The past half-century has witnessed a dramatic increase in the incidence of obesity and diabetes (Jacobson, 2004; Krilanovich, 2004). Both of these occurrences have been accompanied by an increase in the consumption of fructose. Unlike glucose, the metabolism of fructose is not subject to negative feed-back inhibition and can impose stress on intracellular energy stores (Ishimoto et al., 2012; Lanaspa et al., 2014, 2012). In the present study we identify the ability of fructose to increase the sensitivity of pancreatic beta cells to TNFα induced cytotoxicity. Exposure of pancreatic beta cells to fructose induced fructokinase and glut-5 expression, two proteins critical for the metabolism of fructose. Importantly, the increased metabolism of fructose by beta cells was accompanied by an increase in the expression of mitoNEET. MitoNEET is a 2Fe-2S cluster binding protein localized to the outer mitochondrial membrane (Wiley et al., 2007a). The increased expression of mitoNEET mediated an enhanced sensitivity of the pancreatic beta cells to TNFα induced cytotoxicity that was prevented by suppression of mitoNEET expression or pharmacological inhibition of its ability to release its 2Fe-2S cluster.
Type I diabetes is partly due to pancreatic beta cell death mediated by cytokines. Type II diabetes is also characterized by a gradual loss of beta cell mass (Cnop et al., 2005; Mulder and Ling, 2009; Muoio and Newgard, 2008). Fructose use as a sweeter has grown in the past half century and is hypothesized to contribute to the burgeoning epidemic of diabetes in the developed world; however little is known regarding its effects on beta cell viability (Dekker et al., 2010; Goran et al., 2013; Kmietowicz, 2012). The increased use of fructose has been linked to a heightened incidence of Type II diabetes (Goran et al., 2013). In contrast to the effects of fructose on the liver, the consequences of fructose exposure on pancreatic beta cells are not well characterized. Most fructose is metabolized in the liver by fructokinase. However, some fructose does enter the circulation, with fructose blood levels varying widely between 5 µM and 1.9 mM following ingestion, suggesting that other cell types can be exposed to significant levels of fructose.
Accumulating evidence suggests that the gradual loss of beta cell viability in type II diabetes is due to an infiltration of activated immune cells, such as macrophages (Ehses et al., 2008; Jourdan et al., 2013). The metabolism of fructose is not subject to negative feedback inhibition and can therefore alter the metabolic state of a cell, making it more susceptible to activation. It has been demonstrated that fructose can increase the expression of Toll-like receptors 1-9 in F4/80 positive macrophages (Wagnerberger et al., 2012). We have demonstrated that fructose exposure can lead to an increase in the activation of macrophages due to an elevation in the binding of hexokinase II to mitochondria (Shulga and Pastorino, 2014a). In turn, activated macrophages secrete chemokines and cytokines that can cause damage and loss of beta cell viability. In particular, beta cells express TNFα receptors and are capable of undergoing TNFα induced cytotoxicity (Stephens et al., 1999).
TNFα signaling is complex and can eventuate in many outcomes. However, in some instances the cell can be primed to undergo TNFα induced cell death. Indeed, we have shown that fructose can bring about an increase in the sensitivity of hepatocytes to TNFα induced cell killing (Shulga and Pastorino, 2014b). This is facilitated by an increase in the expression of mitoNEET. MitoNEET is a 2Fe-2S cluster binding protein that is localized to the outer mitochondrial membrane and mediates the maturation and shuttling of 2Fe-2S clusters (Lin et al., 2007). However, we have found that under certain circumstances, mitoNEET can mediate mitochondrial injury and cell death during exposure to TNFα. In this instance, the form of cell death is necroptotic, originating from the necrosome complex composed of RIPK-1 and RIPK-3. The necrosome phosphorylates Stat-3, which induces its interaction with Grim-19, a component of complex I of the mitochondrial respiratory chain that also resides partially in the cytosol and nucleus (Shulga and Pastorino, 2012). The resulting Stat-3-Grim-19 complex translocates to the outer mitochondrial membrane where it binds to mitoNEET, inducing a rapid and massive release of its bound 2Fe-2S complex that is quickly degraded. The resulting free iron is taken up by the mitochondria via the calcium uniporter, where it brings about increased reactive oxygen species (ROS) production and mitochondria permeability transition.
