Development of New Macroencapsulating Planar Devices to Inhibit Allorejection of Islet Transformed Cells

Douglas O. Sobel ^*, Keerat Parmar

1. Georgetown University Medical Center, Department of Pediatrics, Division of Endocrinology, Washington, D.C., USA

* Correspondence: Douglas O. Sobel

Academic Editor: Chirag S. Desai

Received: August 31, 2025 | Accepted: March 16, 2026 | Published: April 14, 2026

OBM Transplantation 2026, Volume 10, Issue 2, doi:10.21926/obm.transplant.2602269

Recommended citation: Sobel DO, Parmar K. Development of New Macroencapsulating Planar Devices to Inhibit Allorejection of Islet Transformed Cells. OBM Transplantation 2026; 10(2): 269; doi:10.21926/obm.transplant.2602269.

© 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 1 diabetes mellitus is caused by the autoimmune destruction of the islets of the pancreas, resulting in insulin deficiency [1,2]. This immune pathway is a T-cell-mediated process that includes T-cell recognition and activation, followed by cell-mediated and cytokine-mediated β-cell damage [3].

Despite advances in blood glucose control for people with Type 1 diabetes, the rates of severe diabetic complications, morbidity, and mortality remain high [4,5].

Much work has focused on replacing the destroyed islets to cure the disease through pancreas and pancreatic islet transplantation.

Many poorly controlled patients who undergo whole-pancreas transplantation develop better glycemic control but are usually not free of insulin therapy. Furthermore, the major surgery of transplantation can cause significant side effects, and these patients must be on toxic long-term immunosuppressive drug therapy, which can lead to major organ dysfunction, anemia, and cancer, as well as increased morbidity and mortality from major surgery [6].

Success with animal studies transplanted islets into the portal vein has led to research to transplanting islets in humans [7,8]. Although, the efficacy of human islet transplantation has improved over time, a 10 years study by the GRAGIL network showed that even after multiple islet transplants, only 4.8% of patients are insulin-free even after multiple islet transplants [9].

Further, there are many potential complications with the current procedure of transplanting islets into the portal vein. These complications include surgical and postoperative complications of portal thrombosis, infection, bleeding, and the grave issues of the needed generalized immunosuppression therapy, all of which do not allow it to be the standard of care for Type 1 diabetes [10,11]. Because of the high risk-benefit ratio, islet transplantation is only approved for a specific group of patients with dangerous hypoglycemia unawareness [12].

To avoid immunosuppressive therapy, research has focused on transplanting microencapsulated islets within small-pore macroencapsulated devices. Microencapsulation refers to the encapsulation of a small number of islets within a matrix such as alginate, where the encapsulate is designed to ward off immune attack while still allowing the important flow of nutrients and insulin [13].

There has been some success in transplanting microencapsulated islets in animal studies [14]. In human trials, long-term insulin independence has not been achieved [15].

Transplantation of islets within devices constructed with small-pore membranes that provide a physical barrier to immune cell passage has been shown to inhibit allograft rejection in animals [16]. However, transplantation of such devices has not been successful in humans [17]. Failures in human transplantation are largely due to the sensitivity of islets to hypoxia and the hypoxic conditions at the transplant site [18,19] as well as hypoxic cell death of transplanted cells, which appears related to the fibrotic response to the device [20,21].

Ways to improve encapsulation methods using various biomaterials, and ways to better immunoprotect and provide oxygenation to the transplant have been reviewed [22,23,24,25].

We hypothesize that alterations in device membrane characteristics may improve the properties of the device to inhibit allorejection.

A model to study the inhibition of allorejection of transplanted macroencapsulated devices using an insulin-secreting cell line would be of great value. NIT-1 cells have been used in microencapsulation but failed to normalize blood glucose [26,27]. However, we recently characterized a mouse model of transplantation using MIN-6 cells, a transformed mouse insulin-producing cell line [28,29]. We hypothesize that MIN-6 cells can be used as transplanted cells to further develop macroencapsulation devices that provide a physical barrier to immune cells.

We assessed various membrane pore sizes and types of membranes, and the effect of membrane hydrophilicity on macroencapsulating devices to inhibit allorejection. Furthermore, the development of a multilayered device to limit the required skin surface area and augment the islet load is reported.

