Modulation of JNK and p38 Stress Activated Protein Kinases In Isolated Islets of Langerhans

Abstract

Objective

The objective of this study was to determine the effects of islet isolation and cytokine exposure on e-JUN NH₂ terminal kinase (JNK) and p38 activation and whether insulin or the p38 inhibitor PD169316 could modify the response.

Summary Background Data

Islet transplantation exposes the cells of the graft to a variety of stressful stimuli that could promote β-cell death and lead to graft failure.

Methods

Islets from canine (n = 12) and cadaveric human (n = 6) pancreata were isolated and purified. Islets were cultured in CMRL 1066 with and without 100 ng/ml insulin. The response to cytokine stimulation with tumor necrosis factor (TNF)α and IL-1β and the p38 inhibitor PD169316 was also observed. Islet lysates were analyzed by Western blotting for total and phosphorylated JNK and p38 content. Apoptosis was assessed by TdT-mediated dUTP nick end labeling (TUNEL) assay and by a specific cell death enzyme-linked immunosorbant assay (ELISA).

Results

In unstimulated islets, JNK activity was highest immediately following isolation, declining over 3 days to a low baseline level. The activity of p38 was lowest immediately after isolation, increasing progressively with time. The addition of insulin resulted in a more rapid decline in JNK activity, as opposed to p38, which showed no decrease in phosphorylation in response to insulin. In the cytokine stimulation studies, IL-1β stimulated p38 activation in a dose dependent manner, while JNK was relatively unaffected. PD169316 (100μg/ml) was able to inhibit p38 activation in response to the isolation procedure as well as cytokine stimulation. Apoptotic activity was highest 24 hours after isolation, and was significantly reduced when islets were maintained in insulin-supplemented medium.

Conclusions

Inhibition of the stress-activated protein kinase (SAPK) pathways may be important for the maintenance of islet cell survival following islet isolation for transplantation. This study supports an autocrine role of insulin in this process.

While pancreas transplantation has achieved consistent success, the use of cellular grafts of purified islets of Langerhans has provided only transient reversal of hyperglycemia in a small proportion of diabetic recipients. ¹ A major cause of failure in attempts at clinical islet allotransplantation is primary graft nonfunction, although insufficient islet mass or suboptimal islet viability are thought to contribute as well. In experimental models of islet transplantation, an unexplained loss of graft β-cell mass has been observed ^2-4 and attributed, at least in part, to β-cell apoptosis. ⁵ Apoptotic activity within human islets has been shown to be present and to increase in the immediate post-isolation period. ⁶ The extent to which this activity might represent a natural response to unfavorable conditions in the extracellular milieu both in vitro and at the graft site is unknown.

The emphasis on preservation of islet function has turned attention toward the factor(s) promoting islet engraftment or minimizing graft cell injury. The hyperglycemic environment and the inflammatory response at the graft site are recognized to adversely influence engraftment. ^1,7. The concept of β-cell rest has evolved from the recognition that high metabolic activity predisposes β-cells to damage by inflammatory mediators in vitro, ⁸ in vivo ^9,10 and to type I diabetes in humans. ¹¹ The importance of insulin itself as an autocrine regulator of β-cell function has also been suggested. ¹² Exogenous insulin, either by reducing the metabolic demand on transplanted β-cells, or by directly modifying their response to injury, could be an important mediator of survival in the post-transplant period.

The intracellular pathways activated by a variety of stressful stimuli demonstrate a large degree of signal convergence. The mitogen activated protein (MAP) kinases are serine/threonine kinases that mediate intracellular phosphorylation events triggered by various extracellular signals, including growth and differentiation factors, cytokines and integrin-mediated cell attachment. ¹³ Recently, MAP kinase subgroups have been identified. Both the JNK (SAPK) family ^14-16 and p38 (SAPK2) are activated by extracellular stressors, including cytokines, heat and osmotic shock, endotoxin, and UV irradiation. ^15-21 They differ in the pathways that mediate their activation, and in their substrate specificity.