The present study demonstrates that exposure to fructose resulted in a priming of cultured pancreatic beta cells to TNFα induced cytotoxicity. Moreover, beta cells in pancreatic islets isolated from mice fed a fructose containing diet also exhibited a dramatic increase in sensitivity to TNFα induced cell killing. Fructose exposure induced the expression of mitoNEET, in addition to fructokinase and glut-5 in pancreatic beta cells. Induction of mitoNEET, fructokinase or glut-5 expression was necessary for the sensitizing effect, as their suppression prevented the enhanced susceptibility to TNFα cytotoxicity. The data indicate that the induction of TNFα induced necroptosis in beta cells brought about by fructose exposure, is partly caused by the TNFα induced binding of a Stat3-Grim-19 complex to mitoNEET. Additionally, pre-treatment with the anti-diabetic drug, pioglitazone, which binds to and inhibits the ability of mitoNEET to release its 2Fe-2S cluster, prevented the TNFα induced accumulation of mitochondrial iron, ROS formation and cytotoxicity in fructose exposed pancreatic beta cells.
Fructose exposure promotes TNFα induced necroptosis in pancreatic beta cells
To mimic the post-prandial elevation of glucose and fructose that pancreatic beta cells are exposed to, the cultured beta cells were exposed to glucose and fructose containing spikes for 30 min followed by washout with PBS, and then returned to basal RPMI media. INS-1E cells were exposed to three, 30 min spikes of 1 mM fructose and 10 mM glucose separated by 8 h intervals over 24 h. The cells were then treated with TNFα (10 ng/ml) in basal RPMI-1640 media containing 5 mM of glucose (basal condition) for the times indicated. Treatment with TNFα in cells exposed to 3 spikes of 10 mM glucose alone displayed little cytotoxicity as assessed by loss of plasma membrane integrity (PI positive) [Fig. 1A, labeled TNF(+Gluc)]. Exposure to three spikes of 1 mM fructose and 10 mM glucose in and of itself did not bring about any appreciable loss of cell viability [labeled (+Fruct.+Gluc) in Fig. 1A]. By contrast, when the cells were first exposed to the fructose containing spikes, followed by treatment with TNFα, there was a dramatic loss of cell viability, with 60% of the cells staining positive for propidium iodide at 24 h [labeled TNF(+Fruct+Gluc)]. Importantly, the cell death was prevented by necrostatin, an inhibitor of RIPK-1 (Fig. 1A, +Necro.), with the inactive analogue of necrostatin exhibiting no protective effect (+iNecro). Significantly, there was no staining of TNFα and fructose exposed cells with annexin V at 8 h (Fig. 1B), indicating that the cell death is not apoptotic. By contrast, when non-fructose exposed cells were treated with staurosporine, the cells exhibited annexin V staining by 4 h that was not prevented by necrostatin, but was prevented by ZVAD-FMK, a broad spectrum caspase inhibitor, indicating the mode of cell death in this instance was apoptotic (Fig. 1B). Moreover, caspase inhibition did not prevent the TNFα induced cytotoxicity seen in fructose exposed cells (Fig. 1A). Importantly, substitution of fructose with its non-metabolizable analog, 3-O-methylfructose, did not sensitize the cells to TNFα induced cytotoxicity (not shown).
Fructose exposure also sensitized primary pancreatic beta cells to TNFα induced necroptosis. Pancreatic islets were isolated, the cells dispersed, and incubated in basal RPMI media containing 5 mM of glucose for 16 h. The cells were then exposed to three, 30 min spikes of 1 mM fructose and 10 mM glucose at 8 h intervals. The cells were then treated with TNFα in basal non-fructose containing media for the time periods indicated. Before harvesting the cells, ethidium monoazide or annexin V was added, followed by photo-labeling and washing. The cells were then fixed and assessed for insulin or glucagon content by immunostaining. Importantly, immunostaining with insulin had no effect on annexin V staining as described previously (Medarova et al., 2005). As show in Fig. 1C, in insulin positive islet cells exposed to fructose-glucose spikes and then treated with TNFα, there was a dramatic increase in the number staining positive for ethidium monoazide at 12 and 24 h, indicating a loss of plasma membrane integrity (Fig. 1C, yellow bars). However, the cells did not stain for annexin V after 8 h (Fig. 1D, yellow bars). Importantly, this time point is prior to loss of appreciable plasma membrane integrity, where necrosis would lead to annexin V staining of the inner side of the plasma membrane. As shown in Fig. 1C, pre-treatment with necrostatin prevented the TNFα induced ethidium monoazide staining of the insulin positive islet cells exposed to fructose, but its inactive analog did not. Additionally, the ethidium monoazide staining in fructose and TNFα exposed cells was not prevented by the caspase inhibitor, ZVAD-FMK (Fig. 1C). By contrast, as shown in Fig. 1D, insulin positive islet cells treated with staurosporine displayed annexin V staining by 4 h that was not prevented by necrostatin, but was prevented by the caspase inhibitor. Importantly, islets cells staining positive for glucagon did not show a loss of viability as assessed by ethidium monoazide or annexin V, indicating that the sensitizing effects of fructose are selective for beta cells (not shown).