2. Materials and Methods

Hydrophilic polytetrafluoroethylene (PTFE-HP), lipophilic polytetrafluoroethylene (PTFE-LP), ethylenetetrafluoroethylene (ETFE), fluorinated ethylene propylene (FEP), polyvinylidene fluoride (PVDF), nylon, and polyethersulfone (PES) membranes were obtained from Sumitomo Electric Interconnect Products (San Marcos, CA), Welch Fluorocarbon (West Haven, CT, USA), Dupont (Wilmington, DE), Advantec MFS (Dublin, CA), and CS Hyde (Lake Villa, IL).

2.1 Animals

The animal protocols and procedures were approved by the Institutional Animal Care and Use Committee of Georgetown University. Athymic nude mice and NOD mice were obtained from Taconic Farms (Germantown, NY) and Charles River respectively. All mice were females and Animals were 6 to 8 weeks old at the initiation of the experiments. All animals were housed in the specific pathogen-free environment at Georgetown University. The mice were fed standard rodent chow diet. MIN-6 sensitized mice were generated by SC implanting NOD mice with 1 × 10⁶ MIN-6 cells within a 10 μm pore-sized PTFE device.

2.2 Preparation of Membrane Devices

Single chamber devices: Polymer membrane sheets were cut into strips, folded on itself, and 2 sides were then heat sealed to make a final dimension of 17 mm (L) × 7 mm (W). The integrity of the membrane device was checked with PBS to ensure that there were no leaks. The devices were later loaded with MIN6 cells within a cell matrix at the time of the implant and the fourth side was then heat sealed to create an intact membrane device (Figure 1).

Figure 1 Single chamber macroencapsulation device constructed with PTFE. (a) Construction of device: 1. Polymer membrane sheets are cut, 2. Membrane is folded on itself, 3. Two open sides are heat sealed, 4. Cells are loaded into device, 5. Last open side is heat sealed. (b) Photograph of device.

A multi-chamber device was constructed using single PTFE-HP membrane units. The units were stacked on top of one another, parallel to the skin surface, and then encased within a larger 10 μm PTFE-HP membrane unit. Silicone provided device support (in red) (Figure 2).

Figure 2 Multichamber macroencapsulation device constructed with three PTFE-HP units stacked on top of each other and encased within a larger 10 μm PTFE-HP membrane unit. Silicone provided device support (in red).

The devices were sterilized by gamma irradiation (25 kGy) and then loaded with MIN-6 cells suspended (80% v/v) in Matrigel (BD Catalog# 354230) at the time of implantation (106 cells/0.12 ml). The fourth side was heat-sealed to create an intact membrane device.

2.3 Surgical Implantation Procedure of Macroencapsulation Devices

Mice were anesthetized in a chamber with 3% isofluorane and 4.5 LPM O₂ at 10 bar pressure and 70°F (21°C) and then placed on a sterile heated cushion. The surgical site was shaved and cleaned with alcohol and Betadine. An incision was made in the skin on the dorsal side of the mice at the right thoracic mammary gland. The skin was retracted and a subcutaneous pocket was created by dissection. The device was inserted and then the incision was closed with sterile staples or suture.

2.4 Direct Subcutaneous MIN-6 Cell Transplantation

Without a device, MIN-6 cell suspension in Matrigel was injected SC into the ventral (right flank) or dorsal region of animals using a 22G needle.

2.4.1 MIN-6 Cell Culture

MIN-6 cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) with 25 mM glucose, supplemented with 15% heat-inactivated fetal bovine serum, 1% Pen/Strep, 2 mM L-glutamine, and 70 µM 2-mercaptoethanol. Subculture and maintenance were performed as previously described [29]. All assays used MIN-6 cells grown to 70-80% confluence.

2.4.2 Induction and Assessment of Diabetes

To induce diabetes, streptozotocin (Zanosar^®, NDC# 00703-4636-0) obtained from Teva Parenteral Medicines, Inc., was dissolved in sodium-citrate buffer (pH 4.5) and administered intraperitoneally at 300 mg/kg [30].

2.4.3 Assessment and Diagnosis of Diabetes

Blood glucose was tested by home glucometer test strips. Animals were diagnosed with diabetes when blood glucose levels remained above 14 mM for two consecutive days.

2.4.4 Diabetic Animal Maintenance

Diabetic mice were inject sc with 0.5 U Lantus daily or injected sc with 0.5 U of Degludec insulin every 42 hours when blood glucose is over 350 mg/kg.

