Complete Protection Against Interleukin-1β–Induced Functional Suppression and Cytokine-Mediated Cytotoxicity in Rat Pancreatic Islets In Vitro Using an Interleukin-1 Cytokine Trap

Pancreatic islets from male Sprague-Dawley rats, bred in a local colony (BK Universal, Sollentuna, Sweden), were isolated by a collagenase digestion procedure (19). Groups of 150–200 islets were kept free-floating in culture medium RPMI-1640 (Sigma, St. Louis, MO) supplemented with 10% (vol/vol) FCS (Sigma), benzylpenicillin (100 units/ml), and streptomycin (0.1 mg/ml) for 5–7 days before use (19).

Islets in groups of ∼50 were subsequently transferred to new nonattachment culture dishes containing 2.25 ml RPMI-1640 + 0.25 ml FCS. Human IL-1 trap (18) at different concentrations was added, and after 30 min, human IL-1β (2.5 ng/ml corresponding to 25 units/ml and 150 pmol/l; PeproTech, London, U.K.) or a cytokine combination was added (25 units/ml IL-1β + 1,000 units/ml tumor necrosis factor [TNF]-α [human, PeproTech] + 1,000 units/ml interferon [IFN]-γ [rat; R&D Systems, Abingdon, U.K.]). The islets were subsequently maintained in culture for 48 h before tested. The human IL-1 trap was provided by Dr. Margaret Karow and colleagues (Regeneron Pharmaceuticals, Tarrytown, NY). The molecular weight of IL-1β is 17.0 kDa, and the molecular weight of the cytokine trap is 202 kDa. The study comprised the experimental groups as follows: control = no trap and no cytokine; TL = 150 pmol/l IL-1 trap (1× IL-1β concentration); TM = 1,500 pmol/l IL-1 trap (10 × IL-1β concentration); TH = 15,000 pmol/l IL-1 trap (100 × IL-1β concentration); TC = 150,000 pmol/l Fc protein (1,000 × IL-1β concentration); IL = IL-1β (150 pmol/l); ILTL = IL-1β (150 pmol/l) + IL-1 trap (150 pmol/l, i.e., 1 × IL-1β concentration); ILTM = IL-1β (150 pmol/l) + IL-1 trap (1,500 pmol/l, i.e., 10 × IL-1β concentration); ILTH = IL-1β (150 pmol/l) + IL-1 trap (15,000 pmol/l i.e., 100 × IL-1β concentration); ILTC = IL-1β (150 pmol/l) + Fc protein (150,000 pmol/l, i.e., 1,000 × IL-1β concentration). Cytokine mixture = IL-1β (150 pmol/l) + TNF-α (1,000 units/ml) + IFN-γ (1,000 units/ml). Cytokine mixture + TH = [IL-1β (150 pmol/l) + TNF-α (1,000 units/ml) + IFN-γ (1,000 units/ml)] + IL-1 trap (15,000 pmol/l, i.e., 100 × IL-1β concentration). Cyt. mix. + TC = [IL-1β (150 pmol/l) + TNF-α (1,000 units/ml) + IFN-γ (1,000 units/ml)] + Fc protein (150,000 pmol/l, i.e., 1,000 × IL-1β).

After the 48-h culture, a sample of the culture medium was collected. Then duplicate groups of 10 islets each were incubated during 1 h at 1.7 mmol/l glucose and during a 2nd hour at 16.7 mmol/l glucose (19). The insulin concentrations of the media were measured using a rat insulin enzyme-linked immunosorbent assay according to the manufacturer’s instructions (Mercodia, Uppsala, Sweden). In triplicate groups of 10 islets, the islet glucose oxidation rate was subsequently measured as previously described (20).

To aliquots of the culture medium (250 μl), 100 μl of a reagent solution containing 10 ml 100% acetic acid, 300 μl 10% sulfanilamide (dissolved in acetone), and 30 μl naphtylendiamine (dissolved in redistilled H₂O) was added. The reaction was carried out at 20°C for 15 min, and the absorbance was measured at 540 nm in a Labsystems iEMS spectrophotometer against a nitrite standard curve.

Islets in groups of 50 were cultured without or with the cytokine mixture and IL-1 trap or Fc control protein for 48 h. Then the islets were incubated with 10 μg/ml propidium iodide for 15 min and subsequently dissociated with 0.05% trypsin/Hanks’ balanced salt solution. Changes in light scattering and fluorescence emission were determined by flow cytometry with a Becton Dickinson FACSCalibur equipped with CellQuest software (Becton Dickinson, San Jose, CA). A total of 5,000–10,000 cells per sample were analyzed by exciting the cells at 488 nm and examining fluorescence at 650 nm. Propidium iodide stains both apoptotic and necrotic cells, the latter with high intensity. As a cell shrinks or loses volume, forward-scattered light decreases. Untreated cells were gated on a forward scatter versus propidium iodide fluorescence dot plot, and the same gates were then applied to cytokine-treated cells for calculation of the fractions of apoptotic and/or necrotic cells (referred to as dead islet cells).