JNK and p38 have also been implicated in the induction of apoptotic cell death, both in response to cytotoxic stress and to withdrawal of trophic support. ^14,22 In this regard, it has been suggested that a dynamic balance exists between the JNK/p38 pathways and extracellular regulated kinase (ERK), another member of the MAPK family, which is vital to determining the fate of the cell. ²³ This balance can be favorably influenced by insulin. ^22,24

The importance of autocrine signaling in cell survival has been described, and β-cells are known to possess insulin receptors. ^25,26 Glucose-stimulated β-cells exhibit elevated insulin receptor tyrosine kinase activity, ²⁷ suggesting a mechanism for insulin-mediated autoregulation of cell function. ²⁸

With the discovery of p38 specific inhibitors, p38 signaling in islet cells in response to isolation provides a potential clinical target for the suppression of apoptosis. These pyridinyl imidazole compounds have been shown to block p38 activity without affecting other signaling molecules, including ERK1/ERK2 and JNK signaling. ^29-31 The specificity of these compounds arises from their affinity for the adenosine triphosphate (ATP) pocket of p38. ^31,32 Inhibition of p38 with two different pyridinyl imidazole compounds (SB203580 and SB202190) has been shown to promote survival in embryonic neurons and in nerve growth factor (NGF)-stimulated PC12 cells. ³³ PD169316 has been shown to block apoptosis induced by trophic factor withdrawal in differentiated PC12 cells and Rat-1 fibroblasts. ²²

The importance of stress kinase activation in affecting islet cell survival following isolation is unknown. The purpose of these studies was to determine the nature of the stress kinase response to the pro-inflammatory cytokines IL-1β and TNFα in isolated islets, as well as the ability of insulin itself to modify stress activity in both cytokine stimulated and unstimulated islets following isolation. The ability of insulin to alter apoptotic activity in unstimulated islets was also investigated.

MATERIALS AND METHODS

Human Islet Isolation

Pancreata were retrieved from heart-beating cadaveric donors (age 18 to 45 years) at the time of multi-organ harvest for transplantation. Consent for donation of tissues for research was obtained by the local organ procurement organization. Following vascular flushing with University of Wisconsin solution perfusion, the pancreas was removed and transported to the islet isolation laboratory. Cold ischemia time was between 30 minutes and 5 hours.

Islets were isolated using the method of Ricordi, et al. ³⁴ Briefly, a 30°C solution of 2 mg/ml Liberase enzyme blend (Roche Molecular, Indianapolis, IN) in Hanks’ balanced salt solution (HBSS) (Mediatech, Herndon, VA) supplemented with 0.2 mg/ml DNAse I (Boehringer-Mannheim, Montreal, Que.) was infused into the main pancreatic duct using a syringe. The distended pancreas was placed in a sterilized aluminum digestion chamber (Bio-Rep, Miami, FL) through which HBSS supplemented with Penicillin (100,000 U/L) and Fungizone (2500 μg/L) (Gibco, Burlington, Ontario) was recirculated at 37°C. The extent of tissue digestion was assessed by staining aliquots of digestate with dithizone (Sigma, St. Louis, MO), and visualizing the islets under an inverted light microscope (Nikon, Montreal, Quebec). Cooling the circuit to between 5 and 10°C when the majority of the islets were seen to be free of surrounding acinar tissue terminated the digestion process. The digestate was collected, centrifuged (400 g) and washed three times in HBSS/FCS solution and islets were purified on a discontinuous EuroFicoll density gradient (Eurocollins solution and Polysucrose 400, Mediatech, Herndon, VA; Ficoll, Sigma, St. Louis, MO) using a COBE 2991 Cell Processor (COBE BCT, Denver, CO). ³⁵