2.4.5 IP Glucose Tolerance Test (IPGTT)

After a 4-hour fast, a baseline blood glucose assessment was performed. Following IP glucose administration (200 mg glucose/kg body wt.), blood glucose was measured at 30, 60, 90, and 120 minutes. The rate of glucose disappearance or clearance per minute from 30 to 90 minutes was calculated by the k score.

k = ln(G₁) - ln(G₂) divided by (T₂ - T₁) × 100% where G₁ and G₂ are glucose levels at Time 1 and Time 2.

2.4.6 Scoring of Fibrosis of Device

1+ no fibrosis; 2+ very little fibrosis, no sheath around the device; 3+ clear sheath around the device; 4+ thick fibrotic sheath around the device.

2.5 Statistical Analyses

All statistical analyses were performed using Microsoft Excel and SPSS 25.0 (SPSS, Inc., Chicago, IL). Statistical differences were evaluated using Student’s t-test or One-way ANOVA. When One-way ANOVA was used to compare means between groups, Tukey’s HSD was used to assess statistically significant differences. All error bars represent standard deviations unless otherwise noted.

3. Results

Effect of MIN-6 cells within 1-2 μm pore PTFE-HP membrane transplant devices to lower blood glucose in diabetic nude and diabetic NOD mice Initial studies assessed the ability of a 1-2 μm pore PTFE membrane device to inhibit allorejection (Figure 3). MIN-6 cells (10⁶ cells/device) within a 1-2 μm pore PTFE device were SC transplanted into 3 diabetic NOD mice and 2 control diabetic nude mice (n = 2).

Figure 3 Effect of MIN-6 cells transplanted with in 1-2 μm pore PTFE transplant device. MIN-6 cells (10⁶ cells/device) were SC transplanted within 1-2 μm pore PTFE devices into diabetic NOD mice (◊) (n = 3) and into 2 diabetic nude mice (■) (n = 2). Nude and NOD transplanted mice developed lower mean blood glucose from baseline from Day 5 and 21 respectively until the devices were removed at Day 80. Mean blood glucoses were compared by Student’s T-test.

Mean blood glucose in diabetic nude mice was lower than baseline from Day 5 to Day 80, when devices were removed (p < 0.01). Blood glucose decreased to 70 mg% by Day 40 and remained stable thereafter. Blood glucose rose to diabetic levels after device removal at Day 80.

Diabetic NOD mice transplanted with 1-2 μm pore-size PTFE devices had lower mean blood glucose than baseline from Day 21 to Day 80 (p < 0.01), but it did not fall below 200 mg/dl. At Day 80, the devices were removed, and mean blood glucose rose sharply in all mice.

3.1 Effect of 0.4 μm and 10 μm Pore Transplant PTFE Devices on Short Term Inhibition of Allorejection

The ability of 0.4 μm and 10 μm pore PTFE devices to inhibit allorejection of MIN-6 cells was assessed (Figure 4a).

Figure 4 (a) Effect of using PTFE devices with 0.4 μm and 10 μm pores on the allorejection of transplanted MIN-6 cells. MIN-6 cells were transplanted directly SC (0) in nude mice and transplanted within PTFE devices with 0.4 μm pores (◊) or 10 μm pores (x) in NOD mice. Diabetic control mice (+) received no MIN-6 cells. At 32 days, the 0.4 μm pore devices were removed. At each time point, mean blood glucose were compared using ANOVA with Tukey’s HSD. *p < 0.05 vs mice not transplanted with MIN-6 cells and mice transplanted with MIN-6 cells in 10 μm pore PTFE devices. (b) IPGTT at 24 days of NOD mice not receiving MIN-6 cells (O) and of NOD mice transplanted with MIN-6 cells within PTFE devices constructed with 0.4 μm pores (◊) or 10 μm pores (X) in NOD mice. *p < 0.05 vs not transplanted MIN-6 cells and transplanted MIN-6 cells in 10 μm pores PTFE devices.

The control diabetic nude mice group (n = 2), which assessed MIN-6 cell viability, developed hypoglycemia by 2 weeks, as expected. The mean blood glucose levels of control diabetic NOD mice that received no MIN-6 cells did not change and remained over 400 mg/dl.

Diabetic NOD mice transplanted with MIN-6 cells within a 10 μm pore PTFE device (n = 4) had no change in mean blood glucose levels following MIN-6 cell transplantation and continued to have blood glucose levels over 400 mg% similar to control mice that did not receive any MIN-6 cells.