Pancreatic beta cells were also sensitized to TNFα induced necroptosis by in vivo exposure to fructose. Mice were fed a diet that contained 20% of fructose in their drinking water for 2 days. The concentration of fructose in the serum was measured. The serum fructose concentration of the control fed mice was 1.5±0.3 mM. In the fructose fed mice, the initial serum fructose concentration was similar to that of controls, at 1.8±0.4 mM. However, following 48 h of fructose feeding, the serum fructose concentration rose to 8.4±0.6 mM. Following 48 h, the islets were isolated, the cells dispersed and cultured for 16 h in fructose-free basal media. It is important to note that in this instance, the islets cells were not exposed to fructose while in culture. The cells were treated with TNFα and assessed for ethidium monoazide, annexin V and insulin or glucagon staining. As show in Fig. 2A, insulin positive islet cells showed no difference in viability between control-fed and fructose-fed mice over 24 h. However, upon treatment with TNFα, there was a marked loss of viability in the insulin positive islet cells isolated from fructose fed mice, with a 36% and 68% loss of viability at 12 and 24 h, respectively. By contrast the insulin positive islet cells isolated from control-fed mice were refractory to TNFα induced cytotoxicity, with only a marginal loss of cell viability at 24 h. As shown in Fig. 2B, pretreatment with necrostatin prevented the TNFα induced cytotoxicity in insulin positive islet cells isolated from fructose fed mice, with the inactive necrostatin analog having no effect. Importantly, the broad spectrum caspase inhibitor, ZVAD, did not display any ability to prevent the loss of cell viability induced by TNFα in cells isolated from fructose-fed mice. These results indicate that the TNFα induced cell killing of fructose exposed cells is necroptotic. In agreement with this assertion, treatment with TNFα did not bring about annexin V staining in insulin positive islet cells isolated from fructose-fed mice (Fig. 2C). By contrast, treatment with the apoptosis inducer, staurosporine, brought about robust annexin V staining in insulin positive islet cells isolated from both control and fructose fed mice (Fig. 2C). Moreover, the annexin V staining induced by staurosporine was prevented by the caspase inhibitor ZVAD, but not by necrostatin (Fig. 2C).
TNFα Induced Necroptosis of fructose exposed beta cells is dependent on MitoNEET, fructokinase and glut-5 expression
Fructokinase is expressed at low levels in pancreatic beta cells (Malaisse et al., 1989). However, both glut-5, which is selective for fructose transport, and fructokinase expression have been reported to be induced by fructose exposure (Douard and Ferraris, 2008; Koo et al., 2008). Exposure of INS-1E cells to three, 30 min spikes of 10 mM glucose alone did not bring about induction of glut-5, fructokinase or mitoNEET expression (Fig. 3A, lane #1). By contrast, as show in Fig. 3A lane #2, glut-5, fructokinase and mitoNEET expression were induced by exposure to three, 30 min spikes of 1 mM fructose and 10 mM glucose (lane #2). Twenty four hours after fructose exposure, glut-5 and fructokinase expression declined (Fig. 3A, lane #3). By contrast, mitoNEET expression remained elevated even after 24 h following fructose withdrawal (Fig. 3A, lane #3); a finding that may account in part, for the sustained sensitivity to TNFα induced necroptosis following fructose withdrawal. Significantly, siRNA mediated suppression of glut-5 expression prevented the fructose induced stimulation of both fructokinase and mitoNEET expression (Fig. 3A, lane #6). Also, prevention of fructokinase induction prevented the stimulation of mitoNEET expression, demonstrating that fructose metabolism is required to bring about induction of mitoNEET (Fig. 3A, lane #4). By contrast, suppression of mitoNEET expression had no effect on fructose induced stimulation of glut-5 or fructokinase expression, indicating that mitoNEET induction lies downstream of these two proteins (Fig. 3A, lane # 5). Also, pioglitazone did not prevent the induction of glut-5, fructokinase or mitoNEET expression by fructose, in keeping with its mechanism of binding to and disabling mitoNEET's ability to discharge its 2Fe-2S cluster (Fig. 3A, lane #8).
