Involvement of SIRT1 in Zn2+, Streptozotocin, Non-Obese Diabetic, and Cytokine-Mediated Toxicities of β-cells

Abbreviations: BESTO: β-cell SIRT1 Overexpressing mice; CaEDTA: Ethylene Diamine Tetra-Acetic acid-Calcium salt; GAPDH: Glyceraldehyde-3-Phosphate Dehydrogenase; HD: High-Density cultures; IFN-γ: Interferon Gamma; IL-1β: Interleukin-1 Beta; MEM: Minimal Essential Medium; MLDS: Multiple Low-Dose Streptozotocin; mM: Millimol/L; MTT: 3-(4,5-dimethylthiazol-2-Yl)-2,5-diphenyltetrazolium Bromide; Nampt: Nicotinamide Phosphoribosyl Transferase; Naph: 2-hydroxynaphthaldehyde; N: Nicotinamide; NAD+: Nicotinamide Adenine Dinucleotide; NOD: Non-Obese Diabetic; P: Pyruvate; PARP: Poly-ADP Ribose Polymerase; PDH: Pyruvate Dehydrogenase; ROS: Reactive Oxygen Species; S: Sirtinol; siRNA: Small inhibitory RNA; SIRT1: Sirtuin protein 1; STZ: Streptozotocin; TNF-α: Tumor Necrosis Factor-alpha; T1DM: Ttype-1 Diabetes; T2DM: Type2 Diabetes; uM: Micromol/L; Zn2+: Zinc; [Zn2+]: Intracellular Zinc concentration; ZnT5: Zinc Transporter 5


Introduction
Type-1 diabetes (T1DM) is an autoimmune disease resulting from specific T-lymphocyte-, ROS-, and cytokine-mediated destruction of the insulin-producing β-cells of the islets of Langerhans resulting in dysregulation of blood glucose [1]. The development of T1DM is reduced by treatment with T-cell and cytokine inhibitors, and ROS scavengers [2]. These oxidative processes are suggested to alter the NAD+/ NADH ratio and inhibit proteins involved in energy metabolism and glycolysis causing the accumulation of triosephosphates [3,4]. Inhibitors of NAD+ catabolism have been demonstrated to attenuate diabetic incidence in models of T1DM [5,6]. NAD+ loss is linked to diabetes. Heterozygous knockout of the rate limiting enzyme in NAD+ synthesis (Nampt), causes reduced insulin secretion [7]. Also, Nampt and NAD + levels are reduced in T2DM and the aging or high-fat diet models thereof. Restoration of NAD + , by nicotinamide mononucleotide precursor supplementation, attenuates diabetes in aging and high-fat diet mouse models of T2DM [8].
We recently showed that just prior to becoming diabetic, NOD mice demonstrate increased punctate Zn 2+ staining in islets which is attenuated by a reduced Zn 2+ diet, or zinc transporter 5 (ZnT5) knockout. Triweekly pyruvate or nicotinamide injections, chronic treatment with a reduced zinc diet, or knockout of the zinc transporter 5 (ZnT5) gene delay onset of diabetic incidence and animal mortality by reducing pancreatic zinc and/or maintaining β-cell NAD+ levels and mass [9]. This complements the beneficial effects demonstrated for Zn2+ chelation against acute or multiple low dose streptozotocin exposures [10,11]. Zinc neurotoxicity in vitro or in vivo induces NAD+ loss and glycolytic inhibition resulting in increased triosephosphates and death in a manner exactly equivalent to that seen in β-cells. These results are prevented by genetic or dietary reduction in brain Zn2+, or exogenous pyruvate, nicotinamide, NAD+, or sirtuin inhibition [12][13][14].