Diabetic mice transplanted with 0.4 μm pore PTFE devices (n = 4) developed euglycemia at 2 weeks. Their mean blood glucose was significantly (p < 0.01) lower than that of mice transplanted with no MIN-6 cells and of mice transplanted with MIN-6 cells within a 10 μm device from 3 days through the end of the experiment at 32 days. Removal of 0.4 μm devices containing MIN-6 cells resulted in blood glucose over 400 mg%, no different from control diabetic mice.

3.2 Intraperitoneal Glucose Tolerance Test (IPGTT)

At 32 days, an IPGTT was performed in all groups of mice. Control mice transplanted with no MIN-6 cells and those transplanted with MIN-6 cells within 10 μm PTFE devices had similar mean blood glucose levels following IP glucose infusion (Figure 4b). The mean blood glucose reached the maximum detectable level (600 mg%) at all time points after glucose infusion. NOD mice transplanted with MIN-6 cells within a 0.4 μm device had significantly lower mean blood glucose at every time point compared with control NOD mice. Blood glucose returned to under 150 mg% by 120 minutes.

3.3 Effect of Membrane Hydrophilicity on Long-Term Inhibition of Allorejection

The effect of the lipophilicity of the membrane device on its ability to provide long-term inhibition of allograft rejection was assessed in mice transplanted with either a more lipophilic PTFE membrane (PTFE-LP) or a hydrophilic PTFE membrane (PTFE-HP) device.

Diabetic mice were transplanted with PTFE-HP devices (n = 4) or PTFE-HP devices (n = 4) and assessed for basal blood glucose and glucose responses to a glucose challenge (IPGTT) at 32, 60, 92, and 156 days post-transplant (Figures 5a, 5b).

Figure 5 (a) Effect of hydrophilicity on the membrane device function. Mean basal blood glucose levels of NOD mice (N = 4 per group) transplanted with MIN-6 cells within devices constructed with either PTFE-HP (▲) or PTFE-LP (♦) membranes. At Day 100, membranes from two PTFE-HP (Δ) and two PTFE-LP (◊) devices were punctured. At each time point, mean blood glucoses were compared by Student’s T-test. *p < 0.05 vs mice transplanted with PTFE-LP devices. (b) Mean blood glucose during IPGTT in NOD mice (N = 4/group) transplanted with HP (▲) and LP (●) PTFE devices membranes with 0.4 μm pores at Days 32, 60, 92, and 156 post transplant. At each time point, mean blood glucoses were compared with Student’s T test. *p < 0.05 vs mice transplanted with PTFE-LP devices.

The mean time for blood glucose to fall below 200 mg% in HP and LP devices was 25 days and 35 days, respectively. Mean blood glucose levels were lower (p < 0.05) in mice transplanted with HP devices from Day 11 to Day 47, but later were overall similar to those in PTFE-LP-transplanted mice. At Day 102, one PTFE-HP and one PTFE-LP device were removed, and within 1-2 days the mice's blood glucose rose to above 400 mg% in both mice.

Following the IP glucose challenge, mean peak blood glucose at 30 minutes and blood glucose at two to four time points after the peak were lower in mice transplanted with the PTFE-HP device than with the PTFE-LP device on all days tested (Figure 5b).

Mice transplanted with PTFE-HP devices had similar peak blood glucose levels following IP glucose infusion at each day of IPGTT testing. Peak mean blood glucose ranged from 255 to 290 with no specific trend noted over time. Likewise, the peak mean blood glucoses following IP glucose in mice transplanted with PTFE-LP devices were overall similar at each day of IPGTT testing with mean blood glucose ranging from 380 to 430 mg% with no trend over time.

The glucose clearance k (SD) values for mice with devices with PTFE-HP and PTFE-LP at 32 days were 1.03 (0.09)% and 1.27 (0.11)%; at 60 days were 1.01 (0.05)% and 1.14 (0.08)%; at 122 days were 1.09 (0.09)% and 1.28 (0.02)%; and at 156 days were 1.23 (0.04)% and 1.26 (0.05)%. There was no significant difference at each timepoint.

3.4 Comparison of Different Polymer Membranes Devices to Inhibit Allorejection

We compared the ability of 0.4 μm pore devices constructed with different polymer membranes to lower blood glucose and inhibit allograft rejection in transplanted diabetic NOD mice. Diabetic NOD mice (n = 4) were transplanted with MIN-6 cells within 0.4 μm pore devices constructed with PES, ETFE, PTFE-HP, nylon, FVDF, and FEP. Blood glucose levels were then monitored over time, and an IPGTT was performed at 42 days post-transplant.

