Cellular transplantation in humans has great potential as a viable therapeutic option for a number of hormone deficiency diseases (1). Successes in transplantation of human pancreatic islets for type 1 diabetics (2–4) have generated considerable enthusiasm. However, to date the transplantation of pancreatic islets has required the use of immunosuppressive drugs that expose the patient to serious side effects (5).
An alternative approach is to enclose cells in a semipermeable membrane that protects them from immune system attack and permits the influx of molecules and nutrients necessary for cell function and efflux of the desired cellular product, specifically insulin (6–8), making recipients healthy. Cellular encapsulation has two major potential benefits: 1) transplantation without the need for immunosuppressive drugs, and 2) use of cells from a variety of sources such as allograft (either primary cells or stem cell derived), xenografts (porcine cells or others), or genetically engineered cells (9–12). Potential cell types for immunoisolation include pancreatic islets, hepatocytes, neurons, parathyroid cells, and cells secreting clotting factors.
A perfect immunoisolation system is a membrane that can keep out all high molecular weight immune system components such as immunoglobulin M, while allowing low molecular weight oxygen, nutrients, and hormones to pass through without difficulty. However, the majority of immunoisolation systems are polymeric membranes. Polymeric membranes are random network systems with nonuniform pore sizes. This compromises the immunoprotection function of the membrane. The polymeric encapsulation system was successful in small animal trials, but was less than satisfactory in canine models (13, 14) and in human studies (8, 15). In the past, those experiments have always had to apply some modulating and/or immunosuppressive agents in canine and human transplantation experiments, which complicated the interpretation of results. In large animal transplantation experiments, the possibility of islet regeneration has also been a concern. The result of those complications is the lack of follow-up canine experiments reported in the literature. Successes in large animal transplantation experiments were limited (16–19).
This paper describes a novel immunoisolation system and the successful canine allotransplantation experiments. Transplantation of islets encapsulated by means of the novel system does not require any immunosuppression therapy and thus offers the promise of developing a bioartificial pancreas for type 1 diabetes patients.
MATERIALS AND METHODS
A new encapsulation system of sodium alginate (SA), CaCl2, polymethylene-co-guanidine (PMCG), cellulose sulfate (CS), and poly l-lysine (PLL) was developed for large animal trials. This new system built on the PMCG-CS/CaCl2-SA capsule (20–24) which was successful in nonobese diabetic (NOD) mice trials (25), but not in large animals. To improve the performance, a thin interwoven PMCG-CS/PLL-SA membrane was fused onto the PMCG-CS/CaCl2-SA capsule, forming permanent bonds. This union improves the immunoprotection function without jeopardizing the mass transport of nutrients and insulin. To shield the PMCG and PLL on the surface of the capsule, a third (outer) membrane of CaCl2-SA (12) was added to encase the system.
Capsule Development and Fabrication
Our multicomponent and multimembrane hybrid capsule design requires multiple processing steps. In each step, simultaneous chemical reactions take place and each successive processing step impacts all previous steps. In addition, polymers vary a great deal from vendor to vendor, from lot to lot, and gradually degrade with time. A variation in processing steps or a minor change in polymers can result in an unpredictable change in capsule performance. This makes our capsule difficult to reproduce using a fixed processing regimen. To overcome this, we used the capsule’s diameter, membrane thickness, mechanical strength, and nanopores for processing control. In general, the capsule fabrication process consists of three steps: capsule formation, performance improvement, and biocompatibility improvement. Each step introduces additional chemical components and reaction times. The capsule characteristics were measured after every step. When a new polymer or a new capsule design is being considered, we modify the capsule fabrication parameters such that the final capsule characteristics meet the desired result.
The capsules’ fabrication apparatus developed in our laboratory has been reported before (26, 27), so only a brief description will be given here. A multiloop reactor filled with cation solution was continuously replenished by a cation stream to carry away the anion drops (SA and CS) being introduced into the chamber. The gelling steps were repeated for additional cation solutions until specific capsule performance requirements were met. The capsules were coated with CA-SA to improve biocompatibility.
