Type 1 diabetes mellitus is an autoimmune disorder in which the insulin-producing β-cells of the pancreatic islets of Langerhans are selectively destroyed (1). Transplantation of allogeneic islets offers a novel therapeutic approach for patients with type 1 diabetes (2). The recent success of clinical islet transplantation has been hampered by the fact that large masses of islets are needed to achieve complete graft function and insulin independence (2,3). Many factors reduce islet viability or impair islet function and contribute to the high rate of primary nonfunction generally observed in transplant recipients. Notably, the stress associated with islet isolation, purification and culture, hypoxia, and nonspecific inflammation in the transplant microenvironment all negatively influence islet graft survival (3). In addition, nonspecific inflammation may trigger the immune system and lead to allo- and autoimmune responses.
It is estimated that up to 60% of transplanted β-cell mass is lost to apoptosis even in syngeneic situations (4), implying that interventional strategies with anti-apoptotic genes may be crucial for islet mass preservation in the transplant setting. The introduction of cytoprotective genes ex vivo into islets to reduce their susceptibility to noxious insults in the implant microenvironment has been the subject of much research (5–7). This approach may ultimately lower the number of islets required for successful transplants so islet transplantation may benefit more patients. A number of candidate gene products can extend graft survival, although some are derived from malignant cells in which their anti-apoptotic effects are linked to oncogenicity (8). This has led several groups, including our own, to explore alternate anti-apoptotic genes (9,10).
Our laboratory established the expression of functional erythropoietin (EPO) receptors throughout all cells in primary islets. Furthermore, we demonstrated that preconditioning the media with recombinant EPO protects islets from cytokine-induced destruction in a dose-dependent manner (11). We hypothesized that engineering constitutive EPO expression in the islet microenvironment may promote survival through an autocrine or paracrine mechanism. This study examines whether transduction of human islets with EPO in the context of an adenoviral (Ad) vector (Ad-EPO) protects them from apoptosis.
EPO is a kidney cytokine that regulates hematopoiesis by promoting the survival, proliferation, and differentiation of erythroid progenitor cells (12). Specifically, EPO blocks the normal apoptotic cycle in erythroid progenitors as they progress through maturation (13) by affecting the JAK/STAT5 pathway (14) or the Bcl2 pathway (15). Recently, the role of EPO has been extended beyond hematopoiesis to other tissues where it seems to play a protective role by shielding against hypoxic, ischemic, apoptotic, and necrotic stress (16–19).
Human pancreatic islets were transduced with an Ad vector encoding EPO (Ad-EPO) or green fluorescent protein (GFP) (Ad-GFP). Islets transduced with Ad-EPO expressed high levels of EPO, showed normal regulated insulin secretion, and were significantly protected from apoptosis in vitro. In addition, assessment of the performance of transduced islets transplanted into diabetic immunodeficient mice showed that overexpression of EPO conferred a functional advantage. Our results indicate that EPO is cytoprotective to islets without altering islet function. The ex vivo transfer of cytoprotective genes, such as EPO, to evade islet cell death may promote the engraftment of a higher number of viable islet cells. This approach could ultimately reduce the mass of islets necessary per recipient to observe complete graft function. Although EPO-transduced islets were functional in vivo, animal recipients exhibited side effects possibly because of the high systemic unregulated EPO levels.
MATERIALS AND METHODS
Islet Isolation and Culture
Islets obtained from human cadaveric pancreata were isolated by enzymatic digestion (Liberase; Boehringher-Mannheim, Gaithersburg, MD) using the purification methods previously described (20). After isolation, islets were cultured overnight at 37°C, 5% CO2, in CMRL medium (Mediatech Inc., Herindon, VA) with 10% fetal calf serum (HyClone, Logan, UT), 100 U/mL penicillin, and 0.1 mg/mL streptomycin, 2 mM glutamine, and 25 mM HEPES buffer (Gibco, Grand Island, NY).
Adenoviral Vectors and Transduction
Two recombinant E1-deleted Ad vectors were used: one encoding GFP (Ad-GFP) and the other encoding human EPO (Ad-EPO), both under the control of the human cytomegalovirus promoter. Both viruses were provided by Dr. Andrea Gambotto (Department of Surgery, School of Medicine, University of Pittsburgh, Pennsylvania). Transduction was accomplished as previously described by incubating isolated islets in serum-free medium for 45 min in the presence of virus at a multiplicity of infection of 500 (assuming 1,000 cells/islet or 0.5 virion/islet cell).