Canine Islet Isolation

Canine pancreata were used to overcome the shortage of human tissue donors and because of the acceptable reliability and reproducibility of the canine islet isolation process. All procedures were in compliance with national animal care standards and were approved by institutional Animal Care Committees. Pancreata were harvested from 2 to 4 year old mongrel dogs under general anesthesia and sterile conditions using previously defined protocols. ³⁶ Pancreata were placed in ice-cold saline prior to dissection of fatty and connective tissue. Islet isolation and purification were performed in an identical manner as with human islets, using either Liberase CI enzyme blend or Collagenase P (both Roche Molecular, Indianapolis, IN), Collagenase type 11 (Sigma, St. Louis, MO) or Serva collagenase (Crescent Chemical, Hauppauge, NY).

Culture of Islets

Islets were cultured in suspension in CMRL 1066 (Gibco, Burlington, Ontario) and supplemented with 0.1% bovine serum albumin (Sigma, St. Louis, MO), Penicillin/Streptomycin and Fungizone (Gibco, Burlington, Ont.). Islets isolated in Indianapolis were shipped overnight in serum supplemented CMRL 1066. On arrival, islet equivalent number was reassessed, and medium was changed to serum-free medium before incubation at 37°C in a humidified atmosphere of 5% CO2. For acute cytokine stimulation studies, insulin (Roche Molecular, Indianapolis, IN) was added to the medium for a final concentration of 100 ng/ml, 18 to 24 hours prior to the cytokine exposure. PD169316 was then added to relevant samples at a concentration of 100 μM for 1 hour, after which the islet cells were lysed. PD169316 was obtained from Dr. Allan Saltiel (Parke-Davis). IL-1ß or TNFα (R&D Systems, Minneapolis, MN) in phosphate buffered saline (PBS) were added to the culture medium for a final concentration of up to 100 U/ml and 100 ng/ml, respectively. After the predetermined exposure time, islets were removed from the dish and washed twice in PBS at 4°C.

Islet Cell Lysis

Samples of islets (5000 IEQ per group) were centrifuged, washed twice in PBS at 4°C and the pellet resuspended in a Nonidet-P40 lysis buffer (Nonidet-P40, PMSF, aprotinin, sodium orthovanadate, Sigma-Aldrich Canada, Oakville, Ontario). The cells were sonicated for 15 seconds and centrifuged at 14000 rpm for 20 minutes The supernatant was recovered, frozen and stored at -20°C. Minor variations in sample size were corrected by performing a total protein assay on an aliquot of each whole cell lysate using Bradford protein dye (Bio-Rad, Mississauga, Ontario). For all western blots, equivalent amounts of protein were loaded in each well. Between blots, this varied from 75 to 100 μg/well.

Western Immunoblotting

The islet cell lysates were diluted 4:1 in 6X sample buffer (0.375 M TRIS [pH6.8]), 0.416 M SDS, 30% [v/v] Glycerol, 2.89 mM Bromophenol Blue, 12% [v/v] 2-mercapto-ethenol). The samples were boiled for 5 minutes (100°C) and 15 μ/L aliquots were loaded into 12% (v/v) Tris-glycine gels (Helixx, Toronto, Ontario) for protein separation by SDS-PAGE. The electrophoresed proteins were transferred to nitrocellulose (Bio-Rad, Mississauga, Ontario) and the membranes blocked overnight at 4°C in blocking buffer (26.4 mM Tris, 0.15 mM NaCl, 0.05% (v/v) Tween-20, 1% (w/v) bovine serum albumin, 1% chicken albumin (Sigma-Aldrich, Oakville, Ontario)). The blocked membranes were probed with anti-p38, anti-ERK1, and anti-JNK rabbit polyclonal antibodies (Santa Cruz, Santa Cruz, CA), overnight at 4°C. To evaluate kinase activation, blocked membranes were also exposed to anti-phospho p38 (Thr180/Tyr182), anti-phospho MAP-kinase (Thr202/Tyr204) (Promega, Madison, WI) and anti-phospho JNK (Thr183/Tyr185)(New England Biolabs, Mississauga, Ontario) for 48h at 4°C. Membranes were then washed, (26.4 mM Tris, 0.15 mM NaCl, 0.05% [v/v] Tween-20) and incubated with horseradish peroxidase-conjugated polyclonal anti-rabbit IgG secondary antibody (Amersham, Oakville, Ontario) (in washing buffer with 5% milk) for one hour. Proteins were visualized using the enhanced chemiluminescence (ECL) system (Amersham, Oakville, Ontario). The membranes were stripped for 30 min at 65°C in stripping buffer (0.7 % [v/v] 2-mercaptoethanol, 2% [w/v] SDS, 62.6 mM Tris, pH 6.7), washed, and reblotted. All experiments were repeated at least 3 times, with representative blots selected for quantitative analysis.

