Over 30,000 cases of Type 1 diabetes are diagnosed yearly in the United States.1 Morbidity is severe and life expectancy is shortened. Whole organ pancreas transplantation and islet cell transplantation of 9,000–12,000 islet equivalents (IEQs)/kg can free patients from the need to take exogenous insulin, reduce diabetic complications, and improve quality of life.1 Although successful, pancreas transplantation is a large operation with a high complication rate. Islet cell transplantation is minimally invasive, but the insulin-free survival rate is only about 15% at 5 years.1 Both procedures require immunosuppressive drugs with their attendant costs and side effects.2,3
The limited number of available donor organs has led to the consideration of xenotransplantation of porcine islet cells (PICs).4 Porcine insulin is almost identical to human insulin and was used to treat diabetes for years. Porcine islet cells and human islets respond similarly to glucose changes. Swine are readily available and can be raised to exclude pathogens and to have predictable immunological characteristics.4–6 Unprotected PICs, however, are quickly destroyed when implanted into discordant species by an instant blood mediated inflammatory reaction, by complement dependent preformed circulating antibodies or by induced antibodies to antigens on the implanted graft, and by cellular rejection.4–8 Approaches to control cell destruction include immunosuppressive drugs to destroy or inhibit alloreactive T cells and induction of immune tolerance by altering the bone marrow or the immune system.8–10 Two commonly used immunosuppressants, sirolimus and tacrolimus, are associated with islet cell toxicity. The reliability, toxicity, and cost effectiveness of methods to induce tolerance remain to be established. In addition to the problem of rejection, there is concern regarding the possible transmission of swine pathogens.11
Immunoisolation of xenografted islets by an encapsulating semipermeable barrier has had some success in experimental animals.12–14 Several strategies have been employed: 1) microencapsulation of individual islets (generally by hydrogels),12–14 2) macroencapsulation of large number of islets by polymer membranes,15–19 and 3) hollow fibers containing alginate-encapsulated cells.20,21 An immunoisolating membrane must allow the rapid diffusion of oxygen and nutrients while barring access to immune cells, immune molecules, and inflammatory reactants. Insulin, glucose, oxygen, and CO2 have hydrodynamic diameters <3.5 nm. Prior studies suggest that 30 nm nanochannels can provide relatively effective immunoisolation.17–19 Narrower membrane channels may be preferable since the molecular weights of the major cytokines are 6–70 kDa (hydrodynamic radius of 2–6 nm),22 although the cytokines will not be produced if the foreign antigen is isolated from the host. Graft failure can also occur due to bioincompatibility, nutrient deficiency, accumulation of waste products, damage from nitric oxide and oxygen free radicals, and above all, inadequate oxygen delivery.23 It takes 1–2 weeks for unencapsulated transplanted islets to re-establish capillary density and branching to pretransplant levels.24 The new vessels may lead to a different organization of the endocrine tissue as compared to native islets.25 Some investigators advocate treatment to increase vascularity before or during transplant.26–28 Since encapsulating membranes inherently bar capillary ingrowth, high membrane permeability to oxygen and nutrients is critical. Provision of sufficient oxygen is a major hurdle for survival of encapsulated islets. The oxygen consumption rate of resting normal islets (approximately 3,000 cells/islet) is 2.5–3 nmol/min/100 islets and increases in response to elevated glucose levels.29–31 The partial pressure of oxygen (pO2) is approximately 95 mm Hg in arterial blood, approximately 40 mm Hg in venous blood, and ranges from 35 to 60 mm Hg in the peritoneum and subcutaneous tissues.32–34 The pO2 of native islets and islets in pancreatic grafts is approximately 40 mm Hg, whereas isolated islets transplanted under the renal capsule have markedly lower oxygen tensions (approximately 6 mm Hg).35 First phase insulin secretion is relatively unaffected by hypoxia, but second phase insulin secretion decreases significantly for pO2 <40 mm Hg.36