During the first week, the mean blood glucose levels were lowest in mice transplanted with nylon and PTFE-HP membrane devices, compared with those transplanted with all other membrane devices (Figure 6a). After the first week, the mean blood glucose was lowest in mice transplanted with nylon and PTFE-HP devices. By the end of the experiment, the mean baseline blood glucose fell below 200 mg% only in mice transplanted with devices constrcted from nylon and PTFE-HP.

Figure 6 (a) The effect of different membranes composing 0.4 μm pore transplant devices to inhibit allograft rejection. Mean blood glucose were assessed in diabetic NOD mice (n = 4) transplanted with MIN-6 cells contained within devices constructed with PES (+), ETFE (□), PTFE-HP (-), nylon (o), PVDF (x), and FEP (Δ). At each time point, mean blood glucose levels were compared using ANOVA with Tukey’s HSD. *p < 0.05 vs PES, ETFE, PVDF, and FEP devices. ^#p < 0.05 vs PES, ETFE, PTFE-HP, PVDF, and FEP devices. ^p < 0.05 vs PES, ETFE, nylon, PVDF, and FEP devices. (b) IPGTT at 42 days in diabetic NOD mice (n = 4) transplanted with MIN-6 cells within devices constructed of PES (+), ETFE (□), PTFE-HP (-), nylon (o), PVDF (X), and FEP (Δ). At each time point, mean blood glucose levels were compared using ANOVA with Tukey’s HSD. ^#p < 0.05 vs PES, ETFE, PTFE-HP, PVDF, and FEP devices. ^p < 0.05 vs PES, ETFE, PTFE-HP, PVDF, and FEP devices.

Transplanted membrane devices ranked from lowest to highest basal mean blood glucose at the end of the experiment were as follows: nylon > PTFE-HP > PES = ETFE = PVDF > FEP.

At 42 days, the blood glucose response to the IPGTT was compared. Peak blood glucose in device-transplanted mice from lowest to highest were as follows: nylon < PTFE-HP < PVDF = PES < ETFE < FEP (Figure 6b). At 120 minutes, the mean blood glucose returned to baseline in all groups. The glucose clearance values K (SD) in mic with devices constructed with PVDF, FEP, ETFE, nylon, PES, and PTFE were 1.6 (0.2), 1.59 (0.26), 1.59 (0.25), 1.35 (0.25), 1.64 (0.16), 1.53 (0.28) and were not significantly different from each other.

3.5 Fibrosis Scores of 0.4 μm Membrane Devices

All membrane devices induced a fibrotic response at 42 days. The lowest degree of fibrosis occurred in mice transplanted with PTFE-HP and nylon devices, and the greatest amount of fibrosis was found in PES-constructed devices (Figure 7). The fibrosis scores were significantly lowest in nylon and PTFE-HP devices. The fibrosis scores, from lowest to highest, were Nylon < PTFE-HP < ETFE < PDVF < PES < FEP.

Figure 7 The mean (SD) fibrosis scores surrounding 0.4 μm pore membrane devices at 42 days post transplanted with MIN-6 cells into NOD mice. Mean fibrosis scores were compared using ANOVA with Tukey’s HSD. *p < 0.05 vs nylon. #p < 0.05 vs nylon and PTFE-HP. +p < 0.05 vs nylon, PTFE-HP, and ETFE. ^p < 0.05 vs nylon, PTFE-HP, ETFE and PDVF.

The baseline and peak blood glucose during the IPGTT and the time post-transplant for mean blood glucose to decline to 200 mg% both strongly correlated with the degree of fibrosis of the device at 42 days. The r² and p values were respectively 0.838 (p < 0.02), 0.808 (p < 0.03), and 0.854 (p < 0.02).

3.6 Effect of Allo-Sensitization on the Ability of PTFE-HP Device to Inhibit Allorejection

The ability of the 0.4 μm pore PTFE-HP device to inhibit allorejection in allo-presensitized NOD mice was assessed by comparing blood glucose in MIN-6 cell-sensitized (n = 5) and non-sensitized (n = 4) diabetic NOD mice transplanted with a 0.4 μm pore PTFE-HP device containing MIN-6 cells (Figure 8). Control groups of diabetic mice not transplanted with cells and of mice transplanted with MIN-6 cells using a 10 μm-sized device remained hyperglycemic throughout the study. After transplantation, mean blood glucose in sensitized and non-sensitized mice decreased similarly to non-diabetic levels over time. At the end of the experiment, the transplant devices were removed, and blood glucose levels rose to those observed in non-transplanted mice.

