Abstract
Type 2 diabetes (T2D) arises when the pancreatic beta-cell fails to compensate for increased insulin needs due to insulin resistance. Glucolipotoxicity (GLT) has been proposed to induce beta-cell dysfunction in T2D by formation of reactive oxygen species (ROS). Here, we examined if modeling glucolipotoxic conditions by high glucose-high free fatty acid (FFA) exposure (GLT) regulates beta-cell iron transport, by increasing the cytosolic labile iron pool (LIP). In isolated mouse islets, the GLT-induced increase in the LIP catalyzed cytosolic ROS formation and induced apoptosis. We show that GLT-induced ROS production is regulated by an increased LIP associated with elevated expression of genes regulating iron import. Using pharmacological and transgenic approaches, we show that iron reduction and decreased iron import protects from GLT-induced ROS production, prevents impairment of the mitochondrial membrane potential (MMP) and inhibits apoptosis. This study identifies a novel pathway underlying GLT-induced apoptosis involving increased iron import, generation of hydroxyl radicals from hydrogen peroxide through the Fenton reaction in the cytosolic compartment associated with dissipation of the MMP and beta-cell apoptosis.
Introduction
Chronic nutritional overload is a very recent event in evolution, against which the development of cellular defense mechanisms has therefore not been advanced by selection pressure. In contrast, fluctuations in circulating substrate levels as a result of bouts of fasting and feeding have prevailed throughout human history and have selected for acute accommodations to nutritional fluxes, thereby defining the physiological regulation of metabolism. The pathophysiology of metabolic disorders is thus in part characterized by the adverse effects of inappropriate handling of nutritional overload, and in part by aberrant and chronic activation of normal metabolic pathways. Much is still to be learned about the molecular sensing, signaling and organelle relaying of substrate overload. Three main hypotheses have been proposed to explain the impairment of cellular replication, differentiation, function and survival caused by increased circulating levels of FFAs and glucose now by many referred to as glucolipotoxicity (GLT) (Muoio & Newgard 2008): oxidative stress mediated by excess generation of reactive oxygen intermediates (ROS); lipid partitioning, i.e. redirection of FFA metabolism from beta-oxidation to build-up of toxic lipid metabolites and lipid storage and activation of low-grade inflammation. The molecular mechanisms regulating and executing the consequences of these major pathways are only partially understood.
In highly metabolically active cells such as cardiomyocytes and hepatocytes exhibiting high mitochondrial densities, the oxidative burden is typically counteracted by high mitochondrial oxidative capacity and by the expression of antioxidant enzymes (Lenzen et al. 1996, Lenzen 2017). In spite of lower mitochondrial density, the highly metabolically active pancreatic beta-cells are uniquely sensitive to oxidative stress due to inadequate antioxidant defense (Lenzen et al. 1996).
Iron-sulfur (Fe-S) proteins in the electron transport chain contribute to the generation of ROS from the mitochondria via the Fenton reaction (Thomas et al. 2009) and in the cytosol via NAPDH oxidase activity (Bedard & Krause 2007). Tissue iron overload confers risk for the development of diabetes (Ellervik et al. 2001, 2011, Jehn et al. 2004, Jiang et al. 2004, Montonen et al. 2012), and diabetes and the metabolic syndrome are associated with elevated circulating ferritin, a marker of cellular iron deposition (Ford & Cogswell 1999, Gabrielsen et al. 2012). Additionally, human and rodent beta-cells accumulate iron in conditions of increased circulating iron, as seen in hereditary hemochromatosis and transfusional iron overload, promoting islet apoptosis, decreased islet mass and impaired insulin secretion due to increased ROS production (Lu et al. 1991, Lu & Hayashi 1994, Cooksey et al. 2004). These clinical observations confirm that the beta-cell is particularly vulnerable to aberrant intracellular iron accumulation.
It is not known if GLT plays a causal role in enhancing beta-cell iron transport, thereby leading to excess ROS formation, beta-cell dysfunction and apoptosis. Therefore, we investigated this important science gap of considerable translational importance for the development of novel antidiabetic therapies targeting beta-cell iron handling.
Materials and methods
Animal care
Mice were housed with access to water and food ad libitum, in a 12-h light/darkness cycle (Hansen et al. 2012) (Supplementary materials and methods, see section on supplementary data given at the end of this article).
