Antiapoptotic Effect of Tacrolimus on Cytokine-Challenged Human Islets

Abstract

Our goal was to investigate whether previously related antiapoptotic and anti-inflammatory effects of tacrolimus could be useful in protecting human islets cultured in the presence of several proinflammatory mediators. Human islets obtained from cadaveric donors after intraductal infusion with collagenase, mechanical digestion, and continuous Ficoll gradient purification were cultured in RPMI-1640 medium for 24 h. Escherichia coli lipopolysaccharide (10 μg/ml) or interleukin-1 (50 UI/ml)+γ-IF (1000 UI/ml) and low-dose tacroliumus (5 ng/ml) were added. Homogenized samples (300 IE) from five different donors where assigned to four different experimental groups (control, treatment, cytokines, and cytokines + treatment). To evaluate islet damage and apoptotic response, nucleosome content, Bcl-2 protein levels, caspase-3, -8, and -9 levels, and insulin concentration were measured. Also, TNF-α and IL-6 levels where assessed as indicators of the inflammatory response. All proapoptotic markers, TNF-α, and IL-6 levels were augmented after both LPS and cytokine stimulation. Tacrolimus reduced significantly all of them and restored baseline values of nucleosome and caspase-9 in both experiments and Bcl-2 and caspase-3 when IL-1 + γ-IF was added. Twenty-four-hour insulin concentration diminished when LPS or IL-1 + γ-IF were present. Tacrolimus treatment restored insulin levels in both experiments. These results suggest that in vitro apoptotic events and media insulin concentration decrease after proinflammatory stimulation can be reverted using low-dose tacrolimus.

Introduction

Pancreatic islet transplantation is an effective alternative for type 1 diabetes mellitus patients. Recent advances in donor and recipient selection, isolation procedure, and immunosuppressive protocols have helped significantly to achieve better clinical results in terms of insulin independence and metabolic control (32,34). Along with donor shortage and persistent autoimmunity, allograft rejection remains as the most important problem to solve (31). However, the limited islet graft survival is due to, among others, several factors that generate a nonspecific inflammatory response. In addition, innate immunity seems to be involved and cause an important islet mass loss and graft failure within the first 24–48 h after transplantation (26). In fact, early primary nonfunction occurs also in autologus and syngeneic transplantation without T-cell infiltration (6) and is correlated with a massive release of IL-1, TNF-α, and γ-interferon (γ-IF) (5). Moreover, exposure to endotoxin contained in collagenase blends and host local immune response have also been pointed as important triggers for primary nonfunction mechanisms (41).

Also, it has been shown that there is a proinflammatory cytokine release from intraislet macrophages, ductal and vascular cells, and even β-cells during nonspecific immune response, which activates apoptosis of the islet cells (3,15). Mechanical, chemical, and ischemic stress generated since pancreas procurement and isolation procedures also contribute to apoptosis of the implanted islets cells (3). Not surprisingly, both unspecific inflammatory response amelioration and apoptosis blockade have shown promising results improving islet viability, function, and graft survival (2,24).

Common immunosuppressive protocols used in clinical islet transplantation include low-dose tacrolimus (32). Despite the fact this calcineurin inhibitor has been related to long-term diabetogenic side effects and β-cell regeneration impairment (2), both antiapoptotic protective and anti-inflammatory effects have been observed (36). Thus, such effects could be useful in blocking or ameliorating innate immunity-related events that prevent early graft loss due to primary nonfunction and improving graft survival (27,28).

The main aim of our study was to investigate whether the above mentioned effect of tacrolimus could be useful in protecting human islets cultured in the presence or absence of several proinflammatory mediators. In order to evaluate islet damage we measured different apoptotic markers and insulin concentration. Also, TNF-α and IL-6 levels were assessed as indicators of the inflammatory response.