Importantly, islets isolated from fructose-fed mice also displayed elevations in the expression of mitoNEET, fructokinase and glut-5. Isolated pancreatic islets (200) were harvested, lysates prepared and utilized for western blotting. As shown in Fig. 3B, lane #1, pancreatic islets isolated from mice fed a control, non-fructose containing diet expressed low levels of glut-5, fructokinase or mitoNEET. However, feeding mice a diet with fructose constituting 20% of calories for two days resulted in a dramatic induction in the expression of glut-5, fructokinase and mitoNEET (Fig. 3B, lane #2). Importantly, the induction of mitoNEET expression was prevented by transfection of pancreatic islets with siRNA targeting mitoNEET (Fig. 3B, lane #3).
The induction of mitoNEET, fructokinase and glut-5 expression is required for fructose to sensitize beta cells to TNFα induced necroptosis. INS-1E cells were transfected with siRNAs targeting glut-5, fructokinase, mitoNEET or a non-targeting control. Following 24 h, the cells were exposed to the three fructose-glucose spikes over 24 h. The cells were then treated with TNFα in basal media for the times indicated. As shown in Fig. 4A, suppression of glut-5, fructokinase or mitoNEET expression prevented TNFα and fructose induced cytotoxicity in INS-1E cells. Similarly, pancreatic islets were isolated from fructose-fed mice, dispersed and transfected with siRNAs. Following 24 h, the cells were then treated with TNFα in basal media for the times indicated. As show in Fig. 4B, suppression of glut-5, fructokinase or mitoNEET expression prevented TNFα induced ethidium monoazide staining of insulin positive islet cells. Additionally, INS-1E exposed to fructose or islets isolated from fructose-fed mice were pretreated with 1 µM of pioglitazone for 30 min prior to treatment with TNFα. As shown in Fig. 4A and B, pre-treatment with 1 µM pioglitazone prevented TNFα induced cytotoxicity in both INS-1E cells exposed to fructose and insulin positive islet cells isolated from fructose-fed mice, suggesting that the ability of pioglitazone to bind to and inhibit mitoNEET may account for its pro-survival effects (Paddock et al., 2007; Wiley et al., 2007b; Zuris et al., 2011). Importantly, non-target control siRNA had no effect on preventing TNFα induced cytotoxicity.
TNFα induced necroptosis is mediated by binding of the Stat3-Grim-19 complex to MitoNEET resulting in the release of its 2Fe-2S cluster, accumulation of mitochondrial iron and ROS production
INS-1E cells or pancreatic islets isolated from fructose fed mice were transfected with siRNAs targeting Grim-19, Stat3 or a non-targeting control. For the INS-1E cells, following 24 h, the INS-1E cells were exposed to the three spikes of fructose-glucose. The media was then changed back to basal media and the cells were harvested. As shown in Fig. 5A, in INS-1 cells and islets, the siRNAs targeting Grim-19 or Stat3 selectively suppressed the expression of Grim-19 and Stat3, with the non-targeting control siRNA without effect on the expression of either protein. In Fig. 5B and C, the cells were treated with TNFα for the time periods indicated. Suppression of Stat3 or Grim-19 expression prevented TNFα induced cytotoxicity in fructose exposed INS-1E cells (Fig. 5B). Similarly, as shown in Fig. 5C, in insulin positive islet cells isolated from fructose fed mice, suppression of Grim-19 or Stat-3 expression also prevented TNFα induced cytotoxicity.
We have demonstrated that during TNFα induced necroptosis of hepatocytes, Grim-19 and Stat-3 interact with one another and translocate to the mitochondria (Shulga and Pastorino, 2014). Therefore, INS-1E cells were exposed to fructose-glucose spikes and then treated with TNFα. Following two hours, a time point prior to appreciable cell death, mitochondria were isolated and mitoNEET immunoprecipitated. As show in Fig. 6A, treatment of fructose exposed INS-1E cells with TNFα brought about an interaction between mitoNEET and the Stat-3-Grim-19 complex that was prevented by necrostatin (lane 3# vs lane #4). Similarly, immunoprecipitation of Grim-19 detected an interaction between it, Stat3 and mitoNEET in fructose exposed cells treated with TNFα, which was also prevented by necrostatin (Fig. 6B, bottom panels).