Figure 8 Effect of allo-sensitization of mice on 0.4 μm pore PTFE-HP device inhibition of allorejection. Mean blood glucose of allo-sensitized (◊) and non-sensitized (□) diabetic NOD mice transplanted 10⁶ MIN-6 cells within 0.4 μm pore PTFE-HP devices. A control group of diabetic mice not transplanted with MIN-6 cells (o) and another control group of mice transplanted with MIN-6 cells within a 10 μm sized device (●). Devices were removed at Day 48. Mean blood glucose levels in device transplanted sensitized and non-sensitized mice were compared with Students T-test.

The ability of a multichambered 0.4 μm PTFE-HP device to serve as a transplant device was compared with that of a single-chamber PTFE-HP device.

Single-chamber and multi-chambered sandwich planar 0.4 μm pore PTFE devices (Figure 9) were filled with 1 × 10⁶ cells, and SC were transplanted into diabetic nude mice (n = 4). Control groups included one group of diabetic mice that received no MIN-6 cells and another control group that received SC-administered MIN-6 cells without a device (n = 4). Following the transplant, mean blood glucose levels of mice transplanted with the single- and three-chamber sandwich devices were not significantly different (p < 0.05) at any time point (Figure 8).

Figure 9 Comparison of a single-chambered and a multichambered 0.4 μm PTFE-HP construct as a transplant device. Single-chambered (Δ) and multichambered sandwich (◊) PTFE devices containing 106 MIN-6 cells were transplanted into diabetic nude mice (n = 4). As controls, one group of diabetic mice received no MIN-6 cells (+), and another group received SC MIN-6 cells without a device (X) (n = 4). At each time point, mean blood glucose levels were compared using ANOVA with Tukey’s HSD.

4. Discussion

To improve macroencapsulation device protection against allorejection, we characterized planar transplant devices using MIN-6 cells and explored various device parameters affecting the ability to inhibit allorejection.

Because differences in specific membrane material, membrane tortuosity, geometry and topography, charge, and hydrophilicity could alter immune cell passage, we assessed the effect of different pore size devices on inhibiting allorejection in our specific planar macroencapsulation device [31].

We explored the largest pore size in a PTFE-HP device that would inhibit immune cell passage and allorejection of the transplanted cells. Transplantation of MIN-6 cells within the 10 μm pore PTFE device provided no protection against allograft rejection. This was not surprising, as immune cells readily traverse pores of this size. Thus, mice receiving this 10 μm pore device later served as a negative control. The 1-2 μm pore PTFE-HP device only partially inhibited allorejection, with blood glucose levels in transplanted mice declining but not reaching euglycemic values. Inadequate viability of the transplanted MIN-6 cells was not the cause of this partial protection, as SC administration of MIN-6 cells without a device normalized blood glucose levels in control nude mice. Thus, a smaller pore size may be required, despite immune cell diameters exceeding 4 μm [32].

The 0.4 μm pore transplant device provided complete immune protection against allograft rejection. In fact, all mice developed and remained euglycemic. Furthermore, IPGTT testing showed good glucose clearance, with normoglycemia achieved by 120 minutes. The MIN-6 cells in the device were clearly responsible for the glucose-lowering effect, as mice became hyperglycemic to control levels when the 0.4 μm transplant was removed. Thus, the smallest pore device, the 0.4 μm membrane pore device, was required to fully provide immune isolation and reverse diabetes in the transplanted mice. These data further demonstrated that MIN-6 cells are a feasible model for studying 0.4 μm PTFE planar devices to prevent allograft rejection.

It is important to note that factors other than pore size may also play a role in immune cell transfer and, in turn, allorejection., Liu reported that a nylon membrane with a pore size as small as 0.67 μm only partially inhibited cell penetrationand that a 0.38 μm pore is required to fully inhibit penetration, while pores as large as 1-1.25 μm in another membrane with a different geometry blocked T lymphocytes [20,33].