P38 MAP Kinase Assay

The activity of p38 in the islet cells was evaluated by nonradioactive p38 MAP kinase assay (New England Biolabs, Mississauga, Ontario). A phospho-specific (Thr180/Tyr182) p38 monoclonal antibody was used to immunoprecipitate the activated form of p38 from cell lysates. The immunoprecipitate is incubated with an ATF-2 fusion protein in the presence of ATP (10 mM ATP) and kinase buffer (25 mM Tris [pH 7.5], 5 mM b-Glycerol-phosphate, 2 mM DTT, 0.1 mM Na₃VO₄, 10 mM MgCl₂). Activation is detected by Western blotting with an anti-phospho-ATF-2 (Thr71) antibody. Phosphorylation of the Thr 71 site is required for transcriptional activity of ATF-2. ³⁷

Gel Quantification

The developed film was scanned and then analyzed using Scion Image ß-3b (Scion Corp., Santa Cruz, CA). Relative densities of immunoreactive bands are expressed in arbitrary units and representative bands from at least three independent experiments were selected for analysis.

DNA Fragmentation Assay

Samples of islets from each day in culture were thawed and processed for ELISA using a commercially available kit (Roche Molecular, Montreal, Quebec). The ELISA, based on a sandwich-enzyme immunoassay principle, detects histone-associated DNA fragments in the cell cytoplasm, which are characteristic of the apoptotic process. Samples were incubated in lysis buffer for 30 minutes, and then centrifuged to pellet cellular organelles and nuclei. A mouse anti-histone monoclonal antibody was adsorbed to the wall of a microtiter plate well, and nonspecific binding sites were blocked. The sample supernatant, containing cytoplasmic oligonucleosomes, was added in order to bind the histone component to the immobilized antibody. A mouse anti-DNA, peroxidase-conjugated monoclonal antibody was then added. Visualization was performed using ABTS (2,2’-azino-di-[3-ethylbenzthiazoline sulfonate]) as a substrate, with absorbance measured at 405 nm on a 96-well plate reader. Variations in sample size were corrected by measuring total sample DNA content using the fluorometric DNA assay of Downs and Wilfinger. ³⁸

Terminal Deoxynucleotidyl Transferase-Mediated dUTP Nick End Labeling (TUNEL) Assay

Fixed, unstained histologic sections were dewaxed and rehydrated by sequential washing in toluene, absolute ethanol, 95% ethanol and water. The slides were trypsinized in 0.1% trypsin in PBS at 37°C, and the endogenous peroxidase activity was blocked with 0.3% H₂O₂ in methanol, for 30 minutes at room temperature. The TUNEL reaction mixture containing the terminal deoxynucleotidyl transferase (TdT) enzyme, purified from calf thymus, was added for 60 minutes at 37°C. A solution of sheep anti-fluorescein, peroxidase-conjugated monoclonal antibody was added and visualization was achieved using aminoethylcarbazone chromogen. The sections were counterstained with hematoxylin. The TdT enzyme, nucleotide labeling mixture and anti-fluorescein antibody were purchased as a TUNEL kit (Roche Molecular, Montreal, Que.). An apoptotic index (the fraction of TUNEL positive nuclei) was determined by examining approximately 800 islet cells in each group. Results of the DNA fragmentation assay and the apoptotic index are expressed as mean ± standard error of the mean (SEM).

