (○). * p < 0.05 vs ‘no islet’.

The effect of FGF2 preimplantation of the 10 µm pore device on the ability of islet transplants to decrease blood glucose is shown in Figure 12. Mean blood glucose levels in diabetic mice transplanted with human islets in a 10 µm pore device with or without FGF2 were lower (p < 0.01) than in mice not transplanted with islets on days 2 and 4, and remained lower until the end of the experiment. Mean blood glucose in mice transplanted with 10 µm pore devices with or without FGF2 fell below 150 mg/dl by days 8 and 12, respectively, and remained so until the end of the experiment. Mice transplanted with a 10 µm device preincubated with FGF2 had significantly (p < 0.05) lower mean blood glucose levels than those transplanted with a 10 µm device without FGF2 administration from day 2 until day 12 post-transplant. Thereafter, mean blood glucose levels were similar. The effect of FGF2 on mice transplanted with a 10 µm device was assessed by IP GTT at 20 and 35 days post-transplant. Blood glucose clearance was similar in mice implanted with 10 µm devices with and without FGF2 (Figure 13).

Figure 12 Comparison of blood glucose in diabetic mice transplanted islets within a 10 µm Device with and without FGF2 preincubation. Groups: no Islets (■); 10 µm Device (●); 10 µm Device + FGF2 (Δ). (n = 4 each group). * p < 0.05 vs Device without FGF2.

Figure 13 a and b: IPGTT of mice transplanted with human islets within 10 µm pore devices with device and FGF2 preimplantation and without device preimplantation of with FGF2. Groups: no Islets (■); 10 µm Device (●); 10 µm Device + FGF2 (Δ) at 20 days (a) and 35 days (b).

3.2.5 Long Term Effect of Transplanting Islets within 10 µm PTFE Devices and Silicone Scaffolds

Mice transplanted with islets within a 10 µm pore membrane device and silicone scaffold had lower mean blood glucose than control mice not transplanted with islets, from Day 2 through Day 4 and for the remainder of the experiment (Figure 14). Mean blood glucose under 150 mg/dl was achieved in mice with the 10 µm device and silicone scaffold at 8 and 17 days, respectively. After these times, mice remained euglycemic until the end of the study at 300 days Mean basal blood glucose were lower (p < 0.05) in mice transplanted with the 10 µm pore device than in those transplanted with the scaffold until 28 days, after which it was similar until the end of the experiment.

Figure 14 Mean blood glucose of diabetic mice transplanted with no islets (+) (n = 3), islets into renal capsule (■) (n = 4), islets within SC scaffold (×) (n = 4) and islet within 10 µm pore SC device (▲) (n = 4). * p < 0.05 vs SC scaffold.

Following IP glucose infusion, mice transplanted with 10 µm pore device had lower mean blood glucose than scaffold transplanted mice at 60 and 90 min at 25 days. At 58 days, the mean blood glucose following IP glucose infusion were similar at all time intervals (Figure 15).

Figure 15 a and b: IPGTT of diabetic mice transplanted with no islets (+), islets into renal capsule (■), or islets SC on a scaffold (×) and islet SC within 10 µm pore device (▲) at 25 days (a) and 58 days (b). * p < 0.05 vs Scaffold.

At 25 days, the mean blood glucose following IP glucose infusion in mice transplanted under the renal capsule was lower (p < 0.01) than in mice transplanted with scaffolds at 30, 60, and 90 minutes and in mice transplanted with 10 µm devices at 30 minutes. By 58 days, the mean blood glucose following IP glucose infusion in mice transplanted under the renal capsule was similar to that in mice transplanted with the scaffold and 10 µm device at every time point.

Although the SC-transplanted islets within the scaffold decrease blood glucose more slowly than those transplanted with the 10 µm pore device, both the scaffold and the 10 µm pore macroencapsulation device induce similar long-term euglycemia, similar to islets transplanted into the renal capsule.

4. Discussion

A subcutaneous site for islet transplantation has advantages over the current portal site which include being a non-invasive, simpler procedure with fewer complications, allowing retrieval of transplanted cells, and enabling the development of a local immunotherapy device that permits a lower, less toxic dose of immunotherapy [6,7,16]. However, SC transplantation of human islets has not successfully normalized blood glucose.