Zn 2+ and β-cell death
Zn 2+ is present in the pancreas at the highest concentration anywhere in the body, and within the pancreas is concentrated in the secretory granules of β-cells [10]. In the β-cell, Zn 2+ allows insulin processing and crystallization in the secretory granules [15,16]. Significant amounts of free Zn 2+ are also released from β-cell secretory granules [17,18]. As β-cells contain most of the pancreatic Zn 2+ , its toxic release, by the immune response or exposure to ROS/streptozotocin, would help explain the specificity of β-cell death in type-1 diabetes [19,20]. Zn 2+ released intra-or extra-cellularly by diabetic conditions is postulated to induce detrimental intracellular effects on NAD + levels through activation of the sirtuin pathway. The NAD + loss results in glycolytic inhibition which is prevented by restoration of NAD + levels using nicotinamide, sirtuin inhibition, and pyruvate, but not lactate ( Figure 1). In these studies, we examined the deleterious effects that Zn 2+ -mediated NAD + loss, had in facilitating β-cell death. We propose that immune-, and ROS-mediated dysfunction of energy metabolic pathways could be potentiated by Zn 2+ release, and sirtuin-mediated loss of NAD + . We studied the effects of SIRT1 overexpression or knockdown in MIN6 cultures on NAD + loss and Zn 2+ , STZ, or cytokine toxicities, and the effect of SIRT1 expression on the streptozotocin or NOD in vivo models of T1DM.

Determination of levels of NAD + , and NADH
Measurements of NAD + and NADH were made on cell lysates prepared immediately after 4 hr Zn 2+ , STZ or cytokine exposures. Cells were washed three times to remove compounds followed by lysis in NaOH/EDTA. This lysate was split and part of it directly hydrolyzed at 80°C for 20 min, and the other part acidified followed by hydrolysis at 80°C for 20 min. Alkaline hydrolysis destroys NADH, whereas acid hydrolysis destroys NAD + allowing for their determinations by linked enzymatic cycling reactions. 2 µl of acid extract (~5000 cells) were added directly to 100 µl of NAD + cycling reagent (100 millimol/L Tris-HCl, pH 8.1, 2 millimol/L β-mercaptoethanol, 2 millimol/L oxaloacetate, 0.3 mol/L ethanol, 0.02% BSA, and yeast alcohol dehydrogenase and 0.5 µg/mL malic dehydrogenase) and incubated at 25°C to obtain 500 cycles of NAD + amplification. Termination by heating at 100°C for 5 min was followed by addition of 1 mL of malate indicator reagent (50 millimol/L amino-methylpropanol (pH 9.9), 5 millimol/L L-glutamate, 0.2 millimol/L NAD + , 5 µg/mL malic dehydrogenase, and 2 µg/ mL glutamate oxaloacetate transaminase). This reaction was incubated for 10 min at 25°C. The NADH generated from malate was measured fluorimetrically (excitation at 365 nm, emission monitored at 460 nm) [13,24]. For NAD + additions, a correction was made based on NAD + addition and washout from control cultures.

SIRT1 gene expression
Total RNA from harvested MIN6 cells (10 7 ) exposed to 0 or 40 micromol/L Zn 2+ in serum-free MEM for 3 h was extracted as detailed [25]. The RNA concentration and integrity were verified, and reverse transcription was performed using the iScript cDNA synthesis kit (Bio-Rad, Hercules, CA, USA) on 1 µg of total cellular RNA. Real-time RT-PCR specific primers for SIRT1 and β-actin are as described previously [26]. The reactions were run in duplicate with iQSYBR green Supermix (BioRad) as per manufacturer's instructions. A melt-curve analysis was

Figure 1: Model of zinc toxicity and prevention in Type-1 Diabetes
We hypothesize that ZnT5 and ZnT8 mediate Zn 2+ accumulation in secretory granules. The immune cell response injures β-cells causing Zn 2+ release from granules, and re-uptake, through Ca2+ channels, in neighboring β-cells. Additionally, the ROS may oxidize metallothionein causing Zn 2+ release. The resulting increased [Zn 2+ ]i may cause direct inhibition of mitochondria, and GAPDH, or their indirect inhibition by a reduction in NAD + levels induced by the NAD + catabolizing enzyme SIRT1. Pyruvate, nicotinamide, and sirtuin inhibition prevent NAD + loss and glycolytic inhibition. Black = Toxic, Gray = Therapeutic. performed at the end of each experiment to verify that a single product per primer pair was amplified, and the sizes of the amplified DNA fragments were verified. Samples were compared using the relative CT (the cycle number at the threshold level of log-based fluorescence) method. The percent increase was determined relative to an untreated control culture after normalizing to β-actin expression using 2-ΔΔCT (Livak) method.