Immunocytochemistry

Tissue sections processed for TUNEL were also immunostained for insulin (anti-human insulin antibody [1:100], Dako, CA), using the Streptavidin Biotin Complex and developed with DAB. Negative controls were processed with nonimmune serum substituting for the primary antibodies.

Statistical Analysis

Data were compared using two-way analysis of variance (ANOVA) and a Student’s t test where applicable, with statistical significance accepted at P < .05.

RESULTS

The results of these studies were obtained from two different species; canine islets and human islets. Previous studies suggest that the two types of islets contain comparable levels of expression of MAP-kinases and have a similar response to the mechanical and chemical stress in isolation procedure. ³⁹ In the studies presented here, the responses of canine and human islets were found to be identical. Representative blots for each of the conditions studied were selected for clarity rather than to illustrate any species-specific differences. Therefore, some of the blots illustrated are from canine islets, while others are derived from human. This has been clearly indicated in the legends.

Expression and Phosphorylation of JNK and p38 in Cultured Islets - Effect of Insulin

The aim of this experiment was to determine if the stress kinases, JNK and p38, are abnormally expressed or phosphorylated following the isolation procedure and subsequent incubation in culture medium with or without insulin supplementation. Figure 1a (JNK expression) and Figure 1c (p38 expression) show relatively steady levels of expression of these kinases. The addition of exogenous insulin to the culture medium did not change the expression of p38, while JNK1 and JNK2 expression decreased over time in the presence of insulin. This effect was particularly pronounced for JNK2. In contrast to the pattern of expression of JNK, the degree of phosphorylation of this kinase changed dramatically. Following a remarkably high initial phosphorylation of JNK, there was a marked decrease within 24 hours, with a low or almost undetectable level of phosphorylation thereafter. Insulin, added at the beginning of the culture period, was effective in decreasing the level of JNK phosphorylation at 24 hours (Figure 1b). Conversely, changes in p38 phosphorylation were opposite to those of JNK, i.e., the p38 signals were initially weak and increased gradually over the 7-day culture period (Figure 1d). Moreover, insulin did not appear to have a suppressing effect on p38 phosphorylation. A small early increase in signal intensity was seen with insulin, after which the signal reached a plateau.

Figure 1. Expression and phosphorylation of MAP-kinases in canine islets. Islets were cultured in CMRL 1066 +/- 100 ng/ml insulin, for 7 days after isolation. Islet lysates were analyzed by Western blotting for total and phosphorylated (activated) forms of the stress-activated protein kinases JNK and p38 to determine the effect of insulin on these signaling elements. Representative blots show expression of JNK (A) and p38 (C) and phosphorylation of JNK (B) and p38 (D) during 7 days in culture. medium: CMRL 1066 (ctrl), with 100 ng/ml insulin (ins). Insulin had a suppressive effect on expression of JNK, particularly of the JNK 2 isoform. Insulin administration also diminished the level of activation of JNK 1 and 2 after 1 day in culture.

Differing Responses of JNK and p38 to Cytokine Stimulation

In order to assess the responsiveness of the stress kinases in islet cells following the isolation procedure, canine islets were exposed to IL-1ß or a combination of IL-1ß and TNFα. Exposure of islets to IL-1ß led to a dose-dependent enhancement of p38 phosphorylation (Figure 2b), whereas JNK phosphorylation became evident at low doses of stimulation, and increased only slightly with increasing concentrations (Figure 2a). Treatment with both IL-1ß (100 U/ml) and TNFα (100 ng/ml) was examined. Both p38 and JNK are highly phosphorylated in response to this combined cytokine treatment during the first hour. However, at the later times, the signal for JNK phosphorylation markedly decreases (Figure 2c) whereas p38 remains unchanged (Figure 2d).

