Encapsulation of pancreatic islets has long been a strategy for treatment of diabetes1. Despite the concerted efforts of numerous research groups, a viable encapsulation system leading to prolonged insulin independence has remained elusive. This failure is the result of the large number of factors leading to loss of function for the grafted tissue. Modifications to basic encapsulation systems that seek to address many of these factors individually have shown short-term success but failed to produce a clinically successful system for treatment of human patients. The most obvious factor is immune rejection, but studies comparing survival of encapsulated auto- and allograft tissue showed similar survival time indicating the role of other stresses leading to loss of function2. Some of these stresses include: insufficient diffusion of nutrients and waste products through the capsule resulting in hypoxia and oxidative stress, lack of biocompatibility of the capsule material and the lack of sufficient cell-cell and cell-ECM (extracellular matrix) interactions within the capsule.
The central goal of the work proposed here is to improve the long term function and viability of encapsulated islet cells through the use of genetic modification prior to encapsulation. The intracellular signaling pathways that translate extracellular stresses into cell death (with particular attention given to anoikis) will be targeted for genetic modification to block these processes and in turn extend the viability and function of the graft tissue. The role of hypoxia in graft function following encapsulation will also be studied through genetic modification.
To test the importance of cell-ECM interactions, MIN6 b-cells were infected to overexpress mILK and/or mBcl-2 prior to encapsulation. We hypothesized that integrin-linked kinase (ILK) overexpression would prevent b-cell anoikis caused by the lack of physiological cell-ECM interactions within the PEG capsule. Bcl-2 (an integral protein found in the endoplasmic reticulum, nuclear envelope and mitochondria) is able to block apoptosis stimulated by a range of different stresses. Bcl-2 overexpression should also block anoikis but is less specific than ILK in this regard. Previous work using Bcl-2 overexpression in transplanted islets has shown promising results3. The role of hypoxia is monitored through the use of a hypoxia-responsive fluorescent marker.
EXPERIMENTAL METHODS
Synthesis of poly(ethylene glycol) (PEG) Capsule - A 10% solution of PEG dimethacrylate macromers [molecular weight of 10,000 daltons] containing an ultraviolet (UV) photoinitiator is mixed with islets isolated from BALB/c mice or MIN6 cells (a murine b cell line). This PEG-cell suspension is placed in 1ml syringe (40 ml per syringe) and exposed to light under sterile conditions for 15 minutes. The UV-initiated polymerization forms a cylindrical disk with the islets embedded inside. Smaller and larger capsules can also be formed as desired and previous studies have shown maintenance of viability for a range of cell types encapsulated with this technique4-6.
Genetic Modification of Islets and MIN6 Cells - Recombinant pShuttle adenoviral vectors with murine Bcl-2 (mBcl-2) or murine ILK (mILK) inserted into the multiple cloning site under the control of a cytomegalovirus (CMV) promoter were constructed. Homologous recombination in DH5a bacteria with the pAdEasy-1 Ad5DE1/DE3 plasmid created a new plasmid with mBcl-2 or mILK expression cassette inserted into the E1 region of the adenoviral genome. A previously constructed plasmid identical to the mBcl-2 and mILK plasmids, except with enhanced green fluorescent protein (eGFP) under control of the CMV promoter, was used to establish the infectivity of both Balb/c islets and MIN6 cells. Viral multiplicities of infection (MOI) ranging from 1 to 10000 were used to infect the MIN6 and Balb/c islets with eGFP and expression of eGFP was monitored over a period of 2 months by fluorescent microscopy. A hypoxia sensitive marker virus was also constructed using the pAdEasy system. A hypoxia responsive element (HRE) trimer was placed upstream of a minimal promoter and a red fluorescent protein sequence yielding a hypoxia responsive virus (HRE-Red).
Effect of Genetic Modification on Survival of Encapsulated MIN6 Cells - MIN6 cells were infected with recombinant adenovirus designed to overexpress mILK, mBcl-2, both mILK and mBcl-2 or cyclization recombinase (Cre). Forty eight hours after the infection, the cells were encapsulated in PEG gels as described above and placed in growth media containing 10% fetal bovine serum. Five days after the encapsulation procedure, LIVE/DEAD® Viability/Cytotoxicity Kit (Molecular Probes) staining was performed to determine the viable fraction of cells in each of the groups.