We have demonstrated that the binding of the Stat3-Grim-19 complex to mitoNEET initiates the release of mitoNEET's bound 2Fe-2S cluster. MitoNEET's 2Fe-2S cluster can be monitored by its absorption peak at 458 nm when the 2Fe-2S cluster is bound (Bak et al., 2009; Zuris et al., 2011, 2012). MitoNEET was immunoprecipitated from isolated mitochondria and the absorbance at 458 nm measured. As show in Fig. 7A, treatment of fructose exposed INS-1E cells with TNFα induced a progressive decrease in absorbance at 458 nm, indicating a release of the 2Fe-2S cluster from mitoNEET that was not prevented by a non-targeting siRNA (green bars). The release of the 2Fe-2S cluster from mitoNEET is accompanied by an accumulation of mitochondrial iron; with the content of mitochondrial iron tripling within 4 h after treatment with TNFα in fructose exposed INS-1E cells (Fig. 7B, green bars). Moreover, suppression of Grim-19 expression or pre-treatment with pioglitazone prevented the discharge of the 2Fe-2S cluster from mitoNEET and accumulation of mitochondrial iron induced by TNFα in fructose exposed INS-1E cells (Fig. 7A and B). Suppression of mitoNEET expression also prevented the TNFα and fructose induced overload of mitochondrial iron (Fig. 7B).
The mitochondrial accumulation of iron resulted in a surge of ROS production. INS-1E cells exposed to fructose were treated with TNFα and mitochondrial ROS production was measured utilizing MitoSOX, which is somewhat selective at detecting ROS generated by mitochondria. As show in Fig. 8A, TNFα induced a robust stimulation of ROS production in fructose exposed INS-1E cells that was not prevented by a non-targeting siRNA (green bars). However, suppression of mitoNEET or Grim-19 expression prevented the TNFα induced stimulation of ROS production in fructose exposed cells. Also, pre-treatment of the fructose exposed INS-1E cells with pioglitazone prevented TNFα from stimulating ROS production. Moreover, as shown in Fig. 8B, TNFα induced a similar surge of MitoSox fluorescence in the pancreatic beta cells of islets isolated from fructose fed mice that was prevented by pretreatment with pioglitazone, or suppression of Grim-19 or mitoNEET expression.
The effects of fructose on beta cell viability are not characterized. While the role of mitochondrial failure in beta cell injury and death is established, the signaling pathways and mechanisms that initiate mitochondrial injury are ill defined. The data define mitoNEET, a target of the anti-diabetic class of thiazolidinediones, as a distinct point of initiation and integration of mitochondrial dysfunction during TNFα induced cytotoxicity in fructose exposed pancreatic beta cells, relaying signals to the interior of the mitochondria, where ROS formation promotes opening of the mitochondrial permeability transition pore (PTP), resulting in mitochondrial depolarization and necrosis.
The reported serum concentration of fructose varies widely from 5 µM to 1.9 mM (Abdelmalek et al., 2012; Brymora et al., 2012; Kawasaki et al., 2012). It has been assumed that most fructose is metabolized in the liver and pancreatic beta cells possess little capacity to metabolize fructose. A study utilizing a fructose dehydrogenase assay in conjunction with GC-MS, found the fasting fructose serum concentration at 1.9 mM with a spike of 17.2 mM after ingestion of a fructose and glucose containing drink (Hui et al., 2009). Most studies do show a spike in serum fructose following its injection or ingestion. Importantly, our studies showed that mice fed 20% fructose in drinking water displayed a remarkable spike in the serum fructose concentration in just a 48 h time frame, rising over four fold, from 1.8 mM to 8.4 mM.
Significantly, in our studies, for the 1 mM fructose spike to induce expression of glut5, fructokinase or mitoNEET in beta cells; it had to be accompanied by a spike in glucose (from 5 mM to 10 mM). This is analogous with findings that fructose absorption is increased in the presence of glucose in the intestine, a so called glucose dependent co-transport of fructose (Corpe et al., 1996; Le Gall et al., 2007; Wright et al., 2003). The main glucose transporter for the apical membrane of the enterocytes of the small intestine is glut-2, the same as in pancreatic beta cells. Therefore it is feasible that glucose assists in the initial uptake of fructose into beta cells by glut-2. Once fructose is taken up, it subsequently induces the expression of glut-5, thereby enabling a greater uptake of fructose and induction of fructokinase, with the resultant metabolic effects and subsequent induction of mitoNEET expression. Importantly, the induction of mitoNEET and fructokinase expression, in addition to the enhanced sensitivity incurred on pancreatic beta cells to TNFα induced cytotoxicity, was replicated in mice fed a fructose containing diet, indicating that in the intact pancreas, beta cells are exposed to fructose when it is a component of the diet.