Colony maintenance and trials
The NOD inbred mouse strain (Taconic), the β-cell SIRT1 overexpressing mice (BESTO), and the SIRT1 +/-mice were maintained at LSUHSC's transgenic animal facility. The SIRT1 +/-and BESTO mice were backcrossed onto a C57/Bl6/J background for maintenance. The BESTO mice, line 431-2, showed a 12-fold overexpression of SIRT1 predominantly in β-cells which resulted in transcriptional regulation of target genes [21]. Housing and anesthesia concurred with the institutional Animal Studies Committee guidelines, the PHS Guide for the Care and Use of Laboratory Animals, USDA Regulations, and the AVMA Panel on Euthanasia. SIRT1 heterozygous knockout animals [27] were backcrossed to NOD mice for 10 generations to syngeneity; SIRT1 +/+/NOD animals develop diabetes and mortality at equivalent ages to the parental NOD animals. Upon interbreeding heterozygous animals, SIRT1 -/-animals die peri-natally whether on an SV129 or an NOD background, but survive on a CD1 background [27,28]. This was a double blind trial (both handler and histologist were blind to genotype) of age-matched female NOD mice with either a SIRT1 +/-or a SIRT1 +/+ genotype. Water and food ingestion, and body weight were monitored weekly and did not vary between groups. Fed blood glucose was monitored every Monday afternoon, and fasted blood glucose was also determined periodically (glucose oxidase). Fed and fasted blood glucose gave qualitatively similar results. Animals demonstrating continued akinesia with prodding or inability to eat and drink were euthanized and mortality recorded. MLDS was performed at 4-7 months of age (0.055 g/kg i.p. injection each day for 5 days), and blood glucose was measured on days 0, 1, 4, 7, 14, and 21 (n = 20).

Reagents
Unless otherwise stated, all reagents were from Sigma Chemical Co (St. Louis, MO).

Results
In MIN6 cultures, knockdown of SIRT1 reduced Zn 2+ toxicities, and overexpression of SIRT1 potentiated Zn 2+ toxicities As shown in Figure 2, MIN6 cultures over expressing SIRT1 or an overexpression empty vector control were exposed to A) Zn 2+ in the absence of serum or B) Zn 2+ in the presence of serum. SIRT1 overexpression (760% of control) significantly potentiated Zn2+ toxicity at all levels of exposure in the presence or absence of serum in β-cells, and sirtuin inhibitors (sirtinol or Naph) attenuated this death. As shown in Figure 3, MIN6 cultures over expressing an siRNA to SIRT1 or an siRNA empty vector control were exposed to A) Zn 2+ in the absence of serum or B) Zn 2+ in the presence of serum. SIRT1 knockdown (28% of control) significantly attenuated Zn2+ toxicity at all levels of exposure in the presence or absence of serum in β-cells and sirtuin inhibitors did not further reduce this toxicity. Real-time PCR for β-cell SIRT1 expression was performed on control MIN6 cultures 3 h after exposure to 40 micromol/L Zn 2+ , resulting in a 39 ± 4.5% significant increase in expression relative to untreated control at P < 0.05 by student t-test (n = 6).
Knockdown of SIRT1 reduced STZ and cytokine toxicities, and overexpression of SIRT1 potentiated them MIN6 overexpressing SIRT1, or its siRNA, and control cultures were exposed to 7.5 millimol/L STZ (Figure 4), or mixed cytokines ( Figure 5) for 6 hours (high density condition-7.5x1010 cells/L, 0.04 milliL in each 96-well V-bottom pit) in the presence or absence of nicotinamide, NAD + , or Naph. These compounds were chosen because they have been previously demonstrated to restore NAD + levels, and to attenuate Zn 2+ neurotoxicity. Cell viability was determined by MTT stain-