Microencapsulation is easy to perform and the encapsulated tissue is easily implanted. Microbeads have excellent surface/volume ratios for transport kinetics. Unfortunately, implanted microbeads will likely lose function over time and cannot reliably be retrieved. Currently used hydrogels (typically alginates) have inherently low oxygen permeability. Additional problems include fragility, fouling of pores, fibrosis, broad and ill-defined channel distributions, and lot-to-lot variability of alginates.23 Macroencapsulation allows easy implantation and retrieval; however, previously used materials were essentially stiff ultrafiltration membranes with inadequate biocompatibility, inappropriate permeability, and poor mechanical properties. Attempts to bring oxygen into such chambers by vascular suture have been plagued by thrombosis, fouling, and infection. Recent reports of nanomachined biocapsules reveal similar problems of stiffness, poor oxygen delivery, and small capacity.17–19 Hollow fibers are easy to remove and allow nutrients to reach the cells,20 but the fiber wall and the distribution of islets within the fiber restricts oxygen delivery, reduces permeability, and creates concentration gradients larger than those associated with microspheres.21
The ideal immunoisolatory membrane needs to be: 1) biocompatible, 2) hemocompatible, 3) biostable for at least 6 months, 4) able to transport oxygen and water rapidly and efficiently, 5) nonfouling, and 6) able to allow entry of nutrients and insulin and exit of metabolic wastes while excluding immunologically active molecules and cells. To accomplish this, membrane walls need to be thin and diffusion paths must be short (approximately <300 μm). Further, the device should be easy to manufacture, sterilize, implant, and remove. Water-swollen hydrogels are, in general, biocompatible and therefore potentially useful for immunoisolation but exhibit poor mechanical properties and ill-defined permeability for essential molecules. By contrast, amphiphilic conetworks (APCNs), i.e., networks that contain approximately equivalent quantities of randomly cross-linked cocontinuous hydrophilic and hydrophobic chain elements, which also swell in water, have far superior mechanical properties. Our discovery that certain APCNs are bio- and hemocompatible, and nonfouling in vivo is of considerable significance for a bioartificial pancreas (BAP).37
Our membranes created expressly for immunoisolation consist of cocontinous hydrophilic/hydrophobic-oxyphilic [i.e., poly(N,N-dimethyl acrylamide) (PDMAAm)/polydimethylsiloxane (PDMS)] domains cross-linked with hydrophobic- oxyphilic polymethylhydrosiloxane (PMHS) chains.38 Domain cocontinuity ensures the rapid and simultaneous diffusion of water-soluble molecules across water-swollen PDMAAm domains and O2 across oxyphilic PDMS/PMHS domains. Permeability to O2 in the latter domains is exceptionally high and provides the needed O2 for the encapsulated islets.38 Indeed, our APCNs exhibit all the properties of the ideal immunoisolatory encapsulating membrane. The design, synthesis, and testing of the APCN membranes took some 20 years of development.39 Our latest APCN membranes are much superior to alginate in tensile strength, burst pressures, and oxygen permeability, whereas membrane thickness is significantly lower than those of microencapsulation membranes.40 Our membranes are also appropriately abrasion resistant for soft tissue implantation as shown by recent tests (to be published).
Materials and Methods
The BAP hardware consists of 1) a 5–10 μm thick semipermeable APCN membrane (i.e., a membrane of cocontinuous PDMAAm/PDMS domains cross-linked by PMHS) expressly created for macroencapsulation and immunoisolation of PICs, 2) an electrospun nanomat (NM) of PDMS-containing polyurethane to reinforce the water-swollen APCN membrane, and 3) a perforated hollow-ribbon superelastic nitinol scaffold (S) to stiffen and provide geometric stability to the construct.