Materials and Methods

Islet Isolation

Human islets were obtained from cadaveric donors in our institution. Specific consent was obtained from the donors' relatives and the local Police Court according to Spanish/European Union Regulation. Pancreatic islets were isolated following standard islet transplantation protocols described elsewhere but with research proposal only due to Spanish National Transplantation Organization restrictions. After pancreata were obtained, they were preserved in a Wisconsin University solution at 2°C. After careful dissection and extirpation of surrounding connective tissue, the main pancreatic duct was cannulated using a ¹⁴Fr Abbocath catheter. Collagenase P solution (2 mg/ ml; 1213865; Roche Diagnostics, Laval, Canada) was perfused until complete capsule distension. Enzymatic digestion process took place at 37°C in a closed chamber (Bio-Rep, Miami, FL, USA). Digestion was stopped adding 4°C Hank's solution (PAA Laboratories GmbH, Linz, Austria) four consecutive times when an approximate 200,000 islet-equivalent yield was reached. Because isolated islets were visible after 20 mg/ml dithizone (Sigma-Aldrich, St. Louis, MO, USA) red staining, viability was evaluated every 2 min by direct microscopic vision. Ficoll (Sigma Chemical Co., St. Louis, MO, USA) gradient was applied in a COBE (Denver, CO, USA) cell processor. Remnant islets were collected under direct vision and packed in 300 islet-equivalent (IE) groups.

Pancreatic Islet Culture and Cytokine Addition

Twenty-four hour islet culture was performed using RPMI-1640 (PAA Laboratories GmbH, Linz, Austria) at 37°C, 95% O₂/5% CO₂. Glucose 11.11 mM solution and bovine serum albumin (0.1%; Sigma Chemical Co.) supplements were also included. Pancreatic islet cytokine exposure was achieved after adding to the medium 10 μg/ml Escherichia coli lipopplysaccharide (E5B55 LPS; Sigma-Aldrich) or 50 IU/ml interleukin-1 (IL-1) (Boehringer Mannheim GmbH, Germany) and 1000 IU/ml γ-IF (Boehringer Mannheim GmbH). Low-dose tacrolimus (5 ng/ml; FK506; Fujisawa; Osaka, Japan) was added where indicated.

Experimental Groups

Islets from five different donors where assigned to four different experimental groups (control, treatment, stimulus, and stimulus + treatment).

Measurements

RPMI-1640 cultured samples of 300 IE were homogenized after sequential centrifugation (600 × g × 2; 5500 × g × 2; 2400 × g; 10,000 × g) to obtain microsomal and cytosolic material.

Apoptosis Assessment

For nucleosome, Bcl-2, and caspase-3, -8, and -9 levels, programmed cell death was evaluated by determination of five different related markers. Nucleosome content (IU/islet), Bcl-2-related protein levels (IU/islet), and caspase-3, -8, and -9 concentration (nmol/islet x g24 h) were measured by ELISA/spectrophotometry commercial kits (Calbiochem, La Jolla, CA, USA and Canada).

Insulin Secretion

For each experimental group insulin concentration (ng/islet) was measured in the culture medium after 24 h using a commercial kit for radioimmunoassay (Demeditec, Kiel, Germany).

Proinflammatory Mediators

TNF-α and IL-6 24-h media concentrations (pg/islet) were determined by specific ELISA kits (Biosource, Nivelles, Belgium).

Statistical Analysis

Statistical analysis was carried out with Statgraphics Plus 5.1 package (Statpoint Inc., USA) for Microsoft Windows NT. Mean comparison was performed using factorial analysis of variance (ANOVA) with post hoc Scheffé's test. Results are expressed as mean ± SEM. A 95% confidence level was considered significant.

Results

Nucleosome Apoptosis

IL-1 + γ-IF addition was associated with a significant increase of nucleosome content. In those islets treated with tacrolimus, nucleosome concentration was significantly lower than in cytokine-exposed islets, being comparable to controls. In the same way, endotoxin stimulation produced a significant increase in nucleosome content. In this case tacrolimus also diminished nucleosome content to baseline (Fig. 1).

Figure 1.

Tacrolimus ameliorates the nucleosome content increase in the presence of IL-1 + γ-IF and LPS. Bars show 24-h nucleosome content (mean ± SEM) in RPMI-1640-cultured human islets (300 IE) from five different donors in the presence of IL-1 (50 IU/ml) + γ-IF (1000 IU/ml) or LPS (10 mg/ml) and addition of tacrolimus (5 ng/ml) when indicated. Both cytokine and endotoxin addition to media was associated with a significant increase of nucleosome islet content. Also, tacrolimus showed a protective effect in ameliorating the proapoptotic response to basal values. *p < 0.05 versus controls, FK-506, and FK-506 + stimulus groups.