The apparent pore size of the capsular membrane was determined by size exclusion chromatography (SEC) that measures the exclusion of dextran solutes from the column packed with microcapsules (28). Using neutral polysaccharide molecular weight standards makes it possible to evaluate the membrane properties under the conditions when solute diffusion is controlled only by its molecular dimension. Based on the measured values of solute size exclusion coefficients and known size of solute molecules, one can estimate the membrane pore size distribution.
Using an apparatus developed in our laboratory, the mechanical strength of capsules was measured by placing an increasing uniaxial load on the capsule until the capsule burst (27). The capsule mechanical strength—a function of chemical bonds and membrane thicknesses—can be adjusted anywhere from a fraction of a gram load to many tens of grams load to meet transplantation requirements.
Promising capsules prior to animal studies were tested in a perifusion apparatus (28) using Roswell Park Memorial Institute (RPMI) 1640 medium with 0.1% bovine serum albumin as a perifusate. The islets were equilibrated with perifusate containing 2 mM glucose at 37°C. On average, ∼100 encapsulated islets were loaded into each chamber of a multichannel perifusion system. The flow rate in each channel was ∼1 ml/min. Samples of perifusate were collected over 3-min periods during a 30-min perifusion of 2 mM glucose, a 30-min perifusion of 2 mM glucose+0.045 mM isobutylmethylxanthine (IBMX), and a 60-min perifusion of 2 mM glucose. Samples were assayed in duplicates for insulin using Coat-a-Count Kits with an insulin standard.
We have studied more than 200 capsule designs and mutations in vitro. The capsules chosen for canine transplantation studies are summarized in the capsule parameter columns of Table 1 in the Results section.
Transplantation and Management
All methods and procedures regarding the use of animals were reviewed and approved by the Vanderbilt Institutional Animal Care and Use Committee.
Pancreatic Islet Isolation and Evaluation
Pancreases were surgically harvested from mongrel canines of either sex weighing 17–30 kg (mean 22.4±0.4). Briefly, the animals were anesthetized and a surgical field created. The peritoneal cavity was opened by a long abdominal incision and the viscera exposed. The pancreas was partly immobilized, taking care not to sever any major blood vessels. The pancreatic duct was identified and cannulated at the duodenum. The animal was euthanized, and the excision of the pancreas completed. The gland was infused with cold modified University of Wisconsin (UW-D) organ preservation solution via the ductal cannula. The average harvested pancreas weight was 52±1.0 gm. The glands were transported on ice to the laboratory where the UW-D solution was replaced by a solution of collagenase in UW-D. The glands were then placed in a shaking water bath and digested at 40°C for up to 30 min. The dissociated tissue was then filtered through a 400-μm screen and washed several times with ice-cold media to remove and inactivate the collagenase. Separation of islets and exocrine tissue was performed on a discontinuous Ficoll gradient. After density centrifugation, the islets were collected, washed several times, and put in tissue culture in M199 media supplemented with 10% fetal bovine serum and antibiotics. Islet purity was assessed with dithizone (29) and was 90% or greater. Islet vital staining was performed with calcein AM and ethidium bromide. Islet quantity was determined as islet equivalents (IEQ) using a formula to convert islet populations of differing sizes to islet volume (30). After losses during islet culture, handling, and encapsulation, the average yield was 80,000 IEQ of encapsulated islets per harvested canine pancreas.
Nine healthy mongrel canines of either sex, 6–11 kg (mean 8.33±1.25), were used as recipients. The canines underwent a total pancreatectomy 21–28 days prior to transplantation in order to render the animals totally insulin deficient. Exogenous purified pork insulin was administered to control hyperglycemia. Average exogenous pork insulin requirements were 0.6 and 1.2 U/kg regular and neutral protamine Hagedorn (NPH), respectively, during the pretransplant period. Pancreatic enzymes to aid in food digestion were administered along with the animals’ daily food ration (31).