In this report, we explored various procedures and devices to successfully transplant human islet SC. In turn, we have shown that silicone scaffolds, macroencapsulation devices with 0.4 µm pores that physically inhibit immune cell attack and rejection, and devices with 10 µm pores that may enable local immunotherapy can successfully normalize blood glucose in transplanted mice. Further, we demonstrate methods of FGF2 and device preimplantation to augment successful islet transplantation.

To explore the optimal cell matrix to support human islets, we assessed Geltrex, Matrigel, and alginates, hydrogels commonly used as cell matrices [32,33], and two kinds of Cytodex dextran microspheres, which are also employed as cell matrices. Cytodex 1 is positively charged, and Cytodex 3 beads are coated with collagen [34]. Provona alginate (60% mannuronate monomers), Provona alginate (60% guluronate monomers), Matrigel, and Geltrex provided similar support for human islet viability in vitro, far greater than in standard culture plate conditions. Neither Cytodex 1 nor Cytodex 3 increased islet viability in vitro; in fact, they inhibited it.

We examined the survival of human islets in Matrigel within PTFE macroencapsulation devices and found that the islets remained highly viable within the devices and, in fact, more viable than when incubated on a standard culture plate. A large number of islets, up to 1,000, did not crowd out or inhibit islet viability. Thus, 1,000 human islets per device appeared reasonable and were used in these devices in vivo.

The administration of fibroblast growth factor (FGF2) has been utilized to increase vascular proliferation and prevent tissue ischemia [22,28] and may be useful to support the viability of human islets when transplanted SC. FGF2 is a signaling protein produced by endothelial cells, fibroblasts and macrophages [35]. Along with other functions, FGF2 stimulates capillary cell formation [36].

Since FGF2 can be toxic to cells [37], we determined whether large doses of FGF2, up to 10 mcg/ml, could be harmful to human islets and, in turn, not feasible for SC islet transplantation. We now have evidence that these high doses are well tolerated by human islets and may be used to increase vascularity in SC human islet transplantation.

In vivo studies revealed that transplanted human SC islets are nonfunctional and unable to lower blood glucose in diabetic mice. Similarly, poor survival of rodent islets in the SC space has been previously reported [3,38]. This appears to be due to the low oxygen tension in the SC space [17] and sensitivity of the islets to hypoxia [18]. This harmful hypoxic condition is exacerbated by the destruction of islet vasculature during the process of islet isolation [39].

Since FGF2 administration to the transplant site helps rodent islets survive in muscle, renal capsule, and spleen [40,41,42], we assessed its function in our SC models and found that administering FGF2 at the transplant site 2 weeks prior to islet transplantation enabled transplanted human islets to decrease blood glucose.

However, this improved islet function was only minimal and not sufficient to decrease blood glucose under 300 mg/dl. Previously, Kawakami [43] reported that one week pre-implantation of two SC devices consisting of a polyethylene terephthalate mesh bag coated with polyvinyl alcohol hydrogel containing FGF2, followed by a syngeneic rat islet transplant, induced normoglycemia, a lower blood glucose level than reported here. This difference could be due to the preimplantation of the two PET mesh bags with the FGF2, which also helps to induce neovascularization, the slow release of FGF2, the different animal model, their use of a very much larger number of islets, or the use of rat versus human islets, which may be more difficult to transplant.

Since islets require basement membrane and extra cellular matrix for support and growth [44] and since islet basement membranes are destroyed in the islet isolation process which, in turn, greatly impedes the viability of transplanted islets [4], we hypothesized that transplanting islets onto a surface or scaffold for islet cell adherence and growth will improve islet cell function. Indeed, this is the first report utilizing a silicone scaffold for SC transplantation of islets, which, even without FGF2 administration, allows mice to become euglycemic and remain so until the end of the experiment. This establishes a new, safer, and more effective method to transplant human islets into the subcutaneous space.