Bcl-2 Apoptosis

Bcl-2 expression was significantly lower after IL-1 + γ-IF stimulation. In the presence of tacrolimus Bcl-2 protein levels were restored to basal values. When LPS was added, islets showed an important descent in Bcl-2 values that were restored partially after tacrolimus treatment (Fig. 2).

Figure 2.

Tacrolimus diminishes islet Bcl-2 protein decrease in the presence of IL-1 + γ-IF and LPS. Graphic shows Bcl-2-related protein levels (mean ± SEM) in media of cultured human islets (300 IE; n = 5) in the presence of IL-1 (50 IU/ml) + γ-IF (1000 IU/ml) or LPS (10 mg/ml) and addition of tacrolimus (5 ng/ml) where indicated. Bcl-2 values were significantly lower in the presence of IL-1 + γ-IF and endotoxin. When tacrolimus was added, antiapoptotic protein levels experienced a significant increase, returning to normal levels only when IL-1 + γ-IF was present, but not in those stimulated with LPS. *p < 0.05 versus controls, FK-506, and FK-506 + stimulus groups. **p < 0.05 versus LPS + tacrolimus group.

Caspases Apoptosis

In both IL-1 + γ-IF and LPS groups, caspase-3 levels increased markedly (Figs. 3 and 4). Tacrolimus addition decreased caspase-3 values in cytokine- and LPS-stimulated groups, becoming comparable to baseline levels in IL-1 + γ-IF-exposed islets. Caspase-8 values after both cytokine and endotoxin stimulation demonstrated a sharp increase. Seemingly, tacrolimus addition diminished caspase-8 levels in both experiments. However, this effect did not achieved values comparable to those observed in nonstimulated islets. When caspase-9 levels were measured after IL-1 + γ-IF stimulation, a substantial increase was observed when compared to controls. Caspase-9 levels in stimulated islets also increased when LPS was used. Tacrolimus addition diminished caspase-9 levels down to normal values in both experiments.

Figure 3.

Tacrolimus ameliorates caspase level increase after IL-1 + γ-IF addition. Bars indicate caspase-3, -8, and -9 levels (mean ± SEM) in the media of human 24-h RPMI-cultured islets (300 IE; n = 5) in the presence of IL-1 (50 IU/ml) + γ-IF (1000 IU/ml) and tacrolimus (5 ng/ml) where indicated. All three caspase levels augmented significantly when cytokines were present. Tacrolimus diminished caspase levels in every group, returning to baseline values when caspase-3 and caspase-9 were determined. *p < 0.05 versus all others. **p < 0.05 versus cytokine + tacrolimus group.

Figure 4.

Tacrolimus reverts LPS effects on caspase levels. Figure bars show caspase-3, -8, and -9 concentration (mean ± SEM) in RPMI-cultured human islets in the presence of LPS (10 mg/ml) and/or tacrolimus (5 ng/ml). Caspase-3, -8, and -9 levels increased remarkably in all the experiments. Tacrolimus addition resulted in diminished caspase production in all groups, returning to baseline values when caspase-9 was quantified. *p < 0.05 versus all others. **p < 0.05 versus LPS + tacrolimus group

Insulin Concentration

After both LPS and IL-1 + γ-IF stimulation, 24-h insulin medium concentration decreased significantly. FK-506-treated islets showed higher insulin concentration in comparison with the stimulus-only group in both experiments. Hormone levels after low-dose tacrolimus treatment were comparable to controls in tacrolimus + LPS group (Fig. 5). In addition, control islets showed an acceptable 24-h insulin release, which can be considered as an indirect function assessment.

Figure 5.