Five to seven days prior to islet transplantation, insulin administration was withheld for 36 hr to verify the completeness of the pancreatectomy and absence of circulating insulin. On the day of transplantation, general anesthesia was induced. A 2.5-cm midline incision was performed. Blunt dissection of midline adipose tissue and visualization of the abdominal contents were performed to ensure the intraperitoneal administration of encapsulated islets. Encapsulated islets were administered via a 6.0-Fr cuffed tube over 3–4 min and then the incision was closed. Upon anesthetic recovery (10–15 min), animals were provided their daily food ration; exogenous insulin therapy was discontinued and not reinstated unless the fasting glycemia was greater than 180 mg/dl for three consecutive days. No immunosuppressive or anti-inflammatory therapies were utilized.
Daily Management and Clinical Assessments
The animals’ general physical condition and dietary intake were monitored daily. Venous blood for the determination of glucose and insulin was collected at regular intervals. The animals had intravenous and/or oral glucose tolerance test (0.3 g/kg, 0.7 gm/kg body weight, respectively) performed at monthly intervals. Normal (nonpancreatectomized) animals were used as normal controls for the intravenous glucose tolerance test (IVGTT; n=5) and oral glucose tolerance test (OGTT; n=5) assessments.
Histological Examination and Immunohistochemistry
At the conclusion of the experiments when exogenous insulin requirements returned to pretransplant level, animals were euthanized; omental tissue and encapsulated islets were retrieved and fixed in 10% neutral buffered formalin. Tissue sections were paraffin embedded, trimmed to 5 μm, and placed on charged slides followed by hematoxylin and eosin staining (Richard Allan Scientific, Kalamazoo, MI).
Five-micron sections of paraffin-embedded tissue were placed on charged slides and deparaffinized. The sections were rehydrated and placed in heated citrate target retrieval solution (Labvision, Fremont, CA) for 20 min. Endogenous peroxidase was diminished with 0.3% hydrogen peroxide for 20 min followed by Ultra V block for 5 min (Labvision, Fremont, CA). Sectioned tissues were incubated with guinea pig anti-insulin (Linco Research Inc, St. Charles, MO) diluted 1:5K for 30 min. The Vectastain ABC Elite (Vector Laboratories, Burlingame, CA) system and DAB (DakoCytomation, Carpinteria, CA) were used to produce localized, visible staining. Slides were lightly counterstained with Mayer’s hematoxylin, dehydrated, and cover slipped.
Capsule Pore Size Optimization
Polymeric immunoisolation systems are random network systems. The nanopore size distribution of such membranes is not uniform. As a result, immunoisolation systems using polymeric membranes suffered limited success in large animal studies. Many of these large animal experiments were not reported or could not be reproduced. To overcome this, we developed a new membrane design by fusing the PMCG-CS/PLL-SA membrane onto the PMCG-CS/CaCl2-SA capsule. The strong ionic bonds of the PMCG-CS/PLL-SA system improved the capsule stability and improved nanopore size distribution of the capsule as shown in Figure 1A. This new and more uniform capsule design improved immunoprotection function without jeopardizing capsule mass transport.
Capsule Perifusion Studies
Insulin secretion of encapsulated canine islets was evaluated in a cell perifusion apparatus. The insulin secretion by the encapsulated islets in response to glucose challenge as a function of fraction numbers (or time in minutes) is shown in Figure 1B. The data shows a slight delay in response between the free islets and encapsulated islets. Encapsulated islets harvested 195 days postadministration are viable with signs of central necrosis (data not shown). This is likely due to the fact that nutrient transport by diffusion to the interior of encapsulated islets is inferior to the transport supported by islet vasculature (32). Similar conclusions have been reported by other groups with different capsule designs (11, 33, 34).