Figure 2: SIRT1 overexpression in MIN6 cells potentiated Zn 2+ toxicity
Stably transfected MIN6 cell lines were established which overexpress SIRT1 protein (760% of control). These cell lines were tested for their sensitivity to Zn 2+ toxicity in the presence or absence of serum. A. SIRT1 overexpressing (SIRT1) and empty vector control (Control) MIN6 cell lines were assayed for their sensitivity to Zn 2+ in the absence of serum (MS). Cells were grown to 50% confluence exposed as indicated, and cell death determined by propidium iodide staining ( ing. Pyruvate, nicotinamide, NAD + , sirtinol, and Naph were applied at optimized concentrations as determined by dose responses (data not shown). These compounds each attenuated STZ and mixed cytokine toxicities, with nicotinamide and NAD + having the best efficacy. In SIRT1 overexpressing MIN6 cells, 2x naphthaldehyde (60 micromol/L) was required for efficacy. Zn 2+ , STZ, and cytokines reduced NAD + levels which could be restored by SIRT1 knockdown, nicotinamide, and NAD + MIN6 cultures were exposed to 10 millimol/L streptozotocin, 40 micromol/L Zn 2+ , or mixed cytokines for 4 hrs (prior to cell death); NAD + was isolated and measured. Zn 2+ , streptozotocin, and cytokines induced a significant decrease in NAD + levels and nicotinamide or NAD + restored these levels. In addition, SIRT1 knockdown attenuated the loss of NAD + for Zn 2+ and streptozotocin ( Figure 6). We previously showed that sirtuin inhibition attenuated the loss of NAD + for Zn 2+ in unmodified MIN6 cells [9].

Mice overexpressing SIRT1 in β-cells had increased susceptibility to STZ compared to SIRT1 +/-, and SIRT1 +/-bred onto an NOD background had reduced mortality
Two more BESTO mice were susceptible to multiple low dose streptozotocin exposure than their wildtype littermates (Table 1). SIRT1 knockout mice survive on a CD1 background, but are smaller and sick- Stably transfected MIN6 cell lines were established which overexpress a small inhibitory RNA to SIRT1 resulting in knockdown of its expression (28% of control). These cell lines were tested for their sensitivity to Zn 2+ toxicity in the presence or absence of serum. A. Knockdown of SIRT1 (siRNA) and empty vector siRNA control (Control) MIN6 cell lines were assayed as in Figure 1. Cells were grown to 50% confluence exposed as indicated, and cell death determined by propidium iodide staining (5 microg/mL) 20-22 hrs later compared to the complete death induced by 3 millimol/L Zn 2+ (mean ± SEM, n=12-20 from three independent experiments). B. This cell line was also assayed in the presence of serum (requiring higher Zn 2+ concentrations) as in A. * signifies difference from Zn 2+ exposure alone at P < 0.05. $ signifies difference from similarly exposed Control cultures at P < 0.05. S means sirtinol, uM means micromol/L, Zn means zinc, naph means 2-hydroxynaphthaldehyde.   Figure 5: Cytokine mediated β-cell death was attenuated by SIRT1 knockdown and was potentiated by SIRT1 overexpression MIN6 cells were exposed as indicated to cytokines for 6 hrs at high density. Cell survival was assayed by 0.1% MTT staining. In SIRT1 overexpressing MIN6 cells, 2x naphthaldehyde (60 micromol/L) was required for efficacy. # signifies difference from cytokine exposure alone; $ signifies difference from similarly exposed control cultures at P < 0.05. ly probably due to developmental problems, and did not survive MLDS exposure [28]. SIRT1 knockout mice on SV129 or NOD backgrounds are embryonic lethal. However, SIRT1 +/-heterozygous mice trended towards reduced susceptibility to MLDS, and when put on an NOD background, trended toward reduced diabetic incidence and mortality with a single mortality time point showing a significant difference. When SIRT1 +/-mice were compared to BESTO mice for susceptibility to MLDS, there was a significant difference ( Figure 7, Table 1).

