Zinc ion (Zn2+) is highly concentrated in secretory granules of pancreatic beta islet cells [1, 2]. We have previously reported that Zn2+ in secretory granules of pancreatic beta cells could act as a paracrine effector in pancreatic islet cell death after release from islet beta cells [3]. Consistent with this idea, we have shown that chelation of Zn2+ could decrease the development of diabetes after streptozotocin (STZ) treatment suggesting the role of Zn2+ release in the secondary islet cell death following direct primary death by STZ [3]. This idea was originally proposed in the central nervous system for neuronal death after cerebral ischaemia or prolonged seizures. In such models, Zn2+ concentrated in synaptic vesicles was considered to be translocated to degenerating postsynaptic neurons and chelation of released Zn2+ abrogated neuronal injury [4, 5, 6]. Whereas the mechanism of Zn2+-induced cell injury has not been clearly understood, recent papers suggested a possible role of disturbance in the energy metabolism in Zn2+-induced neuronal cell death because Zn2+ could inhibit key glycolytic enzymes such as glyceraldehyde-3-phosphate dehydrogenase (GAPDH) in vitro and Zn2+-induced neuronal cell death was decreased by pyruvate or other intermediates of glucose metabolism [7, 8]. However, other mechanisms such as repletion of NAD+ or reactive oxygen species (ROS) scavenging by pyruvate have also been implicated.

We conducted this investigation to study if pyruvate could inhibit Zn2+-induced death of pancreatic beta cells. We further explored if pyruvate infusion could inhibit the development of diabetes after STZ treatment in which Zn2+ has been reported to play a role as a secondary death effector [3].

Materials and methods

MTT assay

As a model of pancreatic beta cells, SV40 T-transformed insulinoma cells derived from non-obese diabetic (NOD) mice [9] were used (MIN6N8). MIN6N8 cells were cultured in DMEM-15% FCS containing 2 mmol/l glutamine and penicillin-streptomycin (Gibco-BRL, Gaithersburg, Md., USA). Cell death was measured using the 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT) (Sigma, St Louis, Mo., USA) assay. Cells were seeded in 96-well plates (3×104/well) and treated with ZnCl2 for 24 h; then the medium was removed, and 100 µl of 12.2 mmol/l MTT solution was added to each well. After incubation at 37°C for 4 h, crystals were precipitated by brief centrifugation. The crystals were dissolved in DMSO (Merck, Darmstadt, Germany), and absorbance at 570 nm was measured using an ELISA microplate reader (Molecular Devices, Sunnyvale, Calif., USA). Necrosis of ME-180 cervical cancer cells was induced by treating them with a combination of 100 U/ml IFN-γ and 2700 U/ml TNF-α in the presence of 50 µmol/l N-benzyloxycarbonyl-Val-Ala-Asp.fluoromethylketone (z-VAD.fmk) (Enzyme Systems, Livermore, Calif., USA) [10]. Recombinant human IFN-γ was purchased from R&D Systems (Minneapolis, Minn., USA). Recombinant human TNF-α was generously provided by Dr. T.H. Lee (Yonsei University, Seoul, Korea). All other chemicals were from Sigma unless stated otherwise.

Morphological analysis of dead cells

Cells were double-stained with 4.1 µmol/l Hoechst 33342 and 3.7 µmol/l propidium iodide (PI) (Molecular Probes, Eugene, Ore., USA) to distinguish apoptotic cells from necrotic cells. Cells with intact blue nuclei, condensed/fragmented nuclei and intact pink nuclei were considered as viable, apoptotic and necrotic cells, respectively [10].

Measurement of ATP contents

Intracellular ATP contents were measured using a commercial kit (Sigma). In brief, luminescence from ATP in cells lysed with a premade reagent was measured using a luminometer (Promega, Madison, Wis., USA). ATP contents in the sample were calculated as (ATPInternal Standard×LSample)/(LSample + Internal Standard−LSample).

Measurement of GAPDH activity

We added 25 µg cytosolic sample protein to the 1 ml reaction mixture containing 100 mmol/l sodium pyrophosphate, pH 8.5, 20 mmol/l sodium phosphate, 0.25 mmol/l NAD+, 3 µmol/l dithiothreitol and 16 µmol/l glyceraldehydes-3-phosphate. After incubation at 25°C for 5 min, absorbance at 340 nm was measured and NADH concentration was calculated according to Beer's law [11].