Protective effect of tacrolimus on insulin media concentration in the presence of proinflammatory mediators. Bar charts show insulin concentration (mean ± SEM) in the media of RPMI-cultured human islets in the presence of IL-1 (50 IU/ml) + γ-IF (1000 IU/ml) or LPS (10 mg/ml) and addition of tacrolimus (5 ng/ml) when indicated. Hormone levels diminished significantly when either IL-1 + γ-IF or LPS was present. When tacrolimus treatment was used, insulin levels were restored in both experiments, showing nonstatistical differences with controls in the LPS group. *p < 0.05 versus all others. **p < 0.05 versus IL-1 + γ-IF + tacrolimus group.

TNF and IL-6 Production

Addition of IL-1 + γ-IF to cultured islets resulted in higher TNF-α values. Such increase was neutralized with low-dose tarcrolimus down to control-equivalent values. Moreover, LPS stimulation was also followed by upregulation of TNF-α release and tacrolimus treatment reduced cytokine levels to baseline (Fig. 6). Proinflammatory stimulation with IL-1 + γ-IF resulted in a significant increase of IL-6. In low-dose tacrolimus-treated islets IL-6 values were decreased, becoming comparable to those observed in controls in both experiments.

Figure 6.

Tacrolimus effect on islet cytokine release after proinflammatory stimulation. Graphics show TNF-α and IL-6 levels (mean ± SEM) in the presence of IL-1 (50 IU/ml) + γ-IF (1000 IU/ml) and LPS (10 mg/ml) and addition of tacrolimus where indicated. TNF-α levels were significantly higher when mediators were present. In tacrolimus-treated islets, TNF-α levels diminished not only in the presence of IL-1 + γ-IF but also in the LPS-stimulated group to baseline. A similar effect was observed when IL-6 levels were measured. When tacrolimus (5 ng/ml) was added to culture medium, IL-6 levels diminished down to comparable levels to controls. *p < 0.05 versus all others.

Discussion

Proinflammatory cytokine release and proapoptotic events are responsible of different pathophysiologic phenomena that lead to islet loss and primary nonfunction after transplantation (3,31). In order to reproduce in vitro a proinflammatory environment related to primary nonfunction of islet graft, cultured human pancreatic islets were challenged with cytokines and endotoxin and treated later by adding low-dose tacrolimus.

Little is known about the FK-506-protective effect against primary rejection of transplanted islets. Unspecific anti-inflammatory and antioxidant effects have been attributed to FK-506 while its in vitro and in vivo effect on apoptotic phenomena remain not completely elucidated. Our results indicate that tacrolimus has a powerful protective effect on in vitro human islets when challenged by a proinflammatory environment ameliorating apoptosis and cytokine release and increasing insulin content.

Despite the fact this is an in vitro study and results must be interpreted carefully, islets were obtained from cadaveric donors following standard real-time isolation procedure and the cytokines employed have been related as triggers in islet transplantation primary nonfunction previously (3,6). LPS is a potent activator of innate immunity response via Toll-like receptor 4 activation, among others (8). Additionally, its presence within collagenase and liberase blends has been reported as a contributor to deleterious inflammatory reaction during isolation process (9,41). Cytokine concentration used as stimulus in our experiments was similar to that previously employed (11).

Our results indicate that FK-506 exerts an antiapoptotic effect on mediator-stimulated human islets. Tacrolimus effects on apoptosis are highly variable and are tissue or even cell dependent (20). In spite of its proapoptotic effects on lymphocytes (16), when unspecific proinflammatory stimulation is present, apoptosis blockade can occur. FK-506 does not affect normal apoptotic patterns observed in involution of normal tissues or physiologic atrophy in animal models (38), and can stimulate human tissue regeneration (33). In contrast to our results, downregulation of Bcl-2 and Bcl-XL concentrations in fresh isolated (nonstimulated) islets has been reported after high-dose (50–100 ng/ml) tacrolimus, suggesting a dose–response effect that was ameliorated after promotion of antiapoptotic XIAP gene transduction (18).

As an explanation for tacrolimus effects on apoptosis, intrinsic characteristics of FK binding protein 38 (FKBP-38) have been suggested. FKBP-38 + calmodulin complex varies both expression and function of antiapoptotic agent Bcl-2 (1,10). FKBP activation and linkage to Bcl-2 occurs due to tertiary structure of the later, which presents a “wide loop” between amino acids of its first and second α-helix. FKBP–Bcl-2 interaction ends when Bcl-2 is located near to mitocondrial membrane and exerts an antiapoptotic effect. A putative nuclear migration of NF-κB has been proposed as an explanation (21,23,35).