All the protocols were approved by the Animal Care Committee of the University of Toronto or Danish Animal Inspectorate.
Islet isolation
Islets were isolated after collagenase digestion and hand-picked (Supplementary materials and methods).
Cell culture and treatment
MIN6 cells were cultured in DMEM (Gibco) containing 4.5 g/L glucose, 10% fetal bovine serum (FBS, Gibco), 0.1% penicillin and streptomycin and 50 µmol/L beta-mercaptoethanol (Gibco).
Isolated islets were cultured in RPMI 1640 supplemented with 10% FBS and 0.1% penicillin and streptomycin at 37°C with 5% CO2.
For GLT treatment, culture medium was supplemented with or without endotoxin-free glucose and/or palmitate +/− oleate (both from Invitrogen) dissolved in 80% ethanol. Fatty acids were conjugated to BSA for at least 4 h at 37°C in a molar ratio of 3:1 in culture medium prior to use, and the final concentrations were palmitate (400 µM), combination of fatty acids palmitate (200 µM) and oleate (200 µM) and glucose 30 mM. Islets and MIN6 cells were treated with up to 72 h and human islets for 24 h. Culture medium supplemented with BSA and ethanol was used as control for GLT.
IL-1β (150 pg/mL) (BD Bioscience Pharmingen, San Diego, CA, USA) and IFN-γ (5 ng/mL) (R&D Systems) exposure was carried out for 24 h.
The iron chelator desferrioxamine (DFO) (Sigma) was dissolved in water and deferasirox (D-sirox) (Santa Cruz Biotechnology) diluted in DMSO and added directly to the culture medium (Supplementary materials and methods).
Glucose-stimulated insulin secretion perifusion and total islet insulin content
Glucose-stimulated insulin secretion
Basal insulin secretion was established by pre-incubating islets or MIN6 cells for at least 1 h in 3 mM glucose Krebs–Ringer buffer (KRB). For stimulation, islets or MIN6 cell were incubated in KRB with the given concentrations of glucose for 30 min. KCl-induced insulin secretion was performed with 16.7 mM glucose and 30 mM KCl for 30 min. After glucose-stimulated insulin secretion (GSIS) experiments, islets were lysed in ethanol–HCl at −20°C for at least 24 h. Lysates were dried in a vacuum centrifuge, pellets were dissolved in ultra-pure water and DNA was measured by spectrophotometry using NanoDrop 2000 (Thermo Fisher).
Islet perifusion
Fifty islets from littermates were preincubated in KRB (3 mM Glc) for 2 h in KRB prior to perifusion, and then perifused at a flow rate of 1 mL per minute (Hansen et al. 2012). Total islet insulin was extracted from 10 to 20 islets lysed in ethanol–HCl at −20°C for at least 24 h. Insulin was measured using an insulin homogeneous time-resolved fluorescence kit according to manufacturer’s protocol (Cisbio, Codolet, France) or insulin ELISA (Kekow et al. 1988).
Calcein assay
Islets were incubated in imaging buffer with 0.25 µM calcein-AM (Molecular Probes) for 1 h. Fluorescence was detected and analyzed as described (Hansen et al. 2012) and Supplementary materials and methods. Total islet calcein loading was measured after chelation with 0.015 mM D-Sirox, and changes in LIP was calculated as difference between total calcein load and quenched calcein fluorescence.
Quantitative PCR
RNA was extracted using RNeasy Mini Kit (Qiagen), and RNA was measured using NanoDrop 2000 (Thermo Fisher). cDNA was synthesized using SuperScript II Reverse Transcriptase (Sigma), run with SybrGreen (Applied Biosystems) and analyzed on ViiA7 Real-Time PCR System (Applied Biosystems). Primers were purchased from TagCopenhagen (Copenhagen, DK) (Supplementary materials and methods).
Mitochondrial membrane potential and HIF1a translocation
Mitochondrial membrane potential (MMP) and HIF1a translocation were measured in whole islets as previously described (Chan et al. 2001, Cheng et al. 2010) (Supplementary materials and methods). In brief, islets were loaded with Rhodamine123 (Rhod123; Molecular Probes) (2.5 mg/mL) for 10 min in imaging buffer and the increase in MMP was measured as quenching of Rhod123 fluorescence. Fluorescence was recorded and analyzed as described in Supplementary materials and methods.