In vivo administration of pyruvate

STZ (245.1 µmol/kg in 0.1 mol/l citrate buffer, pH 4.5) was injected intraperitoneally to Sprague-Dawley rats after overnight fasting. Pyruvate solution (150 mmol/l) was started through the tail vein 1 h before STZ injection and was continued for an additional 24 h at a rate of 3 ml·hour−1·kg−1. Control rats were infused with the same amount of normal saline. Glucose was added to the infusion solution 2 h after STZ administration. Blood glucose concentrations were measured using the glucose oxidase method. Non-fasting blood glucose concentrations above 14.4 mmol/l were considered diabetic. All animal experiments were conducted in accordance with an institutional guideline of Samsung Medical Center Animal Facility, an Association for Assessment and Accreditation for Laboratory Animal Care International-accredited facility.

Quantitation of beta-cell mass

Formalin-fixed sections of the rat pancreata were deparaffinized and briefly microwaved in 0.01 mol/l sodium citrate buffer (pH 6.0). They were then incubated with an appropriate dilution of anti-porcine insulin antibody (Ab) (DAKO Japan, Kyoto, Japan) after goat serum blocking. Incubation with biotinylated anti-guinea pig IgG Ab, and then with avidin-biotin-peroxidase complex (Vector, Burlingame, Calif., USA) followed, diaminobenzidine was used as a colour substrate. Point counting morphometry on anti-insulin Ab-stained sections was used to calculate the relative beta-cell volume as a measure of beta-cell mass after STZ treatment [12].

Statistical analysis

Binomial test was used to compare the incidences of diabetes between two groups. Student's t-test was used to compare the mean blood glucose concentrations or relative beta-cell volumes between the groups. Repeated measure analysis of variance (ANOVA) was used to test the effect of pyruvate infusion on the blood glucose concentrations at multiple points. In all cases of multiple statistical analyses, p values were corrected by Bonferroni's method. p values less than 0.05 were regarded as being statistically significant. All results were expressed as means ± SD. All in vitro experiments were carried out more than three times to ensure reproducibility of the experiments. As independent in vitro experiments showed similar tendency, intra-assay means ± SD were used for statistical analysis.

Results

Inhibition of Zn2+-induced insulinoma cell necrosis by pyruvate

Firstly, we explored if pyruvate that has been reported to protect neuronal cells against Zn2+-induced death could affect Zn2+-induced death of MIN6N8 insulinoma cells. Two hundred µmol/l Zn2+ induced death of MIN6N8 cells as reported (Fig. 1A). Zn2+-induced death of MIN6N8 cells was classical necrosis as shown by the uptake of PI without nuclear condensation/fragmentation in 86% of MIN6N8 cells treated with 200 µmol/l Zn2+ (Fig. 1B,C). Pyruvate decreased Zn2+-induced MIN6N8 cell death (necrosis) in a dose-dependent manner and 10 mmol/l pyruvate inhibited more than 80% of insulinoma cell death by Zn2+ (Fig. 1A–C). We then studied if pyruvate could inhibit target cell death in other types of necrosis or apoptosis. Treatment with 15 mmol/l STZ for 24 h induced death (mostly necrosis) of about 60% of insulinoma cells as identified by MTT assay and Hoechst 33342/PI double staining (Fig. 2). However, 10 mmol/l pyruvate did not inhibit MIN6N8 cells death by STZ. It also did not attenuate necrosis of ME-180 cervical cancer cells by IFN-γ/TNF-α combination in the presence of z-VAD.fmk, while more than 75% of target cells underwent necrosis on Hoechst 33342/PI staining . Pyruvate did not inhibit apoptosis of insulinoma by 100 µmol/l etoposide or 1 µmol/l staurosporine, suggesting that the effect of pyruvate is specific for certain types of necrosis such as that induced by Zn2+ (Fig. 2).

Fig. 1A–C.
figure 1

Abrogation of Zn2+-induced necrosis by pyruvate. (A) Pyruvate inhibited Zn2+-induced MIN6N8 cell death in a dose-dependent manner. (B, C) Hoechst 33342/ PI double staining (×400) showed that Zn2+ exerted necrosis on MIN6N8 cells without apoptotic component, which was abrogated by pyruvate (white bars for apoptosis and black bars for necrosis). Results are representative of three or more independent experiments performed in triplicate, showing similar tendency

Full size image
Fig. 2.
figure 2

Effects of pyruvate on other types of necrosis or apoptosis. Pyruvate failed to inhibit necrosis of MIN6N8 cells (black bars) by STZ or that of ME-180 cervical cancer cells (white bars) by IFN-γ/TNF-α combination in the presence of z-VAD.fmk, a pancaspase inhibitor. Pyruvate also did not inhibit apoptosis of MIN6N8 cells by etoposide or staurosporine. Results are representative of three or more experiments carried out in triplicate