From the glucose-stimulated insulin release and glucose oxidation rate experiments, a mean was calculated from each duplicate or triplicate group of islets and then considered as one separate observation. Furthermore, every observation represented different islet donors. Data are expressed as means ± SE, and groups of data were compared, using one-way repeated-measures ANOVA or Student’s t test. Statistical analysis was performed using SigmaStat (SPSS, Chicago, IL).

Addition of 150 pmol/l IL-1β to rat pancreatic islets for 48 h caused a pronounced reduction in the accumulation of insulin in the culture medium compared with control islets not exposed to the cytokine (Fig. 1; IL). This was not affected by the lowest concentration of the cytokine trap (ILTL), while it was fully prevented when the cytokine trap was added at a 10-fold (ILTM) and 100-fold (ILTH) molar excess of the cytokine trap in relation to the IL-1β concentration. As a protein control we added an Fc control protein at a concentration 1,000-fold higher than the cytokine trap, but this failed to counteract the IL-1β action (ILTC). In some initial experiments, we also tested the cytokine trap at a 1,000-fold excess, and thus a full protection against IL-1β was observed. Because we then also observed a similar protection already with 10- and 100-fold excess of cytokine trap, the 1,000-fold concentration of the trap was abandoned in the subsequent experiments; however, the TC group was nevertheless maintained at a 1,000-fold concentration. Addition of the cytokine trap (TL, TM, and TH) and the Fc control protein (TC) alone did not affect the medium insulin accumulation.

After culture for 48 h, the islets were tested in short-term experiments. When the insulin secretory rate was studied at low glucose (1.7 mmol/l), there were no significant differences among the groups, although there was a trend toward a decline in the IL and ILTC groups (Fig. 2A). On the other hand, at high glucose (16.7 mmol/l), glucose-stimulated insulin release was markedly depressed after adding IL-1β (Fig. 2B). The cytokine trap prevented this decline at all three tested concentrations, although the protection in the ILTL group was not complete. Treatment of islets in culture with the cytokine trap or the Fc control protein only did not affect the insulin release at low or high glucose.

As a test on islet glucose metabolism, the glucose oxidation rate after exposure to IL-1β and the cytokine trap was assessed (Fig. 3). The cytokine caused ∼50% reduction in the glucose oxidation rate, and this was also seen in the groups ILTL and ILTC. Again, the inhibitory action of IL-1β was counteracted by the higher concentrations of the cytokine trap.

Next, nitrite accumulation in the culture medium was determined as an indicator if NO formed (Fig. 4). IL-1β induced a strong increase in medium nitrite, and this was also the case in the groups ILTL and ILTC. The elevated nitrite level was prevented by higher concentrations of the cytokine trap. In the control groups, there was no change in the basal nitrite accumulation.

We then aimed to investigate if the IL-1 trap could also affect cytokine-induced islet cell death by fluorescence-activated cell sorter analysis after propidium iodide staining. To elicit a marked islet cell death, we used a cytokine combination (IL-1β + TNF-α + IFN-γ), since IL-1β treatment alone for 48 h does not cause much islet cytotoxicity. A representative experiment is shown in Fig. 5. Non–cytokine-exposed control cells (black area) indicates essentially viable cells (Fig. 5A), whereas cells exposed to the cytokine combination show a distribution with an increased amount of dead cells (white area). In Fig. 5B, the islets have been exposed to the cytokine mixture + IL-1 trap (100-fold × IL-1β concentration), and these cells showed a distribution nearly identical as that seen for control cells, whereas in Fig. 5C, it is shown that the Fc control failed to prevent cytokine-induced cell death. The data from three independent experiments are summarized in Fig. 6, showing the protective effect of the IL-1 trap against the cytotoxicity induced by the cytokine combination.

Finally we tested if the IL-1 trap also could affect functional effects induced by the cytokine mixture under the same experimental conditions as above (Table 1). The cytokine mixture induced a marked nitrite accumulation, and this was counteracted, although not completely, by the IL-1 trap. The glucose oxidation rate and medium insulin accumulation was strongly suppressed by the cytokine mixture, and this was fully prevented by the IL-1 trap. The insulin release at low glucose was elevated after exposure to the cytokine mixture and inhibited at high glucose. The IL-1 trap protected against these effects, although there was a trend that the insulin secretion at high glucose was not fully restored (P = 0.15). The TC control protein could not influence the functional effects induced by the cytokine mixture.

Proinflammatory cytokines and in particular IL-1β have been postulated to be involved in pancreatic β-cell destruction leading to type 1 diabetes (4–9). A large body of experimental data have shown that prolonged in vitro exposure of rodent pancreatic β-cells to elevated concentrations of IL-1β can cause both inhibition of β-cell function and cell death. Regarding isolated human islet cells, a combination of cytokines, e.g., IL-1β + TNF-α + IFN-γ, is required to obtain a similar effect. In the present study, we explored in vitro the efficacy of a multicomponent IL-1 trap in influencing IL-1β–induced suppression of rat pancreatic islet function as well as islet cytotoxicity induced by a cytokine combination. We found that addition of the IL-1 trap exerted a complete protection against IL-1β–induced inhibition of islet metabolism (glucose oxidation rate) and glucose-stimulated insulin secretion. This was most likely achieved by a prevention of IL-1β–mediated increase in NO generation, since elevated NO concentrations can be most harmful to rat islet function (21–23). However, in studies using NO synthase inhibitors and islets derived from inducible NO synthase–deficient mice, it has been shown that IL-1β also exerts suppressive actions that are NO independent (24–27). In the present study, it seems as if also the latter NO-independent suppressive actions by IL-1β were counteracted by the IL-1 trap.