Discussion
In these studies we demonstrated that: 1) sirtuin pathway inhibition attenuated Zn 2+ -, STZ-, and cytokine-mediated toxicity and NAD + loss in β-cells, 2) SIRT1 overexpression in MIN6 cell lines potentiated NAD + loss, and Zn2 + , STZ, and cytokine toxicities, 3) SIRT1 knockdown using an siRNA expressing MIN6 cell line attenuated NAD + loss and these toxicities, 4) diabetic incidence and mortality induced in vivo by streptozotocin or NOD, showed covariance with SIRT1 expression levels upon genetic manipulation.
Zn 2+ is toxic to insulinoma cells and to isolated islets, and that zinc chelation and pyruvate can prevent toxicity. Zn 2+ chelation and pyruvate attenuate hyperglycemia in the acute high dose streptozotocin model though the NAD + -dependent mechanism of action was questioned [10,20]. We have utilized several in vitro and in vivo models of T1DM. These include in vitro exposure of β-cells to zinc, mixed cytokines, or STZ; and in vivo exposure to STZ, or the immune-mediated NOD mouse model. STZ and the NOD mouse cause selective death of insulin-secreting β-cells, inducing reductions in nicotinamide cofactor levels, glucose oxidation, and glucose-induced insulin secretion [29][30][31][32]. The NAD + precursor, nicotinamide, reduces diabetic incidence in both the acute high dose, and the multiple low dose streptozotocin injection paradigms [33,34]. Zn 2+ preferring chelators (CaEDTA and clioquinol) reduce diabetic incidence in the acute high dose and multiple low dose streptozotocin injection models as demonstrated by the reduction in Zn 2+ staining, β-cell death and diabetic symptoms achieved [10,11,20]. We recently demonstrated that pyruvate attenuates multiple low-dose streptozotocin exposure (MLDS) induced-, and NOD-induced diabetes, and that a zinc reduced diet also attenuates NOD-induced diabetes [9].
Our studies in neurons show that an increase in intracellular Zn2+ causes a loss of NAD + levels that may be partially mediated by sirtuin or poly-ADP ribosyl polymerase (PARP) activation depending on celltype [13,35]. The resultant decrease in the NAD + /NADH ratio inhibits the energy metabolic pathway at the susceptible enzymes GAPDH and PDH. Pyruvate, nicotinamide, or exogenous NAD + restore NAD +  Figure 7: SIRT1 +/-mice on an NOD background trend toward reduced susceptibility to diabetic incidence and mortality compared to SIRT1 +/+/ NOD littermates SIRT1 +/-mice were backcrossed into NOD animals for 10 generations, and then intercrossed to generate SIRT1 +/+/NOD, SIRT1 +/-/NOD, but no SIRT1 -/-/NOD littermates. Blood glucose was monitored weekly with a One-Touch Ultra glucose monitor, and animals that demonstrated akinesia with continued prodding by an observer blinded to treatment condition were sacrificed, and mortality recorded. A) The number of animals whose blood glucose (BG) remained below 2.5 g/L for 2 consecutive weeks was plotted as a function of weeks of age. B) The number of surviving animals was plotted as a function of weeks of age. * indicates a significant difference in the age at which 50% of the mice have died compared to SIRT1 +/+/NOD littermate controls at P < 0.05 by a t-test.  levels, and glycolytic flux, and thereby attenuate death. Pyruvate is converted to lactate regenerating NAD + at the expense of NADH [12]. Nicotinamide induces increased synthesis of NAD + , or decreases its degradation by NAD + -catabolizing enzymes [35]. Nicotinamide is effective in diabetic models [5,36], with therapeutic effects observed only if it is given to prediabetic patients [37][38][39][40]. Its mechanism of action is not well defined, with suggestions that it prevents PARP-activation and NAD + depletion, thereby reducing apoptosis of β-cells induced by DNA damage [5]. PARP induced NAD + depletion is also a mechanism for GAPDH inhibition, resulting in triosephosphate accumulation, and β-cell death in diabetes [41]. However, no reduction in diabetes occurs in the NOD/PARP -/-mouse arguing against a causative role for PARP in this animal model of diabetes [42]. The ability of exogenous NAD + to attenuate Zn 2+ , STZ, and cytokine toxicities in insulinoma cultures argues that NAD + levels are involved in the mechanism. The protective effects of sirtinol and SIRT1 knockdown on NAD + levels and beta-cell death suggested that the sirtuin pathway may be involved.