Five to seven days prior to transplantation, exogenous insulin was withheld for 36 hr. Measured circulating insulin levels were not detectable by radioimmunoassay and hyperglycemia returned greater than 450 mg/dl. Within 6 hr after islet transplantation, exogenous insulin administration was not required. Animals exhibited average daily fasting glucose levels of 120–165 mg/dl and circulating insulin concentrations of 3–7 μU/ml at 6–18 hr after meal between 64 and 214 days as shown in Table 1A. Data from a representative animal (no. 141) is shown in Figure 2. Three animals (nos. 259, 108, and 172) received encapsulated islet supplementation upon failure of initial transplant; this was effective in providing extended control for 50, 129, and 56 days, respectively, as shown in Table 1B. Figure 3 shows the course of events and data from a representative animal (no. 172) of the animals receiving islet supplements.
Exogenous insulin therapy was reintroduced when failure of the encapsulated islets to maintain fasting glycemia less than 180 mg/dl was evident for three consecutive days. When exogenous insulin requirements had stabilized at pretransplant levels, animals were euthanized and an autopsy performed. All structures within the C-loop of the duodenum, including the duodenum, portal-hylar, and spleno-mesenteric regions, were closely examined and revealed no pancreatic tissue.
IVGTTs and/or OGTTs were performed on all nine animals. Figure 4 depicts the IVGTT results from control animals (n=5) and four transplanted animals. The glucose concentration in controls rose from baseline to 238 mg/dl within 5 min and returned to baseline by 60 min. In transplanted animals, the glucose levels rose from baseline to an average of 235±7 mg/dl within 5 min, returned to baseline by 95 min, and continued to fall to an average of 70±2 mg/dl at 180 min. Similarly, the insulin level in controls rose to 42±9 μU/ml within 7.5 min and returned to baseline by 35 min; in transplanted animals, it rose to 8±2 μU/ml (60% above basal) within 50 min and returned to baseline by 135 min. Glucose and insulin results from animals that received an OGTT had similar response characteristics as IVGTT. In either the IVGTT or OGTT, animals with encapsulated islets did not demonstrate a first-phase insulin release that was observed in the control animals.
There were no complications associated with the encapsulated islets transplanted into the peritoneal cavity of healthy pancreatectomized dogs. There was evidence of omental neovascularization. Capsules were minimally attached but could be rinsed off the omental surface. Encapsulated islets that were mildly adhered to the omentum or unattached and freely “floating” in the abdomen, were clean, and less than 1% of capsules had a slight amount of fibrin and rare mononuclear cells adhering to the capsule (Fig. 5A). There was no involvement of the encapsulated islets with any other organ system in the splanchnic bed.
Histological evaluation of omental sections with embedded capsules revealed some pericapsular fibrosis ranging from minimal to severe (one section; Fig. 5B). Inflammatory cells (neutrophils, macrophages, and lymphocytes) were noted within the omentum and areas of fibrosis. Multinucleated giant cells were rarely present. The inflammatory process was consistent with a foreign body reaction. Inflammatory cells and fibroblasts/fibrocytes were present within a small number of the capsules. Clusters of cells were embedded within the capsular matrix. These cells were consistent with the morphology of pancreatic islet cells and stained positive by insulin immunostaining (Fig. 5C and D.). The integrity of the cells was variable in the sections evaluated. Islet cells ranged from a normal appearance to exhibiting shrunken, pyknotic nuclei, loss of cellular margin, and mineralization.
This paper demonstrates a novel five-component/three-membrane hybrid capsule system that improves the encapsulated islet survivability and function in pancreatectomized canine allotransplantation experiments. The intraperitoneal administration of encapsulated canine islets functioned sufficiently to permit the withdrawal of daily exogenous insulin and provided fasting glycemic control without immunosuppression in all nine animals. Eight of the animals’ transplanted encapsulated islets functioned for 64 to 106 days and one animal’s islets functioned for 214 days. Early pilot experiments from this laboratory indicated that intraperitoneal dosages of approximately 50,000 encapsulated IEQ per kilogram body weight would be required for our encapsulation system to normalize fasting glycemia. Larger doses of 73,000 to 87,000 IEQ/kg extended the length of time of exogenous insulin independence in six out of seven animals. In a limited number of animals, retransplantation of encapsulated islets was performed and was effective in providing fasting glycemic control after the initial transplantation had run its course. The lack of measurable circulating insulin prior to transplant, the return to pretransplant exogenous insulin requirements at the conclusion of the experiment, the effectiveness of retransplantation, and the absence of any identifiable pancreatic tissue at autopsy affirms that the animals were devoid of their own insulin throughout the course of the experiment.