Previously, a biodegradable polyglycolic acid polymer sheet was found to partially support the SC transplant of syngeneic rat islets, resulting in a slow decrease in blood glucose to under 200 mg/dl at 77 days [45]. However, in the present report, mice transplanted with a silicone scaffold developed euglycemia more quickly and at lower blood glucose levels. In diabetic rats, islets preimplanted 28 days on a poly (D.L-lactide-co-E-caprolacatone) scaffold resulted in euglycemia [46], whereas in the present report preimplanting the silicone scaffold was not required for the development of euglycemia, which occurred faster.

Preimplanting the silicone scaffold with FGF2 had an initial immediate benefit of a faster decline of blood glucoses to euglycemic levels. However, this positive effect was not present soon there after when baseline glucose levels and glucose response to IP GTT were similar at 16 and 20 days respectively. This catch-up in islet function in non-preimplanted mice may be due to the time needed to develop a sufficiently vascularized transplant site. This timing is consistent with the previous report that endothelial growth factors induce islet blood flow by 3 days versus 7 to 14 days in controls [47]. This early-on improvement of islet function with FGF2 and preimplantation could be useful in SC transplantation of human islets with a scaffold, since early oxygenation of the islets is thought crucial for islet viability and success of the transplantation.

Subcutaneously transplanted planar macroencapsulation devices with 0.4 µm pores, which inhibit the influx of immune cells, have been studied as immunoprotective devices to inhibit allograft rejection in animal models [14,15]. This is the first report describing the successful transplant of human islets within a 0.4 µm macroencapsulating device that normalizes blood glucose levels in any diabetic animal.

We explored various ways to administer FGF2 to achieve optimal human islet cell function in the 0.4 µm pore device. Since intra-islet endothelial cells contribute to revascularization of transplanted islets [48] and angiogenic substances can directly induce angiogenesis within the islet [43], we initially examined whether administering FGF2 within the 0.4 µm pore device containing human islets would successfully lower blood glucose in diabetic mice. We found that administering FGF2 within the device containing human islets just before transplant significantly lowered the blood glucose of transplanted diabetic mice, but only to a nadir of 304 mg%.

To improve on this minimal effect of FGF2, we explored other approaches to administer FGF2. Administering FGF2 to the SC tissue outside the device, along with inside the 0.4 µm device, dramatically normalizes blood glucose.

It has been thought that the short half-life of FGF2 necessitates long-term administration of FGF2 [49]. This study shows that a single dose of FGF2, delivered inside and outside a non-preimplanted device, can be highly effective in normalizing blood glucose. Thus, it appears that the biological effect of angiogenesis may persist even after the interstitial FGF2 level declines.

Preimplanting a 0.4 µm pore device with FGF2 in other ways further improved transplant outcomes. Preimplanting a device containing FGF2 for 2 weeks robustly normalized blood glucose levels, whereas transplanted mice not preimplanted did not achieve euglycemia.

The finding that a preimplanted 0.4 µm device containing FGF2 is more effective at lowering blood glucose early on than adding FGF2 to the outside of a non-preimplanted device containing FGF2 suggests that the duration of FGF2 administration prior to transplant may be more important than the locale of FGF2 placement. This may be because a device already vascularized by FGF2 preimplantation will be immediately helpful in oxygenating islets during the critical period just after transplantation, whereas effective vascularization may not be present at the critical time when FGF2 is administered at the time of transplant.

Further, preimplanting the 0.4 µm pore device for 2 weeks with FGF2 inside and outside the device (Group 6) increased islet function and further lowered blood glucose in diabetic mice, compared with mice that were not preimplanted yet were administered FGF2 inside and outside the device at the time of transplant. This improved blood glucose-lowering effect of preimplantation, which was later absent after 28 days, suggested that full device vascularization may have occurred in mice with the non-preincubated device by 28 days.

The blood glucose-lowering effect of the preimplanted 0.4 µm device containing FGF2 was augmented when FGF2 was also administered outside the device at the time of preimplantation. This allowed mice to achieve euglycemia 4-7 days earlier. This modest added effect of FGF2 outside the device could have clinical relevance.