Assessment of Transduction Efficiency
Twenty-four hours posttransduction, islets were examined for GFP expression using confocal microscopy. Quantitative analysis of transduction was performed on disassociated islets (dissociation buffer; Gibco, Grand Island, NY) by acquiring 50,000 cells per condition with a cytometer (FACSCalibur, Becton-Dickinson, San Diego, CA) to determine the percentage of cells expressing GFP. All experiments were performed in triplicate for reproducibility, and the results are expressed as mean with standard deviation.
Measurement of Erythropoietin Expression in Islets and Media
Western blot assays were used to determine and quantify EPO in islets. Islets were solubilized in 20 mM Tris, pH 6.8, containing 1% sodium dodecyl sulfate and facilitated by brief sonication. Proteins were recovered after centrifugation at 12,000 g. Protein concentrations were measured by the BCA Protein Assay (Pierce, Rockford, IL) using bovine albumin standards. Proteins were separated by electrophoresis on 16% polyacrylamide gels (Invitrogen, Carlsbad, CA) under denaturing and reduced conditions and transferred to nitrocellulose membranes, and membranes were used for immunoblotting as previously described (22). Membranes were incubated with 2 μg/mL rabbit anti-EPO (Chemicon International, Temecula, CA) and detected with goat anti-rabbit immunoglobulin (Ig)G-peroxidase conjugate (1:2,000 dilution; Sigma Chemicals, St. Louis, MO) by chemiluminescence (Pierce). Chemiluminescence images were captured with a digital imaging system (FluorChem 8000, Alpha Innotech, San Leandro, CA). Membranes were stripped of bound antibodies by incubation in 2% sodium dodecyl sulfate, 100 mM mercaptoethanol, 62.5 mM Tris, pH 6.7, at 50°C for 30 min, washed, and reprobed with rabbit anti-actin (1:5,000, Sigma Chemicals).
Confocal Microscopy and Immunofluorescence
Islets were fixed in 10% buffered formalin, paraffin embedded, and sectioned for immunohistochemistry. Paraffin sections (5 μm) were deparaffinized, rehydrated, and incubated at room temperature in Universal Blocking Reagent (Bio-Genex, San Ramon, CA) for 10 min to block nonspecific binding. For the simultaneous identification of EPO and insulin antibodies, the following subsequent incubation steps were performed: four hours with anti-EPO antibody (rabbit polyclonal IgG, 1:250 dilution; Santa Cruz Biotechnology, Santa Cruz, CA) and 1 hr with anti-insulin antibody (chicken anti-human insulin, 1:400 dilution; Axell, Westbury, NY). After washing, a secondary goat anti-rabbit (Alexa Fluor 568 goat anti-rabbit IgG, 1:200 dilution; Molecular Probes, Eugene, OR) and goat anti-chicken (Alexa Fluor 647 goat anti-chicken IgG, 1:400 dilution; Molecular Probes) were applied. Omission of the primary antibody served as a negative control. Monkey kidney and human pancreas were used as positive controls for EPO and insulin, respectively. Sections were photographed using a confocal microscope (LSM 510, Carl Zeiss, Germany) with the following lasers: argon, helium-neon 1, and helium-neon 2.
Islet Cell Death Quantification
After transduction, islets were serum-starved for 1 week to induce apoptosis, and islets were harvested and dissociated using enzyme-free dissociation buffer (Gibco). Quantitative analysis of cell death was performed on dissociated islet cells using an annexin V-phycoerythrin apoptosis detection kit (BD Pharmingen, San Diego, CA) and propidium iodide (PI) for necrosis by flow cytometry on a linear scale.
Assessment of Islet Function In Vitro
Glucose-stimulated insulin secretion assay was used to assess β-cell function in untransduced, Ad-GFP–transduced, and Ad-EPO–transduced islets. Triplicates of 50 islet equivalent number (IEN) were incubated at 37oC in 1 mL of Roswell Park Memorial Institute medium without glucose and supplemented with 10% fetal calf serum. Supernatants were collected at the end of each of two sequential 1-hr incubations at low (2 mM) and high (20 mM) glucose concentration. Release of insulin was quantified in supernatants using a commercial Mercodia enzyme-linked immunosorbent assay kit (Alpco, Windham, NH). Stimulation indices were calculated by dividing the insulin output measured after the exposure to high glucose by the one obtained at low glucose.