Figure 2. Dose and time responses of JNK and p38 phosphorylation in canine islets exposed to cytokine stimulation. Canine islets were cultured for 48 hours after isolation and exposed to IL-1ß or a combination of IL-1ß and TNFα. P38 phosphorylation increased with increasing concentrations of stimulus. JNK became phosphorylated at low doses of IL-1ß (10 U/ml) and increasing concentrations had only a slight effect on phosphorylation (A). A more dose-dependent phosphorylation was seen in the response of p38 to increasing concentrations of IL-1ß (B). With the combination of IL-1ß (100 U/ml) and TNFα (100 ng/ml), higher levels of phosphorylation were observed. The JNK response was both rapid in onset and transient (C). The response of p38 was slower and more prolonged (D).

Phosphorylation and Kinase Activity of p38 in Response to Insulin and PD169316

The inhibitory action of PD169316 on p38 in isolated islet cells was evaluated using western blotting and a nonradioactive p38 kinase assay (Figure 3a). Pretreatment of the islets with PD169316 (100 μM) for 1 hour was sufficient to inhibit p38 activation in response to the combination of IL-1ß (100 U/ml) and TNFα (100 ng/ml). PD169316 was also able to block p38 activation in unstimulated islets following isolation and culture (Figure 3a). Insulin also activated p38 in unstimulated islets. Figure 3b illustrates a several-fold concentration dependent enhancement of phospho-p38 signal within the insulin concentration range of 10 to 1000 ng/ml.

Figure 3. (A) p38 kinase activity in cultured islets pretreated with PD169316 and subsequently exposed to cytokines [IL-1β (100 U/ml) and TNFα (100 ng/ml)]. Duplicate samples were obtained from a single isolation. The islets were cultured for 2 days after isolation. Islets were pre-treated with PD169316 (100 μM) for 1 hour, then exposed to the cytokine stimulus. P38 activity was measured by examining phosphorylation of an ATF-2 fusion protein substrate. PD169316 treatment blocked the low level of p38 activation in both unstimulated and stimulated islets, and these groups showed lower levels of phosphorylation relative to the untreated groups. (B) Stimulation of p38 activity by insulin in cultured human islets. Islets were cultured in CMRL 1066 for three days following isolation. They were then exposed to different concentrations of insulin. P38 phosphorylation was examined. A dose response of p38 phosphorylation was observed for the range of 0–1000 ng/ml insulin.

Apoptotic activity in cultured islets

Cytoplasmic oligonucleosome enrichment, as measured by ELISA, was significantly lower in islets cultured in medium supplemented with 100 ng/ml insulin than in islets cultured in insulin-free medium after 24 hours in culture (absorbance 405 nm/mg DNA: 22.66 ± 5.17 vs. 38.44 ± 11.12) and after 7 days (6.46 ± 0.25 vs. 11.04 ± 2.17) (n = 4, P < .05)(Figure 4). When examined using the TUNEL assay, the appearance of islets after 5 days in culture supplemented with 100 ng/ml insulin differed from that of islets in standard medium. In the insulin-free control group (Figure 5a), 49 ± 2% of nuclei were TUNEL positive, with the majority of the cells being centrally located within the islet. Insulin immunoreactivity was not consistently demonstrated in these sections. Therefore, a double labeling technique could not definitively identify the nature of the TUNEL positive cells. However, the topological distribution of TUNEL-positive nuclei suggested that these were primarily ß-cells. In the insulin-treated group (Figure 5b), 5 ± 1% of cells were TUNEL positive. This difference was statistically significant (P < .01).

Figure 4. Apoptotic activity in canine islets. Canine islets were cultured for 7 days in CMRL 1066 with or without supplementation of 100 ng/ml insulin. On days 0,1,3,5 and 7, islets were removed from culture and lysed. Apoptotic activity was measured by a double-antibody ELISA which detects cytoplasmic oligonucleosome content. The results were standardized by measuring total cellular DNA content, and expressed as absorbance at 405 nm/mg DNA. A peak of apoptotic activity on day 1 was followed by a decline in activity thereafter. In the insulin supplemented group ( ), significantly lower levels were observed on day 1 and day 7 relative to the control group ( ).n=4. (* and #:P < .05).