Previously, we examined the IL-1Ra in a study of rat pancreatic islets with a design similar to the present investigation (28). After 48 h of coincubation, a 1,000-fold molar excess of IL-1Ra was required to completely block the effects of IL-1 on β-cell function. Based on measurements of islets DNA contents, IL-1Ra could also prevent IL-1β–induced cell loss. In subsequent studies, we further demonstrated that continuous infusion of IL-1Ra using osmotic minipumps could protect against hyperglycemia in two murine type 1 diabetes models, namely multiple low-dose streptozotocin diabetes (29) and recurrence of disease in NOD mice (10). We also attempted a gene delivery system for IL-1Ra in the recurrence of disease model, but the protective effect was only marginal, probably due to the relatively transient expression of IL-1Ra (29). The IL-1 trap not only has a superior ability to inhibit IL-1β in ex vivo assays, in humans, it also has superior pharmacokinetics, with an in vivo half-life of 5–9 days (30), when compared with IL-1Ra, which has a reported in vivo half-life of 5–6 h (31). Thus, the IL-1 trap may require less frequent dosing and/or exhibit improved in vivo efficacy than IL-1Ra.

In line with the findings using IL-1Ra (32), we observed a protective effect of the IL-1 trap when it was added at a 10-fold or higher concentration compared with IL-1β. It might be expected that a 1:1 ratio would show more of a protective action, compared with IL-1β addition alone, than it actually does. However, it is possible that a small fraction of free IL-1β in the in vitro system still exists at the 1:1 ratio, which can cause a suppressive islet response. Indeed, we have seen that a 10- to 20-fold lower concentration of IL-1β than used herein, i.e., 7.5–15 pmol/l, lead to an inhibition of the glucose oxidation rate and an elevated medium nitrite accumulation by rat pancreatic islets (33); thus, even if 95% of the IL-1β was inhibited, there would still be a signal from the remaining free 5%. Consistent with this, previous results using an MRC5 bioassay, in which the IL-1 trap has an IC₅₀ of 2 pmol/l and IL-1Ra of 73 pmol/l, indicate that when the IL-1 trap is present at 3 pmol/l, a 1:0.75 ratio with the 4 pmol/l IL-1β in the assay, ∼25% of bioactivity remained, whereas at a 1:4 ratio, 95% of bioactivity was inhibited (15). This is in contrast to IL-1Ra, which in this bioassay required a 1,000-fold excess over IL-1β to block 95% of the bioactivity. As a control protein for the IL-1 trap in the islet experiments, we used the Fc protein of human IgG1 (experimental group TC) at 1,000-fold molar excess compared with IL-1β. This protein failed to counteract any of the IL-1β–induced effects examined, suggesting that the protective actions on islet function and viability by the IL-1 trap is due to it specific inhibition of IL-1β and not due to the addition of a nonspecific protein or via cytokine binding to the Fc part of the IL-1 trap.

The IL-1 trap could also block islet cytotoxicity in response to a combination of IL-1β + TNF-α + IFN-γ. It has been speculated that the signals and/or intracellular events leading to suppressed islet cell function can be separated from the mechanisms leading to cytokine-mediated cell killing (34,35). If so, however, addition of the IL-1 trap prevented both functional suppression and cell death. This is in agreement with the view that the IL-1 trap would block the binding of IL-1β to the islet cells. Moreover, it suggests that blockage of IL-1β alone can be sufficient to counteract cell killing induced by a cytokine combination. It is unclear by which mechanism(s) the IL-1 trap reduced the cell death induced by the cytokine combination. A possibility is that a reduced NO generation has contributed to this effect. The fact that the IL-1 trap tended not to fully restore a reduction in glucose-stimulated insulin release in response to the cytokine mixture may be because some elevation of NO production has taken place under this condition. Indeed, it has been reported that the combination of TNF-α and IFN-γ can induce nitrite production and a decrease in insulin secretion from rat islets (26).

In conclusion, the present investigation shows that addition of an IL-1 trap could completely prevent rat pancreatic islets in vitro against noxious effects induced by IL-1β. Because the cytokine has been suggested to have a key role in β-cell degeneration in type 1 diabetes, testing the IL-1 trap in relevant animal models represents an interesting future therapeutic alternative.

This study was supported by grants from the Swedish Research Council (72X-8273), the Novo Nordic Fund, the Swedish Diabetes Association, the European Foundation for the Study of Diabetes, and the Family Ernfors Fund.

We appreciate Dr. Margaret Karow and colleagues (Regeneron Pharmaceuticals) for providing the IL-1 trap and for valuable discussions and comments on the manuscript. We thank Astrid Nordin and Ing-Britt Hallgren for excellent technical assistance.