The requirement for a 0.4 μm pore device may have negative or positive ramifications for its ability to serve as a transplant device. First, different membrane pore sizes may hinder the mass transport of molecules important for the survival and function of the transplant. Slow oxygen diffusion from the SC space into the transplanted device has been a major limiting factor for islet viability [19]. Slowing the influx of glucose into the device and the efflux of insulin from within the device could potentially impede the transplant's ability to provide good glycemic control.

However, because oxygen, insulin, and glucose molecules are relatively small, with hydrodynamic radii of only 0.15 nm, 3 nm, and 0.4 nm, respectively, compared with even the smallest pore in a membrane of 0.4 μm (400 nm), their diffusion would not be significantly altered by pore size alone, given equal porosity (or space of pores/surface area), tortuosity, and thickness of the membrane. This lack of alteration in diffusion with the smallest-pore membrane would not be the case if molecules were not in liquid and traversing a liquid-soaked membrane, as in our case. Without the liquid, diffusion would be impaired even for oxygen through a 0.4 μm pore due to Knudsen diffusion, where molecules would be hindered by molecule collisions with the pores of the membrane [34]. Due to membrane hydrophobicity, insulin has been found to bind to hydrophobic PTFE membranes, an effect that could be more pronounced in smaller-pore [35] 0.4 μm membranes. However, this effect is not expected in the ‘hydrophilic’ PTFE membranes described in this report.

Restricting the molecular diffusion of the following immune effectors (hydrodynamic radii) involved in transplant cell rejection, such as IgG (5.2 nm), IgM (12.7 nm), and cytokines (IL-1β 2.3 nm; IL-2 2.2 nm; TNF 2.2 nm; gamma interferon 5 nm), could be advantageous in inhibiting transplant rejection. However, because these effectors are so small, there would be no significant hindrance even with the smallest pore (0.4 μm) at equal porosity, tortuosity, and thickness.

The viability of cells in membranes devices with different pore sizes will depend on several factors which include: the need for immune isolation, factors allowing for the flow of nutrients and oxygen such as membrane porosity, and cell adhesion to membranes with pores of different sizes since cells generally grow best on membranes with smaller pores (<1 μm) [36]. We compared the functional viability of MIN-6 cells in devices with different pore sizes by examining the ability of the transplant to lower blood glucose in diabetic NOD mice. As mentioned, functional viability was best with the 0.4 μm pore device, which appears to be due to the requirement for the smallest pore, the 0.4 pore, to inhibit immune cell passage and transplant rejection. Further experiments will explore the effect of poor size on cell viability without the confounding factor of immune isolation using nude mice.

The lipophilic and hydrophilic PTFE devices provided very long-term protection for up to 250 days. When assessing the effect of the hydrophilicity of the device membrane on supporting the function of encapsulated MIN-6 cells and inhibiting allorejection, the hydrophilic device was superior. The mean basal blood glucose of mice transplanted within hydrophilic PTFE membrane devices was initially lower than that of mice transplanted with lipophilic PTFE devices. The differences between membranes were more apparent after the IPGTT testing. Blood glucose levels were lower at 30, 60, and 90 minutes in mice transplanted with hydrophilic devices at all time periods, even at 156 days. In both the hydrophilic and hydrophobic PTFE device-transplanted groups, MIN-6 cell function during IPGTT testing from 32 to 156 days showed remarkable consistency over time. This long-term effect is of great interest not only for the validity of this model utilizing MIN-6 cells but also for demonstrating that the loss of insulin-producing capacity of transformed cells, such as MIN-6 cells, that occurs in vitro did not occur over the long period of time in vivo when MIN-6 cells were transplanted in devices [37,38].

Clearance rates were similar in the HP and LP groups, but insufficient to lower blood glucose to baseline in the LP group at Day 32. This impaired glucose handling, resulting in higher blood glucose at 30, 60, and 90 minutes despite good clearance, may be due to a decreased amount of early-phase insulin release as blood glucose initially rises post-glucose infusion [39].

It is unclear why the mice transplanted with hydrophilic devices handled a glucose load (IPGTT) better than those with lipophilic devices. The difference may be due to the better cell support observed with many cells grown on hydrophilic surfaces, the augmented oxygen diffusion that may occur through hydrophilic membranes, or positive interactions of the surface of the HP device and the subcutaneous space or cell matrix [40,41,42,43].

To examine the effect of different polymer materials on the ability of devices to prevent allorejection, we compared devices constructed with nylon (a polyamide polymer), PTFE, ETFE (a similar material to PTFE), FEP, PVDF, and PES, all of which have previously been shown to support cell adherence and/or cell culture [44,45,46,47,48,49].