Apoptosis assay
Twenty-five islets per condition in duplicates were used for apoptosis assay. Apoptosis was detected using Roche Cell Death Detection ELISA (Roche) according to the manufacturer’s protocol.
Reactive oxygen species
Islets were loaded with CM-H2DCFDA (Molecular Probes) (10 µM) for 45 min in imaging buffer, washed once and ROS was detected on Flex Station 3 as described (Friberg et al. 2010) and Supplementary materials and methods).
Quantification of cytosolic and mitochondrial thiol oxidation
Islets were cultured, in batches of ten at different glucose concentrations and/or palmitate with or without DFO (0.1 mM), infected with adenovirus coding mito-roGFP1 or cyto-roGFP1 and transferred to a perifusion chamber in inverted microscope and fluorescence was measured every 30 s. After 20 min, 10 mM DTT was added to maximally reduce roGFP1, followed by 100 μM aldrithiol to maximally oxidize roGPF. The fluorescence ratio during the initial perifusion period was normalized to the ratio in DTT (set to 0%) and in aldrithiol (set to 100%). (Roma et al. 2012) and Supplementary materials and methods).
Statistics
Statistics were performed using Prism (GraphPad). Student’s t-test, and one- and two-way ANOVA with Bonferroni’s post hoc test were used to assess significance. P < 0.05 was considered statistically significant.
Results
GLT increases the beta-cell labile iron pool
Islets isolated from mice fed a high fat diet (HFD) for 8 weeks had significantly increased LIP compared to islets from control-fed mice (Fig. 1A). GLT exposure in vitro for up to 72 h (Fig. 1B), but not palmitate alone (Supplementary Fig. 1A), significantly increased the islet LIP after 48 h compared to vehicle-treated islets. Of note, a combination of palmitate and oleate, considered to be a more relevant model of GLT in vivo, significantly increased the LIP in isolated human islets (Fig. 1C) to an even higher degree than the pro-inflammatory cytokines IL-1β and IFN-γ (Fig. 1D), previously shown to upregulate the iron transporter divalent metal transporter (DMT) 1 in human islets (Hansen et al. 2012).
The beta-cell-labile iron pool is increased by glucolipotoxicity (GLT). (A) Quantification of the labile iron pool (LIP) from isolated islets of mice fed a high-fat diet (HFD) or chow for 7 weeks was measured by calcein assay. Means + s.e.m. of n = 4–5, *P < 0.05, Student’s t-test. (B) Quantification of the LIP from isolated mouse islets exposed to GLT (400 µM palmitate and 30 mM glucose) or vehicle (BSA and ethanol) for up to 72 h using calcein assay. Means + s.e.m. of 4–7 independent experiments, *P < 0.05, Student’s t-test. (C) Quantification of the LIP in human islets exposed to a combination of palmitate (200 µM) and oleate (200 µM) for 24 h using calcein assay. Means + s.e.m. of n = 4, *P < 0.05, Student’s t-test. (D) Quantification of the LIP in human islets exposed to a combination of IL-1β (150 pg/mL) and IFN-γ (5 ng/mL) for 24 h using calcein assay. Means + s.e.m. of n = 3, *P < 0.05, Student’s t-test. (E) Quantification of the LIP in MIN6 cells exposed to with palmitate (400 µM) or vehicle (BSA and ethanol) for up to 72 h using calcein assay. Means + s.e.m. of 3–4 independent experiments, *P < 0.05, Student’s t-test. (F and G) Transferrin (Trfr) and divalent metal transporter 1 (Dmt1) mRNA in MIN6 cells exposed to palmitate (400 µM) or vehicle (BSA and ethanol) for up to 24 h. Means + s.e.m. of six independent experiments, *P < 0.05, ****P < 0.0001, Student’s t-test.