Evaluation of HRE-Red Virus - Islets were infected with either the HRE-Red virus or Cre as a control. Infected islets were placed in a hypoxia chamber which was purged for 5 minutes with a 95% nitrogen and 5% carbon dioxide gas mixture to create a low oxygen environment. Following the purge, the chamber was placed in a tissue culture incubator to maintain the cells at 37º C. Hypoxia studies were carried out with a 6 hour exposure to the hypoxic conditions followed by a return to standard culture conditions (air with 5% carbon dioxide).
RESULTS AND DISCUSSION
The infectivity of MIN6 cells and BALB/c islets with recombinant adenovirus generated using the pAdEasy system was established through this work. A recombinant adenovirus with eGFP constitutively expressed by the CMV promoter was used to infect BALB/c islets and MIN6 cells. Initial studies performed in MIN6 showed expression of GFP following infection and maintenance of signal 2 months after infection. Infection of BALB/c islets was also performed and expression of GFP noted uniformly throughout the islets. A range of MOIs was used in these studies and based on our observations, a MOI of 10 was sufficient for protein expression in the vast majority of cells without causing toxicity.
Initial studies established that dispersed MIN6 cells show low viability in PEG gels after 5 days in culture. MIN6 cells infected to overexpress mILK, mBcl-2 or both showed considerably improved viability under these conditions compared to control transfected cells. Control cells were infected with an adenovirus designed to overexpress Cre. Cre-infected MIN6 cells showed 42 ± 9% viability which was similar to the results found for unmodified MIN6 cells encapsulated for 5 days and significantly lower than the genetically modified cells. After 5 days in the PEG capsule, mBcl-2-infected cells were 76 ± 3% viable, mILK-infected MIN6 cells were 95 ± 3% viable and cells infected with both mILK and mBcl-2 were 88 ± 5% viable. Statistical evaluation showed significant difference between the control cells (Cre-infected) and all three groups receiving mBcl-2 or mILK (p-values <0.001) Also, MIN6 cells infected with mILK showed a significantly higher viability than those infected with mBcl-2 alone (p-value <0.001). Finally, comparing MIN6 cells infected with mILK or mILK and mBcl-2 showed these groups to be no different statistically (p-value > 0.05).
These results suggest that anoikis is a cause of cell death observed in encapsulated dispersed MIN6 cells, and that survival may be improved by activating the pathway triggered by interactions of b1 integrins with the ECM proteins lacking in the PEG hydrogel. We are, in effect, “hotwiring” this pathway to make the cells think they are interacting with the absent ECM matrix. Similar studies are underway with encapsulated Balb/c islets. Initial studies showed longer viability in control groups than that of the MIN6 cells as viability in the control group using Balb/c islets was still over 90% 2 weeks after encapsulation. Studies carried out over a longer duration are being conducted and if a similar increase is seen in the lifespan of infected islets as with infected MIN6 cells then it will be very promising for the initiation of in vivo studies.
The HRE-Red virus was initially evaluated using a hypoxia chamber. Islets infected with Cre or HRE-Red were placed in a low oxygen environment for 6 hours and the fluorescent signal was then monitored after removal from the hypoxic conditions. No red fluorescence was detected in islets infected with Cre and placed in the hypoxia chamber as compared with the HRE-Red islets which showed fluorescence intensity levels corresponding to the MOI used to infect the islets. The higher the MOI, the stronger the intensity of the red fluorescent signal detected. These results show that islets can be modified so that their response to available oxygen can readily be monitored. This will allow for the impact of islet size and capsule composition on oxygen diffusion to be studied without disruption of the capsule.
REFERENCES
[1] Lim F and Sun AM. Science; 210, 908 (1980).
[2] De Vos P et al. Diabetologia; 40, 262 (1997).
[3] Contreras JL et al. Transplantation; 71, 1015 (2001).
[4] Bryant SJ and Anseth KS. J Biomed Mater Res A; 64, 90 (2003).
[5] Burdick JA and Anseth KS. Biomaterials; 23, 4315 (2002).
[6] Nuttelman CR et al. J Biomed Mater Res A; 68, 773 (2004).
ACKNOWLEDGEMENTS
We would like to thank Dr Stephen Langer for his assistance with the construction of the recombinant adenoviruses and Francisco Ramirez-Victorino and Philip Pratt for isolating the Balb/c islets for these studies