Encapsulated islets were responsive to glucose challenges (IVGTT and OGTT) with circulating glucose concentrations returning to baseline within 2 hr. The absence of an initial insulin response was observed in transplanted animals in both glucose challenge scenarios. Possible reasons for this include impairments in transplanted islet physiology and/or transport dynamics related to the site of administration and/or diffusion controlled mass transport mechanism.
The presence of pericapsular fibrosis in capsules adhered to omental tissue was variable with a majority of the capsules exhibiting minimal fibrosis. Capsule breaching by inflammatory cells and fibroblasts/fibrocytes was present in some of the histological evaluations, but the majority of capsules was intact and showed no sign of breaching. In islets exhibiting normal morphology, immunostaining for cytoplasmic insulin was positive, thus indicating islet viability. In all transplantations, including second transplant scenarios, the circulating plasma insulin concentrations decreased during the first week of transplant suggesting early loss of function in a number of encapsulated islets. These observations are evident in Figures 2 and 3, confirming results of others (14–19). The presence of islets exhibiting pathology, the presence of some degree of pericapsular fibrosis on capsules embedded within the omental tissue, the loss of islet viability in the early transplant period, and the absence of first phase insulin release are all potential contributors for a need to return to exogenous insulin therapy. The degree of contribution to the eventual transplant failure cannot be determined from these experiments.
Our immunoisolation system has demonstrated effectiveness in meeting the dichotomous requirements of cellular transplants with efflux of cellular products and without immunosuppression. The development of this immunoisolation system provides impetus for continued research in the treatment of diabetes as well as other diseases caused by the deficiency of cellular products such as hormones, proteins, and neurotransmitters.
The authors thank Dr. A. Powers, Dr. A. Anilkumar, Dr. Z.Y. Chen, Mr. D. Roth, Dr. D. Maron, Mr. John Seiver, Mr. Eric Wang, Mr. Kenneth Wang, and Harvard Medical School 5P30 DK36836 Specialized Assay Core.
1. Nerem R. The Challenge of Imitating Nature, Principles of Tissue Engineering. 2nd ed. St. Louis: Academic Press; 2000.
2. Shapiro AMJ, Lakey JR, Ryan EA, et al. Islet transplantation
in seven patients with type 1 diabetes
mellitus using a glucocorticoid-free immunosuppressive regimen. N Eng J Med
2000; 343: 230.
3. Ricordi C. Islet transplantation
: A brave new world. Diabetes
2003; 52: 1595.
4. Hering BJ, Ricordi C. Islet transplantation
in type 1 diabetes
: Results, research priorities and reasons for optimism. Graft
1999; 2: 12.
5. Shapiro AMJ, Ricordi C, Hering BJ, et al. International trial of Edmonton Protocol for islet transplantation
. N Engl J Med
2006; 355: 1318.
6. Chang TMS. Semipermeable microcapsules. Science
1964; 146: 524–525.
7. Lim F, Sun AM. Microencapsulated islets
as bioartificial endocrine pancreas. Science
1980; 210: 908.
8. Soon-Shiong P, Heintz RE, Merideth N, et al. Insulin independence in a Type 1 diabetic patient after encapsulated islet transplantation
1994; 343: 950.
9. Aebischer P, Hottinger AF, Deglon N. Cellular xenotransplantation. Nature Med
1999; 5: 852.
10. Kuhtreiber WM, Lanza RP, Chick WL. Secretary function of biohybrid pancreas devices containing isolated porcine islets
. Asaio J
1994; 40: M789.
11. Sun Y, Ma X, Zhou D, et al. Normalization of diabetes
in spontaneously diabetic cynomolgus monkeys by xenografts of microencapsulated porcine islets
without immunosuppressant. J Clin Invest
1996; 98: 1417.