We have found that fructose exposure drastically lowers cellular ATP content by 50%. Unlike hexokinase or glucokinase; fructokinase does not experience feedback inhibition by fructose-1-phosphate, so unregulated fructokinase activity depletes ATP (Abdelmalek et al., 2012; Malaisse et al., 1989). Since mitoNEET is a 2Fe-2S transfer protein localized to the mitochondrial outer membrane, it is tempting to speculate that the up-regulation of mitoNEET expression is in response to fructose induced ATP depletion, resulting in an increased need for mitochondrial respiration that requires 2Fe-2S clusters in components of the mitochondrial respiratory chain. Intriguingly, a correlation has been identified between iron overload and the onset of diabetes, with reduction of iron though phlebotomy improving the deleterious effects of diabetes (Simcox and McClain, 2013; Tajima et al., 2012).
Exposure of pancreatic beta cells to either fructose or glucose has been demonstrated to promote onset of the mitochondrial permeability transition (Lablanche et al. 2011). In that work, fructose was found to be a more potent sensitizer to permeability transition pore opening than glucose. In this context it is important to note that opening of the permeability transition pore occurs during TNFα induced cytotoxicity (Pastorino et al., 1996).
In addition to the direct effects of fructose on the sensitivity of beta cells to TNFα induced cytotoxicity, fructose can affect the activation and function of innate inflammatory cells (Leibowitz et al., 2013). Fructose has been reported to induce endotoxemia in primates by providing an increase in the triggers necessary for promoting inflammatory cell activation (Kavanagh et al., 2013). Compounding this is the ability of fructose to increase the sensitivity of immune cells to activation (Tagzirt et al., 2014). These observations are consistent with studies demonstrating that the pancreatic islets in rat models of type II diabetes are infiltrated with activated macrophages (Cucak et al., 2014; Dahlen et al., 1998; Donath et al., 2013; Jourdan et al., 2013). All of these factors conspire to enhance the exposure of beta cells in pancreatic islets to damaging cytokines such as TNFα.
In summary, we find that fructose exposure induces the expression of fructokinase and mitoNEET in beta cells. Induction of mitoNEET expression was necessary for the sensitizing effect, as its suppression prevented the sensitization to TNFα induced cytotoxicity brought about by fructose exposure in the INS-1E and beta cells of pancreatic islets. In both the beta cells of pancreatic islets and INS-1E cells, the cytotoxicity induced by TNFα in the presence of fructose was a necroptotic form of cell killing. The data indicate that the induction of necroptosis in beta cells brought about by fructose exposure is partly caused by the TNFα induced binding of a Stat3-grim-19 complex to mitoNEET. Upon TNFα treatment, the binding of the Stat3-Grim-19 complex causes a discharge of mitoNEET's 2Fe-2S cluster, resulting in an accumulation of mitochondrial iron and ROS production. Suppression of either Stat3, Grim-19 or mitoNEET expression prevented the TNFα induced ROS production and necroptosis in beta cells exposed to fructose. Intriguingly, pre-treatment with the anti-diabetic drug, pioglitazone, which binds to and inhibits the ability of mitoNEET to release its 2Fe-2S cluster, prevented the TNFα induced accumulation of mitochondrial iron, ROS formation and cytotoxicity in fructose exposed pancreatic beta cells.
MATERIALS AND METHODS
Culture, isolation of pancreatic islets, and exposure of pancreatic beta cells to fructose spikes and TNFα
The pancreatic beta cells line, INS-1E, was cultured in basal RPMI media containing 5 mM of glucose at 50,000 cells per well in 24 well plates for cell viability assays and at 250,000 cells per well in 6 well plates for harvesting and isolation of mitochondria. Where indicated, the cells were exposed to a 30 min spike of fructose (1 mM) and glucose (10 mM). In total, the cells were exposed to three, 30 min spikes of 1 mM fructose and 10 mM glucose over 24 h, with each spike separated by 8 h. Importantly, between the fructose-glucose spikes, the cells were washed with PBS and placed back into basal RPMI media prior to the next fructose-glucose spike. The cells were then treated with TNFα at 10 ng/ml for the time periods indicated in the figures.