In our study, low-dose tacrolimus augmented insulin medium content after proinflammatory stimulation without negative effects on hormone release. Our measurements include insulin accumulation in an 11 mM solution that may be considered glucose-stimulated insulin release. Nevertheless, comparison between insulin release into high (16 mM) versus low (2.8 mM) media could be useful in order to identify possible glucose-stimulated insulin release variations due to tacrolimus addition.

There is little evidence about real long-term systemic and insular side effects of low-dose tacrolimus (31). Adverse side effects of FK-506 on glycemic metabolism are well known and could explain failure to achieve long-term insulin independence after islet transplantation under novel immunosuppressive protocols (4). The present article does not include a dose–response study, but similar FK-506 concentrations used in the present article have been tested in murine and pig models without short-term impairment of either baseline or after-stimulus insulin secretion. Unfortunately, these results are not maintained after long-term follow-up (19). It has also been observed an inhibitory effect of tacrolimus on insulin secretion due to a hypothetic selective glucokinase blockade (40) without decrease of insulin content, which is consistent with our findings. Conversely, previous experiments have shown a decrease in transcription of the insulin gene in rats exposed to high-dose (0.1 mg/kg; 50 ng/ml) tacrolimus. However, this phenomenon was only seen after 5 days of tacrolimus treatment (17). These data suggest that although there are important side effects on insulin secretion after high-dose/long-term tacrolimus treatment, low-dose immunosuppression could be useful in preventing adverse primary nonfunctionrelated events within the first hours after transplantation without impairment of islet insulin release. Nowadays, promising results using nondiabetogenic FK-506 analogues such as FTY 720 have been reported (2).

A protective effect was also observed by diminishing proinflammatory mediators after stimulation. Pancreatic islets are very sensitive to cytokines and many cytoprotective strategies have been proposed to improve graft survival. The role of TNF-α in acute inflammatory response in the very first hours after islet transplantation is widely accepted (14,15,29) and its liberation during primary rejection of islet graft has different sources: host and islet-resident macrophages, β-cells, and hepatic endothelium (3). Thus, TNF-α blockade is one of the most important targets to improve islet allograft tolerance. Both monoclonal antibodies (infliximab) and soluble receptor analogues (etarnercept) have been used. Few reports about use of TNF-α blockade in human islet transplantation have been published since the Miami group experience (12). Low-dose tacrolimus effects on TNF-α release have not been thoroughly studied. Tacrolimus has shown a blockade effect on cytokine release in different unspecific inflammatory models such as severe acute pancreatitis (30) or renal cortex cells stimulated with LPS, IL-1, and TNF-α (42). Also, a protective effect against immediate blood-mediated inflammatory reaction has been found after tacrolimus administration specially after neural posttraumatic injury (37).

Tacrolimus protective effects on immediate inflammatory response have also been studied in hepatic ischemia-reperfusion models, by blocking oxidative stress and in situ liberation of IL-1 and TNF-α (13). Thus, potential amelioration of inflammatory events within the liver can be another positive effect of tacrolimus immunosuppression in islet transplantation. Previous reports have shown that immune cells proinflammatory cytokine release (IL-1, IL-6) after LPS stimulation diminishes significantly after tacrolimus treatment (22,25) but effects on stimulated pancreatic islets have not been carefully investigated. Role of IL-6 on pancreatic islets remains obscure. Although it has been accepted that this cytokine had negative effects on islets (39), Choi and collaborators have observed protective effects by diminishing other proinflammatory mediators (7). Our results indicate that FK-506 can ameliorate IL-6 after proinflammatory stimulation similar to observed in primary rejection phenomena, helping to create a favorable islet environment.

This article shows that tacrolimus protects islets from proapoptotic phenomena, insulin release impairment, and cytokine release observed after proinflammatory stimulation. Although the clinical relevance of these findings is not known, they could be influential enough to suggest modifications in the immunosuppression strategy employed in the first 48–72 h after islet transplantation.