Figure 5. (A) Apoptotic activity in human islets as visualized by TUNEL on fixed, paraffin-embedded sections. Islets were cultured for 5 days in CMRL 1066 with or without supplementation of 100 ng/ml insulin. The TUNEL reaction was used to examine apoptotic nuclei. The terminal deoxynucleotidyl transferase enzyme labels single stranded DNA with dUTP, which is then detected by peroxidase-labeled antibody. Apoptotic nuclei, which have a large amount of cleaved DNA, are seen as TUNEL positive and stain red. An apoptotic index was calculated by counting approximately 800 nuclei in each group. In islets cultured in standard insulin-free medium, 49 ± 2% of nuclei were TUNEL positive, with the majority of the cells being centrally located within the islet. (B) In medium supplemented with 100 ng/ml insulin, 5 ± 1% of cells were TUNEL positive.

DISCUSSION

The transplantation of islets promises to restore normal glucose homeostasis to the insulin-dependent diabetic recipient. Primary functional graft failure continues to be a principal impediment to the success of this form of therapy. ¹ Even grafts demonstrating some function post-transplant eventually fail, as the viable β-cell population diminishes. The isolation process leads to the activation of apoptotic cell death within the islet, and this could account for the decline in β-cell mass observed with time in culture and in recipients following clinical islet transplantation. ^3,5,6

During isolation and after transplantation, islets are exposed to adverse conditions that could lead to cell death and ultimate graft failure. The pathways mediating the cellular response to these stressors include activation of the stress kinase subgroup of the MAPK’s. ^14-16 Insulin and other growth factors are known modulators of the stress kinase signal, ^23,24 and stress kinase activation may be important in the initiation of apoptotic death in response to growth factor withdrawal. ^22,23

The data presented proposes that insulin itself may promote islet cell survival, as well as helping modulate the effects of a variety of noxious stimuli. There has been no published evidence for an autocrine effect of insulin in this regard, although an autocrine effect on ß-cell function has been demonstrated, both in vitro ^25,40 and in vivo. ⁴¹ A negative effect of insulin on insulin secretion has been recognized for some time. Studies in human subjects have also confirmed the negative feedback effect. ^42-44 However, a mechanistic explanation for a short-loop feedback inhibition of insulin secretion has been elusive.

The presence of insulin receptors (IR) on intact islet α, ß-, and δ-cells ²⁶ has been established. It was subsequently shown that glucose and insulin administration induce IR tyrosine phosphorylation, as well as phosphorylation of certain elements of the insulin signaling pathways in insulinoma cell lines ²⁷ and isolated rodent islets. ⁴⁵ It was suggested that the responses seen to glucose stimulation were due to the effect of secreted insulin.

A positive effect on insulin mRNA production and insulin cell content has been more recently demonstrated, ⁴⁶ and the components of the glucose/insulin signal were further dissected by Leibiger et al. ⁴⁷ Insulin gene transcription was stimulated by insulin exocytosis, whether this resulted from glucose metabolism or not. Pathways identified as vital to this effect involved phosphatidylinositol (PI) 3’-kinase, Ca²⁺/calmodulin dependent (CaM) kinase, and p70s6 kinase. The survival mediating extracellular response kinase (ERK) and p38 did not appear to be essential for the insulin-mediated increase in insulin transcription. The role of p38 is clearly multifaceted and controversial, as it has been implicated in pathways governing insulin gene transcription via an activating effect on the promoter PDX-1 in human islets. ⁴⁸ Macfarlane et al. present more recent evidence implicating p38 in transcription of PDX-1 itself, as well as its translocation from the cytoplasm to the nucleus. They also suggest that the effects of chemical stress mimic those of glucose and that p38 is the common signal involved in these responses. ⁴⁹ Our data would suggest that insulin itself may play a role in β-cell survival and that this could take place partly through the activation of the p38 pathway, thus linking the known autocrine effects of insulin to processes governing survival.