We found that transplanting devices constructed with nylon polymer and PTFE-HP membranes more effectively inhibited allorejection, with the nylon device performing best. Blood glucose decreased faster and fell below 200 mg% only in mice transplanted with the nylon and PTFE-HP devices. The effectiveness of the different membrane devices in inhibiting allorecognition, as determined by both the lowest mean blood glucose at the end of the experiment and the lowest peak blood glucose during IPGTT, from best to least effective was nylon > PTFE-HP > PVDF = PES > ETFE > FEP.

The degree of fibrosis in the devices at the end of the experiment appeared to depend on the membrane material used to construct the devices. The significant inverse correlation between the fibrosis score of the implanted device at the end of the study and the time post-transplant for the blood glucose to decline to 200 mg, as well as the baseline and peak blood glucose during the IPGTT at 42 days, supports the hypothesis that the fibrotic process could be related to the ability of the membranes to function and the viability of the MIN-6 cells.

This is consistent with previous studies suggesting that the fibrotic foreign-body response around the device and resultant cellular hypoxia are important reasons why macroencapsulation devices have failed [19,20]. Future studies exploring techniques to decrease the fibrotic response, such as coating the devices, may further improve transplant viability [20,50,51].

Other possible factors contributing to the differential ability of membrane devices to regulate blood glucose and inhibit allograft rejection include solute membrane permeability, cell surface chemistry, polymer structure (such as tortuosity, geometry, and charge), and differential adhesion of the membrane to matrix or interstitial proteins [52,53,54].

People receiving second allotransplants have a higher risk of allorejection, which appears to be due to allosensitization from the previous transplant(s), including islet transplant, or from previous blood transfusion or pregnancy [55,56,57].

Thus, a method to successfully transplant allogeneic tissue in sensitized subjects would be invaluable. We found that, even in allosensitized mice, transplantation of MIN-6 cells into 0.4 μm PTFE-HP devices inhibited allorejection. In fact, the device was equally effective in preventing allorejection in sensitized and nonsensitized mice, underscoring the value of such technology.

These findings also support the contention that a cell-mediated process plays the primary role in allorejection in both allosensitized and nonsensitized mice. Humoral immunity appears to play no direct role, since IgG and IgM molecules can easily pass through due to their relatively small diameters of approximately 14.2 nm and 35 nm, respectively.

A major obstacle to the use of subcutaneous planar devices to prevent islet allorejection in humans is the large volume of islet cells required, which would necessitate an excessive amount of skin surface area, since planar devices must be thin to allow sufficient oxygen transfer through the membrane [58]. To address this issue, we constructed and tested a sandwich, multichambered, planar 0.4 μm pore PTFE device containing MIN-6 cells and found that this double-layered multichambered device decreased blood sugar levels, as did the single planar device. This suggests the device provides sufficient oxygenation and may be useful in reducing the excessive skin surface required for transplanting planar devices.

There are potential risks to long-term device implantation, such as membrane degradation, indirect activation of immune cells, and fibrosis even though PTFE and nylon are known to be highly inert and resistant to degradation [59]. The occurrence of fibrosis surrounding the device could limit oxygen diffusion into the device and result in hypoxic death to the transplant. If immune activation and fibrosis became an issue, coping strategies that can be considered include: adding substances to tor on the device that have been shown to inhibit fibrosis or inflammation and altering the additives within the polymer device fabrication [60].

In conclusion, this is the first report describing a nylon macroencapsulating device that inhibits allorejection. Furthermore, the 0.4 μm pore nylon device and, secondly, the PTFE-HP membrane devices were the most effective at inhibiting allorejection and at inducing the least fibrotic response. The protection was robust enough to inhibit allorejection in allosensitized mice. Our development of multi-chambered devices will enable increased islet load by reducing the required skin surface area. Furthermore, the MIN-6 cell has been demonstrated to be useful for studying allograft rejection of transplanted macroencapsulation devices. To support the translational relevance of this work, we have demonstrated that human islets transplanted in similar PTFE devices can reverse diabetes in nude mice [61]. Future studies with nylon membranes are warranted.

Author Contributions

Douglas Sobel: conceptualization; methods; writing, project administrator; funding; statistics. Keerat Parmar: conducted research; methodology, review manuscript; help edit Revision.

Competing Interests

The authors have declared that no competing interests exist.