The scaffold (S) defines minimum diffusion distances and keeps them constant. The S is mechanically robust, flexible, biocompatible, easily filled with islets, hermetically sealable, sterilizable, and securely adheres to the APCN/NM membrane. Scaffolds were precision laser machined by Burpee Materials Technology (Eatontown, NJ). The hollow ribbon superelastic nitinol scaffold is 100 μm thick, perforated with laser-cut hexagonal 400 μm wide openings. The openings allow communication between the islets and the surrounding milieu through the APCN membrane. The width of the struts is 50 μm and the thickness of the struts is 100 μm. The hollow device is approximately 600 μm high which can accommodate approximately four layers of islets (average diameter 150 μm). Thus, the maximum diffusion path from the surface to the center of the device is approximately 300 μm. Figure 1A and B shows a schematic of the BAP.
The diameter of the hydrophilic channels of the APCN membranes is 3–4 nm. The average glucose permeation rate is 9.4 mg · cm−2 · h−1, and that for monomeric insulin is 8.2 μg · cm−2 · h−1.40 Details of electrospinning of NMs and preparation of membranes have been described.40 Briefly, PDMS-based polyurethane nanofibers were deposited onto the scaffold, which was then impregnated with an APCN prepolymer solution and cured by heating to produce the reinforced immunoisolatory membrane: the APCN/NM. Impurities were removed by exhaustive extractions and the device ends were sealed with a silicone elastomer sealant (Kwik-Sil, Precision Instruments). A Teflon tube was fixed to the ends of the device to create approximately 1.5-mm holes for suturing. Figure 1C shows the assembled device.
Islets were injected into the scaffold lumen with 20-gauge needles inserted through the sealant. Islets were loaded to approximately 35% of capacity, allowing them to move freely in the device. A pancreatectomized 10 kg dog requires 0.3–0.4 mg (8–10 U) insulin/d. Assuming insulin is released within 1 hour after food ingestion two to three times a day, our devices must deliver 100–150 μg insulin/h. We can readily provide this amount by implanting approximately 130,000 IEQs using seven of our 3.5 cm long devices.
Porcine islet cells (PICs) from retired breeder sows were kindly provided by Drs. M. Trucco and R. Bottino of the University of Pittsburgh. Islet cell isolation was performed by collagenase digestion of the donor pancreas with the semiautomatic Ricordi method. Islet cells were purified on a discontinuous iso-osmolar Ficoll gradient using an automated cell separator (COBE 2991). The cells were washed in Hanks balanced salt solution and islet identification verified by dithizone staining. Viability was verified with a two-color fluorescence technique (calcein AM and propidium iodide). Islets were cultured overnight in CMRL medium containing 10% fetal calf serum, 100 IU/ml of penicillin, and 100 μg/ml of streptomycin and shipped to Cleveland. Islet integrity was verified by light microscopy and islets were then loaded into the sterilized BAP with a syringe.
Assembled APCN/NM/S BAPs were used to implant 60,000–150,000 porcine IEQs in three pancreatectomized 7–12 kg mongrel dogs without immunosuppression. Glucose tolerance tests (500 mg/kg glucose), complete blood counts, and metabolic profiles were done pre- and postpancreatectomy and at weekly intervals after implantation. Blood samples were taken for glucose and insulin at 0, 5, 10, 15, 20, and 30 minutes. After 3 weeks, the devices and tissues around the implants were removed under general anesthesia and the animals were euthanized with 1 ml/5 kg barbiturate/phenytoin (Beuthanasia-D, Schering-Plough, Kenilworth, NJ).