Islet Transplantation in Immunodeficient Mice
Islet function was assessed in a transplantation setting. Islet graft recipients were 4- to 6-week-old athymic nude-nu (nude) mice purchased from Harlan (Indianapolis, IN). Diabetes was induced by a single intravenous injection of 200 mg/kg of streptozotocin (Sigma Chemicals) 2 weeks before islet transplantation. Diabetic mice were used as recipients of Ad-EPO–transduced, Ad-GFP–transduced, or unmanipulated islets only if nonfasting blood glucose values greater than 350 mg/dL were recorded (Glucometer Elite; Bayer, Tarrytown, NY). Two animals per group underwent transplantation using “suboptimal” grafts (1,000 IEN), and three animals per group underwent transplantation using “optimal” grafts (2,000 IEN) implanted under the left kidney capsule as previously described (23).
Assessment of Islet Function In Vivo
Assessment of graft function in the “suboptimal” islet graft model was determined using an intraperitoneal glucose tolerance test (IPGTT) 10 days after transplantation. Animals were fasted for 16 hr, and glycemic levels were measured before the administration of an intraperitoneal injection of glucose solution in saline (2 g/kg of body weight). Glucose levels were measured at the indicated time points for 2 hr after glucose bolus using a portable glucometer. Hematocrit and other blood chemistries were assayed by standard clinical techniques. Animals were kept in the University of Miami animal facilities and used in compliance with the United States Department of Agriculture and National Institutes of Health regulations, and all manipulations were conducted and monitored under protocols reviewed and approved by the Institution Animal Care and Use Committee.
Adenovirus Transduction of Human Islets Is Highly Efficient
Transduction efficiency was quantified by analyzing GFP positive cells after dissociation of islets transduced with Ad-GFP fluorescence using cytometry analysis. The mean and standard deviations of Ad-GFP–transduced islet cells in eight independent experiments were as follows: Ad-GFP 72.0±11.0; Ad-EPO 2.7±1.2; control 1.6±0.3. This confirms that Ad transduction of human islets is efficient (21).
Insulin Synthesis and Secretion Is Not Negatively Affected in Adenoviral-Erythropoietin–Transduced Islets
To assess whether transduction or transgene expression affects the insulin content or the ability of islets to secrete insulin in response to glucose stimulation, insulin content and glucose-stimulated insulin secretion were measured 24 hr posttransduction. Table 1 shows that both insulin content and glucose stimulation indices were comparable between control, nontransduced islets, and Ad-EPO–transduced islets but were significantly different (P =0.006) when comparing Ad-EPO– and Ad-GFP–transduced islets. These results indicate that the Ad-induced cytotoxicity is counteracted by EPO inasmuch as metabolic function is affected in Ad-GFP–transduced but not in Ad-EPO–transduced islets.
Islets Transduced with Adenoviral-Erythropoietin Synthesize and Secrete High Levels of Erythropoietin
As shown in Figure 1A, EPO was expressed in islets transduced with Ad-EPO but not in nontransduced islets or in islets transduced with Ad-GFP, confirming specific expression of the transgene. Ad-EPO–transduced islets secrete high levels of EPO as assessed by Western blot analysis 24 hr after infection by the presence of immunoreactive EPO in the conditioned media (Figure 1C). On the basis of these data, we estimated that after 24 hr in culture there was threefold to fivefold more EPO in the medium than present in the cells. Furthermore, densitometry of the immunoblot bands showed that EPO accounted for 11%±2 % of the protein present in the islet homogenate (Figure 1D and E).
Erythropoietin, Green Fluorescent Protein, and Insulin Are Expressed in Islets
Transgene expression was assessed using immunohistochemistry on isolated transduced islets. The presence of immunoreactive insulin was observed in islets regardless of transgene expression as shown in Figure 2. For co-imaging of insulin immunofluorescence (red) and EPO immunofluorescence (green), a blocking step was used before sequential antibodies (see Materials and Methods). For co-imaging of insulin and GFP, islets were immunostained only with anti-insulin in which the GFP expression is determined by fluorescence of the transgene. Both transgenes (EPO and GFP) were found to be appropriately expressed in the transduced islets with clear coexpression of EPO with insulin as seen in Figure 4. Colocalization of EPO-insulin is seen in yellow and demonstrates transduction of β-cells, whereas non-β-cells that are transduced are seen in green.