The results presented here show distinct differences in the p38 and JNK responses in canine and human islet cells. Following isolation, both p38 and JNK are highly activated, but the duration of the signal is remarkably different; i.e., the phospho-JNK signal is immediate and transient, whereas that of phospho-p38 is delayed and sustained. While insulin is known to increase cell survival by suppressing p38 activation in growth factor deprived PC12 cells, this appears not to be the case in either canine or human islet cells. A modulatory role for insulin on stress kinases in islet cells is inhibitory in the case of JNK, and stabilizing in the case of p38, supporting the notion that cell context is an important factor in determining the response to insulin.

A peak of apoptotic activity occurring on day 1 after isolation followed the peak of JNK activity closely. The activity of JNK declined more rapidly when culture medium was supplemented with 100 ng/mL of insulin. Apoptotic activity determined by a specific cell death ELISA was significantly lower in the insulin supplemented medium after 1 and 7 days in culture, and a pronounced reduction in TUNEL positive nuclei was seen in islets cultured with supplemental insulin after 5 days. The correlation between JNK activity and apoptosis has been firmly established in other cell systems ⁵⁰ (as has the ability of insulin to inhibit JNK activation ^23,51) and JNK pathway activation by cytokines has been demonstrated in the RINm5F insulin producing cell line. ⁵² This is the first report of such activity in islet cells. Furthermore, the data appear to suggest that in order to have a maximal impact on JNK activation (and consequently on cell survival), the intervention to inhibit JNK must be initiated as early as possible, perhaps during the isolation process itself. Moreover, faster deactivation (dephosphorylation) of JNK by insulin might remove a death signal that is dependent on synergy between JNK and p38 when both are active.

In contrast to the high JNK phosphorylation (activity), p38 phosphorylation (activity) was markedly diminished on the day of isolation. This finding was consistent across isolations. The activity of p38 increased with time in culture, and this was not the result of an increase in p38 expression. It is possible that the effect of insulin on p38 is to maintain an intrinsic level of activity, thus limiting the magnitude of changes in response to other stimuli. Evidence to support this position comes from the observation that the p38 response to cytotoxic stressors in hepatocytes in vivo is one of deactivation rather than activation. ⁵³ Thus the p38 signal may behave in a bipolar manner, where the extent of activation or deactivation mediates different cellular responses.

The nature of the isolation procedure does not permit easy technical or pharmacological manipulation of the islets in order to reduce apoptosis. For this reason, cell loss caused by isolation has been an inevitable reality, both experimentally and in the limited clinical experience. Islet cell death in response to isolation could be mediated at first by transient JNK activation, then by sustained elevation of p38. Treatment with p38 or JNK suppressors could therefore be beneficial in the long-term culture of islets. Our results support the feasibility of such a notion because the treatment with PD169316 significantly reduced p38 phosphorylation and kinase activity in response to a cytokine stimulus.

Clearly, cell loss prior to engraftment, as well as continued cell loss thereafter, is a significant barrier to the success of islet transplantation, because the majority of clinical islet transplants never demonstrate any graft function. ¹ Isolation and purification of islets exposes them to mechanical, osmotic, and ischemic stresses, the precise consequences of which have been poorly characterized. In addition, isolation may remove important cell-cell and cell-matrix interactions ^54–56 contributing to maintenance of the ß-cell population in the intact pancreas. Prior to this study, the roles of the stress kinases pathways in islet cells had not been extensively studied, and nothing was known about their response to isolation, purification, and transplantation. They are, however, clearly linked to the processes of apoptosis and survival, and the ability to manipulate them may prove important (if not crucial) to the maintenance of a functioning islet cell graft.