For isolation of pancreatic islets, eight to twelve week-old male C57/BL6J mice are anesthetized with isoflurane. An abdominal incision was made to expose the liver and intestines. The ampulla is clamped with surgical clamps on the duodenum wall to block the bile duct pathway to the duodenum. Under a stereo-microscope, a 5 ml syringe with a 30G1/2-G needle is inserted into the common bile duct though the joint side of the hepatic and cystic ducts, reaching the middle of the common bile duct. The pancreas is distended by injecting a solution of collagenase XI (1000 U/ml) in Hanks Balanced salt solution (HBSS). The pancreas is removed and placed in 50 ml of HBSS containing collagenase XI. The tube is incubated for 15 min at 37.5°C for digestion of the pancreas. The tube is shaken two or three times during the incubation. Following incubation, the pancreas is disrupted by shaking until a homogenous suspension is formed. The pancreatic digestion is stopped by placing the tube on ice and adding 25 ml of 1 mM CaCl2 in HBSS. The tube is centrifuged at 300 g for 30 s at 4°C. The supernatant is decanted and the pellet resuspended in 25 ml of chilled HBSS containing 1 mM CaCl2, wherein it is centrifuged for a second time at 300 g for 30 s. The supernatant is decanted and the pellet resuspended in 15 ml of HBSS containing 1 mM of CaCl2. The resuspended pellet is poured over a 70 µm cell strainer. The contents of the strainer are deposited in a petri dish containing RPMI 1640 media. Under a stereo-microscope, islets are picked utilizing a wide mouthed pipette tip. The islets are placed into RPMI-1640 media and then into a 5% CO2 incubator. The identity of islets cells was determined by immuno-staining for insulin (beta cells) or glucagon (alpha cells). Pancreatic islets were isolated from mice fed a non-fructose containing diet and from mice fed a fructose containing diet for two days. The mice were randomly divided into two groups, the control group and fructose-drinking water group. Both groups were fed a fructose-free rat chow diet. However the control group was given water while the fructose group was administered 20% fructose in the drinking water for two days.
Determination of fructose serum concentrations
The serum concentration of fructose was measured utilizing a spectrophotometric technique as described by Hui et al. (2009). A solution of fructose dehydrogenase from Gluconobacter sp (125 U/ml) was prepared in 1% Triton X-100, 1 mM 2-mercaptoethanol in citric phosphate buffer (0.05 M citric acid, 0.09 M dibasic sodium phosphate, pH 4.5). MTT [3-(4,5-dimethylthiaze-syl)-2,5-diphenyltetrazolium bromide] was prepared in citric phosphate buffer at a concentration of 0.6 mg/ml, with phenazine methosulfate (PMS) prepared at 2.4 mg/ml. Assays were performed in flat-bottom, 96-well microtiter plates; 20 µl of standard or serum was added to a well, and the assay initiated by the addition of 200 µl of mixing reagent. The plates were then incubated for 30 min at 37°C, followed by absorbance reading at 570 nm.
Determination of cell viability and ROS production
For INS-1E cells, viability was determined by the ability of the cells to exclude propidium iodide, an indicator of plasma membrane integrity. Following treatments, cells were harvested and centrifuged at 700 ×g. The cell pellet was re-suspended in phosphate buffered saline to which was added 5 µM of propidium iodide. After 5 min incubation, the cells were pelleted and re-suspended in PBS. The percentage of viable cells was determined utilizing a Cellometer (Nexcelom Bioscience LLC, Lawrence, MA, USA) as the ratio of the number of cells in the fluorescent images (propidium iodide positive) to the bright field images. To differentiate programmed, primary necrosis from secondary necrosis that is seen in late stages of apoptosis, annexin V staining was utilized concurrently. Cells undergoing programmed necrosis do not stain positive for annexin V at early time points and prior to loss of plasma membrane integrity, whereas apoptotic cells undergoing secondary necrosis will stain positive for annexin V.
Caspase activity was determined using NucView 488 Caspase-3 activity kit (Biotium Hayward, CA, USA). Floating and attached cells were collected and resuspended in RPMI media containing 5 µM of the NucView 488 substrate and then incubated at room temperature for 30 min protected from light. After incubation, the cells were washed once with ice cold PBS, and then re-suspended in PBS. Caspase activity was detected by an increase in the intensity of the DNA binding dye using Cellometer Vision (Nexcelom Bioscience). For determination of phosphatidylserine (PS) externalization, floating and attached cells were collected and re-suspended in 100 µl of binding buffer at 1.0×105 cells/ml. FITC-Annexin-V (5 µl) was added, and the cells were incubated for 15 min at room temperature. PS positive cells were determined by flow cytometry.