Erythropoietin Expression Is Cytoprotective In Vitro
To determine whether apoptosis or necrosis levels are altered by expression of EPO, transduced and control islets were cultured for 1 week in normal serum and then serum-starved for 1 week to induce apoptosis. Although apoptosis levels of human islets vary between preparations, islets cultured for 2 weeks in serum-containing medium have levels of apoptosis ranging between 9% and 16%, which is therefore considered “acceptable” under normal culture conditions. Cell death was analyzed on dissociated cells by flow cytometry using both annexin V staining, a hallmark of apoptosis, and PI, which reflects necrosis. Untransduced islets were compared with islets transduced with Ad-GFP or Ad-EPO. Figure 3 summarizes the results of six independent experiments, showing that the proportion of cell death was 88%±9% (29%±6% annexin V+; 58%±3% PI+) in control, 63%±4% (26%±2% annexin V+; 37%±4% PI+) in Ad-GFP–transduced islets, and 49%±3% (5%±3% annexin V+; 43%±4% PI+) in Ad-EPO–transduced islets (P <0.01). These data indicate a cytoprotective effect of EPO on islets and confirm that Ad transduction does not induce islet death at this multiplicity of infection. A consistent and unexpected observation of our study was that in culture, the survival of Ad-GFP–transduced islets was higher than the untransduced control islets in the six independent experiments that were performed.
Improved In Vivo Function of Adenoviral-Erythropoietin–Transduced Islets
To determine whether EPO is cytoprotective in vivo, Ad-GFP–transduced, Ad-EPO–transduced, or nontransduced islets were transplanted under the kidney capsule of chemically induced diabetic nude mice. Two animals per group were transplanted using “suboptimal” grafts (1,000 IEN), and they showed nonfasting blood glucose values in the hyperglycemic range during the 10-day posttransplant observation period, regardless of islet treatment. In our experience, transplantation of 1,000 IEN of human islets into chemically diabetic immunodeficient mice generally results in delayed or no reversal of diabetes. With this model, islets expressing a cytoprotective gene should be characterized by increased cell viability, better glucose-stimulated insulin release, and improved in vivo function when compared with the unmanipulated control.
When the glucose challenge test was performed 10 days posttransplant, fasting glycemic values were measured in the graft recipients. Assessment of fasting blood glucose is a useful tool to determine the function of implanted islets when the islet mass implanted does not result in normalization of nonfasting blood glucose levels. Animals receiving Ad-EPO–transduced islets demonstrated fasting blood glucose levels in the normal range of values (<200 mg/dL; time 0 shown in Figure 4), whereas animals receiving Ad-GFP–transduced or untransduced islets demonstrated elevated fasting glycemic values, indicating the metabolic advantage provided by the overexpression of EPO. Assessment of glucose disposal rate during the IPGTT allows for the evaluation of the functional islet mass and for the quantitative comparison between experimental groups receiving identical islet masses. In our study, animals bearing Ad-EPO–transduced islets showed improved response to glucose challenge when compared with animals receiving a comparable mass of Ad-GFP–transduced islets (Figure 4). These results indicate that the glucose clearance rate was improved by EPO overexpression.
Immunostaining for insulin throughout the recipient pancreata from all experimental animals was negative (data not shown), demonstrating no residual endocrine function after streptozotocin treatment.
Effects of Erythropoietin Overexpression In Vivo
An unexpected finding of our experiments was that animals undergoing transplantation with the “optimal” islet mass of 2,000 IEN Ad-EPO–transduced islets showed a deterioration of their general condition, paralleled by thickening of blood and splenomegaly. To explore the possible role of EPO overexpression in the observed phenomenon, blood samples were collected 35 days after transplantation from chemically diabetic nude mice (n=3/group) that had received Ad-EPO– or Ad-GFP–transduced islets. Erythrocytes, hemoglobin, and platelet counts in the blood were significantly increased in the animals receiving Ad-EPO–treated islets, probably as a consequence of the high level of EPO expression by the grafted islets (Table 2).
The introduction of cytoprotective genes into pancreatic islets before transplantation has been suggested as an important approach to improve the clinical outcome of islet transplantation. Our group recently demonstrated that there are functional EPO receptor islets and that recombinant EPO added to media was cytoprotective to these cells in a dose-dependent manner (11). These results, combined with the fact that a variety of cells synthesize EPO resulting in both paracrine and autocrine protective effects (21,24), led us to test the ability of transduction with Ad-EPO to protect human islets in vitro and in vivo.