Full size image

Protection of intracellular energy metabolism by pyruvate

We also studied if Zn2+ or pyruvate affects the energy metabolism of MIN6N8 cells because Zn2+ has been reported to inhibit glycolysis by blocking GAPDH. Treatment of MIN6N8 cells with 200 µmol Zn2+ induced an abrupt decrease in ATP contents. Pyruvate dramatically inhibited the decrease in ATP contents by Zn2+ treatment, suggesting that failure in energy metabolism plays an important role in Zn2+-induced insulinoma cell death and pyruvate decreases Zn2+-induced cell death by normalizing the energy metabolism (Fig. 3A). The decrease in ATP contents of MIN6N8 cells after Zn2+ treatment and its rescue by pyruvate paralleled the temporal pattern of MIN6N8 cell death (Fig. 3B). To further prove that Zn2+ induces MIN6N8 cell death by inhibiting GAPDH and disrupting the energy metabolism, we directly measured GAPDH activity after Zn2+ treatment of MIN6N8 cells. As hypothesized, Zn2+ treatment for 6 to 24 h decreased GAPDH activity of MIN6N8 cells in a similar time frame (Fig. 3C).

Fig. 3A–C.
figure 3

Effects of Zn2+ and pyruvate on intracellular ATP contents and GADPH activity. (A) Zn2+ (squares) induced a rapid decline of intracellular ATP contents, which was abrogated by the addition of pyruvate(triangles). Temporal change in ATP contents induced by Zn2+ and pyruvate (A) paralleled that in MIN6N8 cell viability (B). Pyruvate alone (circles) did not significantly affect ATP contents or viability (A, B). (C) Treatment of MIN6N8 cells with Zn2+ decreased GAPDH activity in a similar time frame

Full size image

As these results suggested the failure of ATP production as a probable cause of Zn2+-induced insulinoma cell death, we studied the effect of tricarboxylic acid (TCA) cycle intermediates on Zn2+-induced MIN6N8 cell death. The introduction of 10 mmol/l of oxaloacetate or α-ketoglutarate increased MIN6N8 cell viability after Zn2+ treatment, suggesting that acute depletion of ATP is responsible for insulinoma cell necrosis after Zn2+ treatment and its correction protects them from Zn2+-induced necrosis (Fig. 4). Also 5 mmol/l dimethylester of another TCA cycle intermediate succinate (succinic acid dimethyl ester, SAD) that penetrates efficiently into pancreatic islet cells inhibited Zn2+-induced MIN6N8 cell death (Fig. 4). As pyruvate, oxaloacetate and α-ketoglutarate are ROS scavengers and their protective effect might be related to the removal of oxygen radicals produced by Zn2+, we asked if antioxidants could inhibit Zn2+-induced insulinoma cell death. Classic antioxidants such as 100 µmol/l Trolox and 10 µmol/l butylated hydroxyanisole (BHA) did not significantly affect MIN6N8 cell death by Zn2+. In addition, the effect of α-ketobutyrate that has ROS scavenger properties but does not act as an energy substrate on Zn2+-induced MIN6N8 cell death was much less than that of pyruvate, oxaloacetate or α-ketoglutarate, suggesting that the effect of pyruvate or TCA cycle intermediates is not related to the scavenging of oxygen radicals (Fig. 4).

Fig. 4.
figure 4

Effects of energy substrates and antioxidants on Zn2+-induced MIN6N8 cell death. Oxaloacetate, α-ketoglutarate or SAD as TCA cycle intermediates or their cell-permeable ester substantially inhibited the necrosis of MIN6N8 cells by Zn2+, while antioxidants such as BHA, Trolox or α-ketobutyrate did not notably affect their death compared to pyruvate, oxaloacetate or α-ketoglurarate. Results are representative of three or more experiments carried out in triplicate

Full size image

As pyruvate could increase the cytoplasmic content of NAD+ through its conversion to lactate in the presence of sufficient lactate dehydrogenase (LDH) and the increase in NAD+ contents has been suggested to be an important mechanism of pyruvate-mediated protection of neuronal cells, we hypothesised if methylpyruvate that is not metabolized in cytoplasm but acts as an insulin secretagogue by generating ATP in mitochondria could protect insulinoma cells against Zn2+ -induced necrosis. The addition of 10 mmol/l methylpyruvate abrogated Zn2+-induced insulinoma cell death like pyruvate (Fig. 5A). It also protected ATP levels in Zn2+-treated insulinoma cells, which suggests that the protective effect of pyruvate against Zn2+-induced insulinoma cell death is due to the restoration of ATP levels rather than an increase in NAD+ contents and might be related to an insufficient LDH in islet/insulinoma cells (Fig. 5B). Additionally, 10 mmol/l α-ketoisocaproate (KIC) that is exclusively metabolized in mitochondria and stimulates ATP production like methylpyruvate inhibited Zn2+-induced MIN6N8 cell death, indicating the role of ATP metabolism in Zn2+-induced insulinoma cell death. The addition of 10 mmol/l lactate that is not converted to pyruvate in insulinoma/islet cells because of insufficient LDH did not inhibit Zn2+-induced necrosis of insulinoma cells (Fig. 5A).