Surgery was performed after a 12-hour fast under an IACUC protocol following the guidelines of AALAC. Animals were premedicated with atropine/glycopyrrolate (0.04/0.02 mg/kg subcutaneously) and acetylpromazine (0.05 mg/kg intramuscularly) and given penicillin benzathine 450,000 Units/penicillin procaine hydrochloride 450,000 Units IM. Total pancreatectomy was accomplished through a midline incision under general anesthesia with N2O/O2/isoflurane taking care to avoid duodenal ischemia. The abdomen was then closed with 0 Vicryl fascial sutures and 3-0 Vicryl subcuticular skin sutures. The animals were recovered and kept on intravenous Ringer's lactate. Pain was relieved by a fentanyl patch (50 μg/24 h) continued for 3 days. Serum glucose levels were checked every 6 hours. Amoxicillin (250 mg orally bid) was given for 3 days postoperatively. On postoperative day 2, dogs received chow with one teaspoon pancreatic enzymes (Pancreved, Vedco Inc, St. Joseph, MO) twice daily. Animals were then treated with approximately 0.5–0.7 units of 70% NPH and 30%. Regular (Novolin 70/30, Novo Nordisk) insulin/kg twice daily to keep fasting blood sugars under 250 mg/dL. Three weeks after pancreatectomy, when insulin requirements had stabilized, the animals were implanted with BAPs under general anesthesia.
Prior to pancreatectomy, fasting glucose averaged 76 mg/dl ± standard deviation (SD) 10.4 mg/dl. Average fasting insulin measured 4.1 ± 5.2 μU/ml and rose to 30.1 ± 22.2 μU/ml 5 minutes after a glucose challenge (500 mg/kg). After pancreatectomy, fasting glucose levels averaged 348 mg/dl ± SD 24.9 mg/dl and insulin was undetectable in all three animals.
The first dog received approximately 130,000 IEQs using three 70 mm long BAPs implanted in the omentum. Device volume was 0.14–0.16 ml, designed to accommodate approximately 33,000–45,000 porcine IEQs (assuming solids in the suspension are 35%–50% of the volume). Fasting glucose normalized at 12 hours (86 mg/dl), but hyperglycemia returned at 24 hours. Renal function tests (serum creatinine and blood urea nitrogen), liver function tests (aspartate aminotransferace, alanine aminotransferace, bilirubin, alkaline phosphatase), and blood counts remained normal. The devices were removed after 3 weeks. Surprisingly, since superelastic nitinol resists kinking, the BAPs fractured because they were too long. The tissues over the devices remained transparent except at the fracture points. Adhesions were present only on the polypropylene sutures used to anchor the devices. A thin (approximately 600–800 μm) capsule with abundant neovascularization was noted adjacent to the implants. There was no lymphocytic infiltrate, mast cells, or eosinophils; however, neutrophils were seen adjacent to the area of the fracture. The next two dogs were implanted with shorter (35 mm) scaffolds internally reinforced with a 28 gauge 316L stainless steel tube (outer diameter 0.321 mm, thickness 0.08 mm, and the same length as the scaffold). The pin was positioned between the struts on one side of the BAP along the long axis. After insertion, the sharp end of the pin was dulled, and it was permanently fixed to the scaffold at both ends with Kwik-Sil. The reinforced BAPs were more resistant to major deformation, but remained flexible and elastic. None broke, even after several weeks of implantation.
The second dog received approximately 130,000 islets using seven BAPs; four in the omentum, one in the subhepatic space, and two in the intramuscular layers of the abdominal wall. Unfortunately, the dog had a small bowel obstruction at the time of device implantation and lysis of adhesions resulted in bowel injury. Serum glucose levels fell from 533 to 72 mg/dL 4 hours after surgery. Glucose was then administered intravenously at 2.5 g/h. The blood glucose rose to 600 mg/dL and the insulin level rose from undetectable levels preoperatively to100 μU/ml. The dog developed peritonitis and died from a bowel perforation 48 hours after implantation. At necropsy, the peritoneum and omentum showed an acute inflammatory reaction with extensive bacterial contamination. Despite this, the islets were still 100% viable by Trypan blue staining.