In this article, the expression of EPO in pancreatic islets was examined in the context of an Ad vector. The results of the experiments described demonstrate that when human islets are engineered to express EPO, cell death is significantly reduced in serum-starvation conditions, whereas both function and insulin content are retained. Expression of the EPO gene resulted in islets that not only retained glucose-regulated insulin secretion in vitro but also did so significantly better than islets transduced with a marker protein (GFP). It should be noted, however, that lower levels of necrosis were consistently observed in islets transduced with Ad-GFP (Fig. 3), a result that was unexpected. Recent data by Zhang et al. (25) showed that transduction of a variety of cells with Ad-GFP induces heat-shock protein 70, a stress-responsive gene important for cell survival. Heat-shock protein 70 overproduction has been shown to be protective in a variety of cells both in vitro and in vivo because of a reduction of cytochrome c release with subsequent DNA fragmentation (26).
Ad-EPO–transduced islets showed improved function in vivo as demonstrated by the IPGTT, when compared with control islets. Taken together, these data indicate that the synthesis and secretion of EPO by cultured islets may be beneficial in maintaining or improving cell engraftment and function. This is a novel potential gene for genetic engineering of islets and may prove to be a viable cytoprotective protein for islets.
Although these data are promising, the use of adenovirus with a viral promoter has limitations in terms of clinical gene therapy. An important obstacle is the high level of uncontrolled expression of transgenes introduced using Ad vectors. Indeed, our results showed that the unregulated expression of EPO, although protective of islet function, affected erythropoiesis in the animals that underwent transplantation, leading to their unexpected general loss of health. Recipients of 2,000 IEN EPO-transduced islets showed significant increase in erythrocytes, hemoglobin, and platelets, which is associated with splenomegaly. Obtaining controlled levels of circulating EPO by the use of inducible promoters may prove a viable strategy to pursue. Furthermore, examining the potential of a more therapeutically relevant gene delivery system to introduce EPO, including the adeno-associated virus, feline immunodeficiency virus, or non-viral vectors may be advantageous.
We have shown that a non-oncogenically derived gene confers cytoprotection to islets without metabolically damaging the cells. The unique properties of EPO when compared with other anti-apoptotic molecules is that recombinant EPO is approved for clinical use in a variety of diseases not limited to erythropoiesis (27). In transducing islets with Ad-EPO, we examined possible cytoprotective effects of the gene. Our findings suggest that EPO protects pancreatic islets both in culture and in the transplant setting.
The effect of EPO has been shown to be mediated exclusively by its binding to the EPO receptor in which cytoprotective signals are mediated (28). It remains to be determined whether EPO causes a STAT-5 translocation in islets as is seen in other cells or whether EPO is acting through alternate signal transduction pathways. Given the high level of EPO secretion into the islet microenvironment, the cytoprotection seen in this study is probably paracrine or autocrine in nature. Because the mitochondrial apoptotic pathway and release of oxidative products are activated by serum deprivation (29), EPO in this scenario likely protects against oxidative damage. It will be important to determine the long-term safety of this gene in transplanted islets and to control its level of expression before clinical applications are to be envisioned.
Our sincere appreciation to Kevin Johnson for histologic preparations, Beata Frydel and Brigitte Shaw for confocal microscopy, and the University of Pittsburgh vector core facility for cloning and generating the adenovirus used in this study.
1. Notkins AL. Immunologic and genetic factors in type 1 diabetes. J Biol Chem 2002; 277: 43545–43548.
2. Shapiro AM, Lakey JR, Ryan EA, et al. Islet transplantation in seven patients with type 1 diabetes mellitus using a glucocorticoid-free immunosuppressive regimen [see comments]. N Engl J Med 2000; 343: 230–238.
3. Pileggi A, Ricordi C, Alessiani M, et al. Factors influencing Islet of Langerhans graft function and monitoring. Clin Chim Acta 2001; 310: 3–16.
4. Biarnes M, Montolio M, Nacher V, et al. Beta-cell death and mass in syngeneically transplanted islets exposed to short- and long-term hyperglycemia. Diabetes 2002; 51: 66–72.
5. Giannoukakis N, Mi Z, Rudert WA, et al. Prevention of beta cell dysfunction and apoptosis activation in human islets by adenoviral gene transfer of the insulin-like growth factor I. Gene Ther 2000; 7 ( 23): 2015–2022.
6. Contreras JL, Bilbao G, Smyth CA, et al. Cytoprotection of pancreatic islets before and early after transplantation using gene therapy. Kidney Int 2002; 61 (Suppl 1): 79–84.
7. Rabinovitch A, Suarez-Pinzon W, Strynadka K, et al. Transfection of human pancreatic islets with an anti-apoptotic gene (bcl-2) protects beta-cells from cytokine-induced destruction. Diabetes 1999; 48: 1223–1229.