Fig. 5A, B.
figure 5

Inhibition of Zn2+-induced MIN6N8 cell death by methylpyruvate. (A) Methylpyruvate that is not metabolized in cytoplasm but generates ATP in mitochondria or KIC that is exclusively metabolized in mitochondria and stimulates ATP production abrogated MIN6N8 cell necrosis by Zn2+. (B) Methylpyruvate protected intracellular ATP contents after Zn2+ treatment as efficiently as pyruvate. Results are representative of three or more experiments carried out in triplicate

Full size image

Inhibition of STZ-induced diabetes by pyruvate infusion

Finally, we studied if pyruvate could inhibit the development of diabetes after STZ administration that is considered to be partly mediated by Zn2+. All six control rats to which normal saline was infused became diabetic 48 h after intraperitoneal injection of STZ. In contrast, continuous pyruvate infusion for 24 h decreased the incidence of diabetes to 37.5% (3/8) at both 48 h and 72 h after STZ treatment as revealed by binomial test with Bonferroni's correction (p<0.05, p<0.05). When the blood glucose concentrations were compared, repeated measure ANOVA showed that general profile of the glucose levels was changed by pyruvate infusion (p<0.05). Individual Student's t-test with Bonferroni's correction showed that the glucose levels were lower in pyruvate-infused rats (n=8) compared to control rats (n=6) at 72 h after STZ treatment (12.1 mmol/l±7.6 mmol/l vs 23.7 mmol/l±6.6 mmol/l) (p<0.05) (Fig. 6). At 48 h, pyruvate infusion lowered glucose concentrations after STZ treatment but the difference was not statistically significant (13.5±8.8 mmol/l vs 21.6±3.8 mmol/l, p>0.1). The relative beta-cell volume 3 days after STZ treatment was 0.580±0.289% (n=7) in pyruvate-infused rats, which was higher than that in control rats (0.250±0.224%, n=6) to which normal saline was infused (p<0.05), suggesting that pyruvate inhibits diabetes after STZ treatment through the inhibition of islet cell death by Zn2+ released after the initial islet cell injury (Fig. 7A,B).

Fig. 6.
figure 6

Effects of in vivo pyruvate administration on STZ-induced diabetes. Pyruvate infusion decreased the incidence of diabetes at 48 h and 72 h after STZ injection (* p<0.05). Blood glucose concentrations at 72 h after STZ administration were also lower in pyruvate-infused rats (triangles) compared to control rats (squares) to which normal saline was infused (# p<0.05)

Full size image
Fig. 7A, B.
figure 7

Attenuation of STZ-induced beta-cell injury by pyruvate infusion. Beta-cell volume measured by point counting of the cells stained with anti-insulin Ab 3 days after STZ treatment (A) was higher in pyruvate-infused rats compared to control rats (* p<0.05) (×200) (B)

Full size image

Discussion

We have shown that pyruvate inhibited insulinoma cell necrosis by Zn2+ in vitro, which is similar to the effect reported in central neurons [8]. Simple chelation of Zn2+ by pyruvate is unlikely because the log stability constant for zinc pyruvate is very low [21]. Pyruvate also has been reported to inhibit neuronal injury by death effectors other than Zn2+ [14, 22] and reperfusion necrosis of cardiac cells [23, 24]. An acute decrease in ATP contents by Zn2+ and its reversion by pyruvate observed in this investigation is consistent with previous papers showing the inhibition of GAPDH by Zn2+ in other types of cells [7, 8]. Previous reports have suggested the role of ATP as a switch between apoptosis and necrosis in that depletion of ATP contents below 50%, suppressed caspase activation and DNA fragmentation [25, 26]. Consistent with such contention, treatment of MIN6N8 cells with etoposide did not induce an early decrease in ATP contents, while it already exerted substantial apoptosis 12 h after treatment. However, other forms of necrosis were not inhibited by pyruvate such as STZ-induced insulinoma cell necrosis or cytokine-induced necrosis of ME-180 cells in the presence of caspase inhibitors [10]. The reason for the inhibition of only a certain type of necrosis by pyruvate is not clearly understood. While detailed biochemical consequences or pathways of necrosis are not clearly dissected, necrosis as a morphological definition could entail complex heterogeneous events. For instance, necrosis of beta islet cells by STZ is due to poly(ADP-ribose) polymerase (PARP) activation followed by NAD+ depletion [27, 28], whereas that of lymphocytes or certain cancer cells by TNF family members without caspase activation reportedly involves receptor-interacting protein (RIP) or lysosomal protease such as cathepsin B [29, 30]. Thus, only a certain type of necrosis might be critically affected by pyruvate or specific energy substrates. Pyruvate also could be able to correct energy metabolism only in certain types of cells.