The sirtuin family (SIRT) and the pancreas
Inhibition of the sirtuin pathway attenuates Zn 2+ or STZ toxicities of β-cells in part by preventing NAD+ depletion [9]; as we have also shown for Zn 2+ neurotoxicity. The sirtuin pathway is involved in Zn 2+induced neuronal, and β-cell NAD + depletion and toxicities [9,13]. The sirtuin family of proteins are NAD + -dependent protein deacetylases resulting in NAD + -catabolism, transcriptional silencing, and transcriptional regulation [43]. SIRT1 is ubiquitously expressed; within the pancreas, SIRT1 is expressed strongly in the cytoplasm of β-cells and weakly in both the nucleus and cytoplasm of β-cells [21]. Young β-cell SIRT1 transgenic (BESTO) mice have increased glucose tolerance under basal conditions [21], but lose this effect with age [44]. This suggests that under young physiologic conditions, SIRT1 overexpression may be beneficial, but under pathophysiologic aged T1DM diabetic conditions (MLDS), SIRT1 may be detrimental perhaps due to potentiation of zinc toxicity. Sirtuins appear to mediate part of the NAD + loss after Zn 2+ and STZ exposures of MIN6 cells, as evidenced by the partial restoration of NAD + levels by sirtuin inhibition for zinc and STZ exposures, but not for cytokine exposures [9]. Sirtuins also act through transcriptional modulation, which may be the predominant mechanism in Zn 2+ neurotoxicity, and cytokine toxicity in β-cells. SIRT1 overexpression was shown to attenuate IL-1β and IFN-γ induced toxicity in RIN β-cells [45]. However, this study was not done in the presence of TNF-α, or under high-density conditions where zinc release and toxicity could play a role. Recently, SIRT1 was shown to attenuate pancreatic β-cell expansion [46], and to decrease hepatic insulin responsiveness [47]. However, the SIRT1 and AMP kinase activator, resveratrol, suppresses T-cell immune responses, and attenuates diabetic incidence in NOD mice [48]. These effects of resveratrol may be SIRT1 independent, mediated instead by oxidant scavenger or AMP kinase activation mediated mechanisms [49,50]. In age or high-fat diet models of T2DM, SIRT1 activation partially mediates the beneficial effects of NAD + restoration in the peripheral tissues affected by T2DM (liver, WAT, skeletal muscle), though effects on the pancreas or beta-cells were not presented [8]. Similar effects in models of T1DM were not presented.
In these studies, we have implicated the sirtuin pathway and SIRT1 in the pancreatic NAD + loss and beta-cell death induced by the in vitro models of T1DM: Zn 2+ , STZ, or mixed cytokine exposures. We have also suggested that SIRT1 plays a role in the diabetic incidence and mortality induced by the STZ or NOD in vivo models of T1DM. These results differ from those reported for SIRT1 on T2DM induced affects in peripheral tissues, where SIRT1 activity appears to mediate the ef-fects of reducing or increasing NAD + levels [8]. The roles of SIRT1 in pancreatic versus peripheral tissue, and in T1DM versus T2DM appear to be varied and complex, and will require further studies to unravel.