For isolated pancreatic islets, ethidium monoazide in conjunction with annexin V staining was utilized to determine the viability and mode of death for cells staining positive for insulin or glucagon. Ethidium monoazide is a fixable fluorescent photoaffinity label, which after photolysis, binds covalently to DNA and, like propidium iodide, only enters cells with compromised plasma membranes (Avlasevich et al., 2006).
As a positive control for apoptosis, beta cells were treated with staurosporine, a compound that brings about apoptosis in beta cells (Collier et al., 2011). Staurosporine induced caspase-3 activation along with annexin V staining before breakdown of plasma membrane integrity, as assessed by propidium iodide or ethidium monoazide.
For measurement of ROS production, 5 µM of MitoSOX was added to cells 10 min before harvesting. The cells were pelleted, re-suspended in PBS. In cells with active production of ROS, MitoSOX is oxidized to a fluorescent species. The percentage MitoSOX positive cells was determined utilizing a Cellometer which calculated the ratio of the number of MitoSOX positive cells in the fluorescence images to the number of cells in the bright field images.
Isolation of mitochondrial and cytosolic fraction
Following treatments, approximately 500,000 cells were harvested by trypsinization and centrifuged at 700×g for 10 min at 4°C. The cell pellets were washed once in PBS and then resuspended in 3 volumes of isolation buffer (20 mM Hepes, pH 7.4, 10 mM KCl, 1.5 mM MgCl2, 1 mM sodium EDTA, 1 mM dithiothreitol, and 10 mM phenylmethylsulfonyl fluoride, 10 µM leupeptin, 10 µM aprotinin) in 250 mM sucrose. After chilling on ice for 3 min, the cells were disrupted by 40 strokes of a glass homogenizer. The homogenate was centrifuged twice at 1500×g at 4°C to remove unbroken cells and nuclei. The mitochondria-enriched fraction (heavy membrane fraction) was then pelleted by centrifugation at 12,000×g for 30 min. The supernatant was removed and then filtered through a 0.2 µm and then a 0.1 µm Ultrafree-MC filter to obtain the cytosolic fraction.
Measurement of chelatable iron
Reagents were treated with Chelex 100 Resin (0.3 g/10 ml). The following reagents were incubated for 1 h at 37°C in a 1 ml final reaction volume: 100 mM Tris-HCl pH 7.4, 0.5 mg calf thymus DNA, 0.075 U bleomycin, 5 mM MgCl2, 1 mM ascorbic acid, 50 µl of cell lysates or FeCl2 standard. The reaction was stopped by the addition of 60 mM butylated hydroxytoluene. An aliquot of the mixture (0.4 ml) was incubated with 0.2 M phosphoric acid and 0.11 M thiobarbituric acid (TBA) for 45 min at 90°C. The upper organic layer was extracted with 1 ml N-butanol and absorbance was measured at 532 nm. A FeCl2 concentration curve was constructed to determine the iron concentration in cellular lysates.
Transfection with siRNA
Cells or islets were transfected with the indicated siRNAs targeting mitoNEET, fructokinase, glut-5, Stat-3, Grim-19 or a non-targeting control utilizing a lipid-based method supplied from a commercial vendor (Gene Therapy Systems) at a final siRNA concentration of 50 nM. After formation of the siRNA-liposome complexes, the mixture was added to the cells for 8 h. Afterwards, the medium was aspirated, and complete medium was added back for a further 16 h, after which time the cells were utilized for experiments.
Immunoprecipitation of Grim-19 and Stat-3
Grim-19 or Stat-3 were immunoprecipitated from mitochondrial extracts (Grim-19 antibody from Novus Biologicals and Stat-3 antibody from Cell Signaling). The immunoprecipitates were then run out on SDS-PAGE gels and -blotted onto PVDF membranes. The western blots were developed with antibodies against Stat-3 or Grim-19.
Results are expressed as means±s.d. of at least three independent experiments. Statistical significance was defined at P<0.05.
The authors declare no competing or financial interests.
N.S. conducted experiments. J.G.P. conceived and performed experiments. J.G.P. prepared the manuscript.
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
- Received September 3, 2015.
- Accepted September 14, 2015.
- © 2015. Published by The Company of Biologists Ltd
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution and reproduction in any medium provided that the original work is properly attributed.