8. Schattner EJ. Apoptosis in lymphocytic leukemias and lymphomas. Cancer Invest 2002; 20: 737–748.
9. Garcia-Ocana A, Takane KK, Reddy VT, et al. Adenovirus-mediated hepatocyte growth factor expression in mouse islets improves pancreatic islet transplant performance and reduces beta cell death. J Biol Chem 2003; 278: 343–351.
10. Cebrian A, Garcia-Ocana A, Takane KK, et al. Overexpression of parathyroid hormone-related protein inhibits pancreatic beta-cell death in vivo and in vitro. Diabetes 2002; 51: 3003–3013.
11. Fenjves ES, Ochoa MS, Cabrera O, et al. Human, nonhuman primate, and rat pancreatic islets express erythropoietin receptors. Transplantation 2003; 75: 1356–1360.
12. Wu H, Klingmuller U, Acurio A, et al. Functional interaction of erythropoietin and stem cell factor receptors is essential for erythroid colony formation. Proc Natl Acad Sci U S A 1997; 94: 1806–1810.
13. Sato T, Maekawa T, Watanabe S, et al. Erythroid progenitors differentiate and mature in response to endogenous erythropoietin. J Clin Invest 2000; 106: 263–270.
14. Socolovsky M, Fallon AE, Wang S, et al. Fetal anemia and apoptosis of red cell progenitors in Stat5a-/-5b-/- mice: a direct role for Stat5 in Bcl-X(L) induction. Cell 1999; 98: 181–191.
15. Silva M, Grillot D, Benito A, et al. Erythropoietin can promote erythroid progenitor survival by repressing apoptosis through Bcl-XL and Bcl-2. Blood 1996; 88: 1576–1582.
16. Celik M, Gokmen N, Erbayraktar S, et al. Erythropoietin prevents motor neuron apoptosis and neurologic disability in experimental spinal cord ischemic injury. Proc Natl Acad Sci U S A 2002; 99: 2258–2263.
17. Juul SE, Joyce AE, Zhao Y, et al. Why is erythropoietin present in human milk? Studies of erythropoietin receptors on enterocytes of human and rat neonates. Pediatr Res 1999; 46: 263–268.
18. Ogilvie M, Yu X, Nicolas-Metral V, et al. Erythropoietin stimulates proliferation and interferes with differentiation of myoblasts. J Biol Chem 2000; 275: 39754–39761.
19. Siren AL, Fratelli M, Brines M, et al. Erythropoietin prevents neuronal apoptosis after cerebral ischemia and metabolic stress. Proc Natl Acad Sci U S A 2001; 98: 4044–4049.
20. Lehmann R, Fernandez LA, Bottino R, et al. Evaluation of islet isolation by a new automated method (Coulter Multisizer Ile) and manual counting. Transplant Proc 1998; 30: 373–374.
21. Giannoukakis N, Mi Z, Gambotto A, et al. Infection of intact human islets by a lentiviral vector. Gene Ther 1999; 6: 1545–1551.
22. Harr SD, Uint L, Hollister R, et al. Brain expression of apolipoproteins E, J, and A-I in Alzheimer’s disease. J Neurochem 1996; 66: 2429–2435.
23. Pileggi A, Molano RD, Berney T, et al. Heme oxygenase-1 induction in islet cells results in protection from apoptosis and improved in vivo function after transplantation. Diabetes 2001; 50: 1983–1991.
24. Spivak JL. Erythropoietin use and abuse: when physiology and pharmacology collide. Adv Exp Med Biol 2001; 502: 207–224.
25. Zhang F, Hackett NR, Lam G, et al. Green fluorescent protein selectively induces HSP70-mediated upregulation of COX-2 expression in endothelial cells. Blood 2003; 102: 2115–2121.
26. Tsuchiya D, Hong S, Matsumori Y, et al. Overexpression of rat heat shock protein 70 is associated with reduction of early mitochondrial cytochrome C release and subsequent DNA fragmentation after permanent focal ischemia. J Cereb Blood Flow Metab 2003; 23: 718–727.
27. Cerami A. Beyond erythropoiesis: novel applications for recombinant human erythropoietin. Semin Hematol 2001; 38: 33–39.
28. Bittorf T, Seiler J, Ludtke B, et al. Activation of STAT5 during EPO-directed suppression of apoptosis. Cell Signal 2000; 12: 23–30.
29. Roucou X, Antonsson B, Martinou JC. Involvement of mitochondria in apoptosis. Cardiol Clin 2001; 19: 45–55.