Our observation that TCA cycle intermediates such as oxaloacetate or α-ketoglutarate inhibited Zn2+-mediated insulinoma cell necrosis further supports the critical role of energy metabolism in Zn2+-induced necrosis. SAD, an ester of another TCA cycle intermediate succinate, that efficiently penetrates into pancreatic islet cells and participates in energy metabolism [13] also inhibited insulinoma cell death by Zn2+. On the other hand, α-ketobutyrate, a structural homologue of α-ketoglutarate, which has ROS scavenging effect without metabolic function or other antioxidants such as BHA or Trolox failed to enhance insulinoma cell viability after Zn2+ treatment. These results suggest that ROS does not play an important role in Zn2+-induced necrosis of MIN6N8 cells. In contrast, previous papers reported roles for oxygen radicals in Zn2+-induced death of neuronal cells [31, 32] and the capability of pyruvate as an ROS scavenger [14]. These discrepancies might reflect the difference in the cell types studied. The role of NAD+ as a mediator of protection by pyruvate against Zn2+-induced injury also has been reported in neuronal cells [8]. However, no substantial amount of NAD+ is likely to be produced directly from pyruvate in islet/insulinoma cells because LDH is scarce in islet cells [18]. The protective effect of oxaloacetate or α-ketoglutarate against Zn2+-mediated cell death also cannot be explained by the changes in intracellular NAD+ levels. Nicotinamide or 3-aminobenzamide that inhibits STZ-induced islet cell death by inhibiting PARP and protecting NAD+ levels [27, 28] also did not enhance insulinoma cell viability after Zn2+ treatment. If pyruvate protects insulinoma cells against Zn2+-induced necrosis by conserving energy metabolism, it should be in mitochondria and produce ATP. Pyruvate is able to penetrate into mitochondria [18, 33] and is well metabolized to yield ATP in islet cells [15, 33], which might be related to the low activity of LDH and high activity of mitochondrial glycerol phosphate dehydrogenase in islet cells allowing channeling of pyruvate and NADH toward mitochondrial oxidation [16, 17]. Our observation that methylpyruvate inhibited Zn2+-induced decline in ATP levels as efficiently as pyruvate is consistent with previous reports showing similar levels of ATP production by pyruvate and methylpyruvate [33]. Furthermore, effective inhibition of Zn2+-induced insulinoma cell death by methylpyruvate suggest that pyruvate most likely inhibits Zn2+-induced insulinoma cell death by replenishing ATP rather than generating NAD+ because methylpyruvate is known to enter mitochondria without metabolic conversion in cytoplasm and will not generate NAD+ in cytoplasm [15]. Furthermore, KIC that directly stimulates ATP production in mitochondria [19, 20] also protected insulinoma cells from Zn2+-induced death. The inability of lactate to inhibit Zn2+-induced insulinoma cell death is similar to a previous paper using neuronal cells [8]. Particularly in islet cells, lactate is not easily converted to pyruvate due to the lack of LDH [18]. MIN6 insulinoma cells have also been reported to have much lower LDH activity compared to non-beta cells, albeit slightly higher when compared to beta cells [16, 34].

Amelioration of STZ-induced diabetes by pyruvate infusion is similar to the protection of brain tissue by pyruvate against ischaemic injury [35]. The inhibition of STZ-induced diabetes by pyruvate infusion was not due to the possible effects of pyruvate on insulin secretion because the effect of pyruvate was observed even at 48 h after the cessation of pyruvate infusion (thus, at 72 hours after STZ treatment) and pyruvate is a relatively poor secretagogue of insulin secretion [33]. This result and our previous report that CaEDTA, a Zn2+ chelator, inhibited diabetes after STZ administration [3], suggests a role for Zn2+ released after primary islet insult in the development of STZ-induced diabetes. While Zn2+ is not the primary effector for islet cell death in STZ-induced diabetes and pyruvate does not affect islet cell death by STZ, Zn2+ liberated from secretory granules of beta cells might aggravate beta-cell destruction and contribute to the development of diabetes.