A device containing 11,000 IEQs was inadvertently left in a closed tube of RPMI medium without antibiotics for 24 hours at 37°C. The device was removed from the heavily contaminated medium, incubated another 24 hours in sterile medium, and then utilized for an in vitro glucose challenge test. After incubation in 13 ml of glucose-free RPMI-1640 medium for 30 min, the device was placed in 2 ml of RPMI-1640 medium (4 mg/ml glucose) and shaken. Samples were taken every 20 minutes for 1 hour without replacing the medium with fresh medium. After each sampling at 1, 1.5, 2, 2.5, and 3 hour, the device was moved to fresh medium (4 mg/ml glucose) to maximize insulin diffusion (insulin build-up in the medium slows down insulin diffusion). Concentrations were determined by ELISA (Linco Research, St. Charles, MO). Results are shown in Figure 2. The total amount of insulin released at 3 hours was 51 mU. Despite a pO2 estimated to be near the critical level needed to support intracellular metabolism (1–3 mm Hg), the membranes' extraordinary oxygen permeability allowed the islets to survive and to retain their insulin secreting ability. We conclude that our membranes not only protected the islets from acute rejection, but also from environmental stresses that normally would result in cell death.
The third dog was implanted with seven 35 mm long BAPs; four in the omentum and three subcutaneously. The islet preparation was only approximately 60% pure. The estimated number of viable islets implanted was only approximately 70,000 IEQs. The dog showed a transient improvement in glycemic control with measurable insulin on days 3 and 5; however, insulin was undetectable after day 5 and there was no improvement in the GTTs at 1 and 3 weeks. No erythema or fluid collections were noted. Renal and liver function tests remained normal. Three weeks after implantation, the devices were easily separated from the omentum (Figure 3) and the subcutaneous tissues. There was no inflammation or rejection in the omentum or the subcutaneous tissues. Neovascularity was present adjacent to the devices (Figure 4). Islets appeared to be clustered in the openings in the scaffold where oxygen and nutrient delivery should be maximal (Figure 5). Trypan blue staining (Figure 6) indicated that 50%–60% of cells were viable.
Our BAP kept xenografted islets alive for 3 weeks without prevascularization or immunosuppression. Thus, it provided effective immunoisolation. The 3–4 nm channels also prevented bacterial entry into the BAP. We did not achieve relief of hyperglycemia, likely due to the lack of sufficient islet mass. Other possible explanations are a slow insulin permeation rate and capsule formation. Given that encapsulated PICs survived hypoxic in vitro and in vivo environments, oxygen delivery appears more than adequate. The in vitro insulin permeation rate of our BAP is in the designed target range (8.2 μg/cm2 · h). The in vivo permeation rate should be even higher because unlike natural insulin, commercially available insulin contains 0.5% Zn2+ which delays insulin transit across the membrane. Since Zn2+ concentrations are negligible in vivo, the permeation rate of insulin should increase fivefold, i.e., to approximately 41 μg/cm2 · h. Thus, the insulin permeability of our immunoisolatory devices should be appropriate for a clinically acceptable BAP. Release rate depends on the concentration of islets in suspension, and it takes time for the insulin concentration to reach its maximum. Diffusion across the BAP membrane and the placement of the islets outside the circulatory system also delays response. The observed neovascularization and the prolonged survival of the islets suggest that nutrients and insulin are able to pass through the membranes, but further studies will be necessary to confirm this.
A conceptually novel BAP was used to implant live porcine islets into dogs without immunosuppression of any kind. Islets remained viable up to 3 weeks without evidence of rejection and were protected in environments that normally would lead to cell death. Normoglycemia was not obtained, probably due to insufficient islet cell mass. Further studies are warranted.
The authors thank the NSF (DMR-0243314), The University of Akron, and the Cleveland Clinic for financial support, and Dr. Rita Bottino and Dr. Massimo Trucco of the University of Pittsburgh for kindly providing islet cells. The authors also thank the Ohio Department of Development Third Frontier program for funding of this research through University of Akron, Center for Multifunctional Polymer Nanomaterials and Devices (CMPND).
1. Fiorina P, Secchi A: Pancreatic islet cell transplant for treatment of diabetes. Endocrinol Metab Clin North Am
36; 999–1013, 2007.