12. Lanza RP, Kuhtreiber WM, Ecker D, et al. Xenotransplantation of porcine and bovine islets
without immunosuppression using uncoated alginate microspheres. Transplantation
1995; 59: 1377.
13. Lanza RP, Ecker DM. Transplantation
using microencapsulation: Studies in diabetic rodents and dogs. J Mol Med
1999; 77: 206.
14. Calafiore R. Transplantation
of minimal volume microcapsules in diabetic high mammalians. Ann NY Acad Sci
1999; 875: 219.
15. Scharp D, Schwartz S, Mulgrew P, et al. Encapsulated human islet allografts: Phase I/II clinical trial. Am Diabetics Meeting
16. Weir GC, Bonner-Weir S. Scientific and political impediments to successful islet transplantation
1997; 46: 1247.
17. Gorka O, Hernandez R, Gascon A, et al. Cell encapsulation
: Promise and progress. Nature Med
2003; 9: 1.
18. De Groot M, Schuurs TA, van Schilfgaarde R. Causes of limited survival of microencapsulated pancreatic islet grafts. J Surgical Res
2004; 121: 141.
19. De Vos P, van Hoogmoed C, van Zanten J, et al. Long-term biocompatibility, chemistry, and function of microencapsulated pancreatic islets
2002; 24: 305.
20. Wang TG, Lanza RP. Bioartificial Pancreas: Principles of Tissue Engineering. 2nd ed. St. Louis: Academic Press; 2000.
21. Wang TG. Microencapsulation Methods. PMCG Capsules
: Methods of Tissue Engineering. St. Louis: Academic Press; 2002.
22. Wang T. New Technologies for Bioartificial Organs, Artificial Organs. Boston: Blackwell Science, 1998.
23. Roth DJ, Jansen ED, Powers AC, Wang TG. A novel method of monitoring response to islet transplantation
: Bioluminescent imaging of an NF-kB transgenic mouse model. Transplantation
2006; 27; 81: 1185.
24. Deng Q, Anilkumar AV, Wang TG. Role of viscosity and surface tension in bubble entrapment during liquid drop impact onto a deep liquid pool. J Fluid Mech
2007; 578: 119.
25. Wang T, Lacik I, Brissova M, et al. An encapsulation
system for the immunoisolation of pancreatic islets
, Nature Biotech
1997; 15: 358.
26. Anilkumar AV, Lacik I, Wang TG. A novel reactor for making uniform capsules
. J Biotech Bioeng
2001; 75: 581.
27. Lacik I, Brissova M, Anilkumar AV, et al. New capsule tailored properties for the encapsulation
of living cells. J Biomed Mat Res
1998; 39: 52.
28. Brissova M, Lacik I, Powers AC, et al. Control and measurement of permeability for design of microcapsule cell delivery system. J Biomed Mat Res
1998; 39: 61.
29. Latif ZA, Noel J, Alejandro R. A simple method of staining fresh and cultured islets
1988; 45: 827.
30. Ricordi C, Hering BJ, London NJM, et al. Islet isolation assessment. In: RG Landes, ed. Pancreatic Islet Cell Transplantation
. 1992; 13: 132.
31. Williams PE, Flakoll PJ, Frexes-Steed M, Abumrad NN. In vivo methods for measurements of amino acid metabolism. In: Nissen S, ed. Regulation of Protein and Amino Acid Metabolism. North Holland: Elsevier, 1990.
32. Brissova M, Fowler M, Wiebe P, et al. Intraislet endothelial cells contribute to revascularization of transplanted pancreatic islets
2004; 53: 1318.
33. De Vos P, Van Straate JFM, Nieuwenhuizen AG, et al. Why do microencapsulated islet grafts fail in the absence of fibrotic overgrowth? Diabetes
1999; 48: 1381.
34. Duvivier-Kali VF, Omer A, Parent RJ, et al. Complete protection of islets
against allorejection and autoimmunity by a simple barium-alginate membrane. Diabetes
2001; 50: 1698.