This positive effect of 2 weeks of preimplantation of the 0.4 µm pore device in the present report contrasts with previous reports demonstrating no positive effect of preimplantation of a planar 0.4 µm pore macroencapsulation device 2 weeks prior to islet (rodent) transplantation [50] and with findings that a positive effect of preimplantation was observed only after 3 months of preimplantation of the device [51,52]. The positive effect with only 2 weeks of preimplantation found in the present report may be due to differences in the structure of the macroencapsulating devices, the use of human islets, or, more likely, the additional use of FGF2 in our studies, which can accelerate the vascularization of the device.

The work outlined in the present report brings us closer to successfully transplanting islets in a device designed to immunoprotect islets from allograft rejection by a physical barrier.

Since a small 0.4 µm pore membrane immunoprotective device that inhibits immune leukocyte passage may not be required in a device to protect islets against allograft rejection using local immunotherapy, a larger-pore macroencapsulating device that may allow better islet oxygenation was studied. We found that SC transplantation of human islets within a 10 µm pore device causes a brisk lowering of blood glucose to euglycemic levels by 8 days. The brisk blood glucose-lowering effect with the 10 µm pore device could be due to improved oxygenation of the transplant.

FGF2 preimplantation with the 10 µm device caused an even more rapid decline in blood glucose to euglycemic levels. However, by 23 days, there was no difference in mean blood glucose if FGF2 was not administered. Nonetheless, the use of FGF2 could be advantageous if improved early oxygen support of islets post-transplant is critical to islet survival.

The improved blood glucose-lowering effect from preimplanting the 10 µm device with FGF2 was not as great as the improved blood glucose-lowering effect from preimplanting a 0.4 µm pore device with FGF2. This may be because the oxygenation of islets is more severely impaired with the 0.4 µm pore device.

Since both the scaffold and the 10 µm pore macroencapsulation device supported short-term viability of human islets and lowered blood glucose in transplanted mice, we assessed the long-term ability of these devices to reverse diabetes. Both the scaffold and the 10 µm pore PTFE device robustly maintained euglycemia in transplanted mice for 300 days. However, the macroencapsulating device was initially superior to the silicone scaffold in supporting islet function, with lower basal blood glucose and a better blood glucose response at 25 days.

It is important to note that the device was made from PTFE. This material is very biocompatible and inert, more inert than other polymers such as nylon and PVDF, which are FDA-approved device materials. Athough PTFE induces very little inflammatory reaction, it is possible that material degradation and foreign body response could activate immune cells and a foreign body fibrotic response which could lead to hypoxic damage to the grafted cells [53].

Scaling such research to human studies presents several challenges, including regulatory barriers, device retrieval, immunologic differences among subjects, and manufacturing constraints. Additionally, the theoretical surface area of the skin that may be required could be relatively large; approximately 500 cm² for a single planar device and 250 cm² for a double planar device.

A limitation of this study was the lack of in vivo experiments that verified the return to hyperglycemia following graft removal in transplanted mice, although the study had appropriate controls of chemically induced diabetic mice not treated with devices in every experiment. Another limitation of the study was the lack of a full histological examination of transplant vascularity and fibrosis.

5. Conclusion

The SC transplantation of a silicone scaffold with human islets enabled normalization of blood glucose levels, and preimplanting a scaffold with FGF2 further hastened the reduction of blood glucose.

Among the different procedures for transplanting a 0.4 µm -pore PTFE macroencapsulating SC device, a device designed to protect transplanted cells from allorejection, the most effective approach is to place FGF2 outside a preimplanted device that contains FGF2 prior to transplantation.

A 10 µm-pore macroencapsulating device, a device potentially useful for local immunotherapy, was the most successful method for supporting SC human islet transplantation and allowed rapid normalization of blood glucose that continued for 300 days. The aforementioned approaches using scaffolds and devices will be helpful for successful studies of SC human islet transplantation.

Author Contributions

Douglas Sobel: Conceptualization, writing – original draft. Wanxing Cui: Conceptualization, writing – review and editing, providing islets.

Competing Interests

The authors have declared that no competing interests exist.