2. Maffi P, Bertuzzi F, De Taddeo F, et al
: Kidney function after islet transplant alone in type 1 diabetes: Impact of immunosuppressive therapy on progression of diabetic nephropathy. Diabetes Care
30: 1150–1155, 2007.
3. Torpey N, Bradley JA, Fung JJ: Immunosuppressive therapy in solid organ transplantation, in Zbar AP, Guillou PJ, Bland KI, et al (eds), Immunology for Surgeons.
London, Springer-Verlag, 2002, pp. 127–166.
4. Cooper DK, Gollackner B, Sachs D: Will the pig solve the transplantation backlog? Annu Rev Med
53: 133–147, 2002.
5. White DJG: Islet xenotransplantation. Curr Opin Organ Transplant
12: 148–153, 2007.
6. Hoerbelt R, Madsen JC: Feasibility of xeno-transplantation. Surg Clin North Am
84: 289–307, 2004.
7. Cantarovich D, Blancho G, Potiron N, et al
: Rapid failure of pig islet transplantation in non human primates. Xenotransplantation
9: 25–35, 2002.
8. van der Windt, DJ, Bottino RJ, Casu A, et al
: Rapid loss of intraportally transplanted islets: An overview of pathophysiology and preventive strategies. Xenotransplantation
14: 288–297, 2007.
9. Mirenda V, Golshayan D, Read J, et al
: Achieving permanent survival of islet xenografts by independent manipulation of direct and indirect T-cell responses. Diabetes
54: 1048–1055, 2005.
10. Jiang S, Herrera O, Lechler RI: New spectrum of allorecognition pathways: Implications for graft rejection and transplantation tolerance. Curr Opin Immunol
16: 550–557, 2005.
11. Binette TM: Porcine endogenous retroviral nucleic acid in peripheral tissues is associated with migration of porcine cells post islet transplant. Am J Transplant
4: 1051–1060, 2004.
12. Dufrane D, Goebbels RM, Saliez A, et al
: Six-month survival of microencapsulated pig islets and alginate biocompatibility in primates: Proof of concept. Transplantation
81: 1345–1353, 2006.
13. Schneider S, Feilen PJ, Brunnenmeier F, et al
: Long-term graft function of adult rat and human islets encapsulated in novel alginate-based microcapsules after transplantation in immunocompetent diabetic mice. Diabetes
54: 687–693, 2005.
14. Duvivier-Kali VF, Omer A, Lopez-Avalos MD, et al
: Survival of microencapsulated adult pig islets in mice in spite of an antibody response. Am J Transplant
4: 1991–2000, 2004.
15. Jesser C, Kessler L, Lambert A, et al
: Pancreatic islet macroencapsulation: A new device for the evaluation of artificial membrane. Artif Organs
20: 997–1007, 1996.
16. Honiger J, Darquy S, Reach G, et al
: Preliminary report on cell encapsulation in a hydrogel made of a biocompatible material, AN69, for the development of a bioartificial pancreas. Int J Artif Organs
17: 46–52, 1994.
17. Leoni L, Desai TA: Micromachined biocapsules for cell-based sensing and delivery. Adv Drug Deliv Rev
56: 211–229, 2004.
18. La Flamme KE, Popat KC, Leoni L, et al
: Biocompatibility of nanoporous alumina membranes for immunoisolation. Biomaterials
28: 2638–2645, 2007.
19. Smith C, Kirk R, West T, et al
: Diffusion characteristics of microfabricated silicon nanopore membranes as immunoisolation membranes for use in cellular therapeutics. Diabetes Technol Ther
7: 151–162, 2005.
20. Altmah JJ: The bioartificial pancreas: Macroencapsulation of insulin secreting cells in hollow fibers. J Diabet Complications
2: 68–74, 1988.
21. Thrash ME Jr: An analysis of the oxygen transport mechanism in a bioartificial pancreas, Master Thesis, Cleveland State University, 1997.
22. Haddad JJ: Cytokines and related receptor-mediated signaling pathways. Biochem Biophys Res Commun
297: 700–713, 2002.
23. DeVos P, Van Hoogmoed CG, van Zanten J, et al
: Long-term biocompatibility, chemistry, and function of microencapsulated pancreatic islets. Biomaterials
24: 305–312, 2003.
24. Jones GL, Juszczak MT, Hughes SJ, et al
: Time course and quantification of pancreatic islet revascularization following intraportal transplantation. Cell Transplant
16: 505–516, 2007.
25. Morini S, Brown ML, Cicalese L, et al
: Revascularization and remodelling of pancreatic islets grafted under the kidney capsule. J Anat
210: 565–577, 2007.
26. Sigrist S, Mechine-Neuville A, Mandes K: Induction of angiogenesis in omentum with vascular endothelial growth factor: Influence on the viability of encapsulated rat pancreatic islets during transplantation. J Vasc Res
40: 359–367, 2003.
27. Yin Z, Wu W, Fung JJ, et al
: Cotransplanted hepatic stellate cells enhance vascularization of islet allografts. Microsurgery
27: 324–327, 2007.
28. Langlois A, Bietiger W, Mandes K, et al
: Overexpression of vascular endothelial growth factor in vitro using deferoxamine: A new drug to increase islet vascularization during transplantation. Transplant Proc
40: 473–476, 2008.
29. Carlsson PO, Liss P, Andersson A, et al
: Measurement of oxygen tension in native and transplanted rat pancreatic islets. Diabetes
47: 1027–1032, 1998.
30. Zacharovova K, Berkova Z, Spacek T, et al
: In vitro assessment of pancreatic islet vitality by oxymetry. Transplant Proc
37: 3454–3456, 2005.
31. Hellerstrom C: Effect of carbohydrate on the oxygen consumption of isolated pancreatic islets of mice. Endocrinology
81: 105–112, 1967.
32. Transport of oxygen and carbon dioxide in the blood and body fluids, in Guyton AC (ed), Textbook of Medical Physiology
, 8th ed. Philadelphia, W.B. Saunders Co., 1991, pp. 433–442.
33. Ward WK, Wood MD, Slobodzian EP: Continuous amperometric monitoring of subcutaneous oxygen in rabbit by telemetry. J Med Eng Technol
26: 158–167, 2002.
34. Klossner J, Kivisaari J, Niinikoski J: Oxygen and carbon dioxide tensions in the abdominal cavity and colonic wall of the rabbit. Am J Surg
127: 711–715, 1974.
35. Carlsson EO, Palm F: Oxygen tension in isolated transplanted rat islets and in islets of rat whole-pancreas transplants. Transpl Int
15: 581–585, 2002.
36. Dionne K, Colton CK, Yarmush ML: Effect of hypoxia on insulin secretion by isolated rat and canine islets of Langerhans. Diabetes
43: 12–21, 1993.
37. Chen D, Kennedy JP, Kory MM, et al
: Amphiphilic networks II: Biocompatibility and controlled drug release of poly[isobutylene-co-2-(dimethylamino) ethyl methacrylate]. J Biomed Mat Res
2: 1327–1342, 1989.
38. Kang J, Erdodi G, Kennedy JP: Third-generation amphiphilic conetworks. III. Permeabilities and mechanical properties. Part XXX of the series “amphiphilic conetworks”. J Poly Sci Part A Polym Chem
45: 4276–4283, 2007.
39. Karanukaran R, Kennedy JP: Novel amphiphilic conetworks by synthesis and crosslinking of allyl-telechelic block copolymers. (Part XXXIII of the series “amphiphilic conetworks”). J Polym Sci Part A Polym Chem
46: 4254–4257, 2008.
40. Erdodi G, Kang J, Yalcin B, et al
: A novel macroencapsulating immunoisolatory device: The preparation and properties of nanomat-reinforced amphiphilic conetworks deposited on perforated metal scaffold. Biomed Microdevices
11: 297–312, 2009.