Citation: Journal of Molecular Endocrinology 61, 2; 10.1530/JME-17-0262
The beta-cell-labile iron pool is increased by glucolipotoxicity (GLT). (A) Quantification of the labile iron pool (LIP) from isolated islets of mice fed a high-fat diet (HFD) or chow for 7 weeks was measured by calcein assay. Means + s.e.m. of n = 4–5, *P < 0.05, Student’s t-test. (B) Quantification of the LIP from isolated mouse islets exposed to GLT (400 µM palmitate and 30 mM glucose) or vehicle (BSA and ethanol) for up to 72 h using calcein assay. Means + s.e.m. of 4–7 independent experiments, *P < 0.05, Student’s t-test. (C) Quantification of the LIP in human islets exposed to a combination of palmitate (200 µM) and oleate (200 µM) for 24 h using calcein assay. Means + s.e.m. of n = 4, *P < 0.05, Student’s t-test. (D) Quantification of the LIP in human islets exposed to a combination of IL-1β (150 pg/mL) and IFN-γ (5 ng/mL) for 24 h using calcein assay. Means + s.e.m. of n = 3, *P < 0.05, Student’s t-test. (E) Quantification of the LIP in MIN6 cells exposed to with palmitate (400 µM) or vehicle (BSA and ethanol) for up to 72 h using calcein assay. Means + s.e.m. of 3–4 independent experiments, *P < 0.05, Student’s t-test. (F and G) Transferrin (Trfr) and divalent metal transporter 1 (Dmt1) mRNA in MIN6 cells exposed to palmitate (400 µM) or vehicle (BSA and ethanol) for up to 24 h. Means + s.e.m. of six independent experiments, *P < 0.05, ****P < 0.0001, Student’s t-test.
Citation: Journal of Molecular Endocrinology 61, 2; 10.1530/JME-17-0262
The beta-cell-labile iron pool is increased by glucolipotoxicity (GLT). (A) Quantification of the labile iron pool (LIP) from isolated islets of mice fed a high-fat diet (HFD) or chow for 7 weeks was measured by calcein assay. Means + s.e.m. of n = 4–5, *P < 0.05, Student’s t-test. (B) Quantification of the LIP from isolated mouse islets exposed to GLT (400 µM palmitate and 30 mM glucose) or vehicle (BSA and ethanol) for up to 72 h using calcein assay. Means + s.e.m. of 4–7 independent experiments, *P < 0.05, Student’s t-test. (C) Quantification of the LIP in human islets exposed to a combination of palmitate (200 µM) and oleate (200 µM) for 24 h using calcein assay. Means + s.e.m. of n = 4, *P < 0.05, Student’s t-test. (D) Quantification of the LIP in human islets exposed to a combination of IL-1β (150 pg/mL) and IFN-γ (5 ng/mL) for 24 h using calcein assay. Means + s.e.m. of n = 3, *P < 0.05, Student’s t-test. (E) Quantification of the LIP in MIN6 cells exposed to with palmitate (400 µM) or vehicle (BSA and ethanol) for up to 72 h using calcein assay. Means + s.e.m. of 3–4 independent experiments, *P < 0.05, Student’s t-test. (F and G) Transferrin (Trfr) and divalent metal transporter 1 (Dmt1) mRNA in MIN6 cells exposed to palmitate (400 µM) or vehicle (BSA and ethanol) for up to 24 h. Means + s.e.m. of six independent experiments, *P < 0.05, ****P < 0.0001, Student’s t-test.
Citation: Journal of Molecular Endocrinology 61, 2; 10.1530/JME-17-0262
To verify that the increase in islet LIP was beta-cell derived, we exposed insulin-producing mouse MIN6 cells to palmitate for up to 72 h. We did not manipulate the glucose concentration in MIN6 cell experiments, since MIN6 cells are normally grown in 25 mM glucose. Palmitate significantly increased LIP in MIN6 cells after 24 h of exposure (Fig. 1E). This increase in beta-cell LIP was associated with decreased glucose-induced insulin secretion and increased ROS production and apoptosis (Supplementary Fig. 1B, C and D).
Next, we investigated if the GLT-mediated increase in the beta-cell LIP was associated with changes in expression of iron-handling genes in MIN6 cells exposed to palmitate for up to 24 h (Fig. 1F and G and Supplementary Fig. 1E, F and G). Palmitate significantly increased expression of the iron import genes DMT1 and transferrin receptor (Trfr) after 18 and 24 h, while the iron storage gene ferritin heavy chain (Fth) was transiently downregulated at 6 h. Ferritin light chain (Ftl) and the iron export gene ferroportin (Fpn) were not regulated by palmitate exposure. Since increased ROS production has been suggested to induce beta-cell dysfunction through altered transcriptional activity (Prentice et al. 2014), we measured the expression of the beta-cell transcription factor pancreatic and duodenal homeobox 1 (Pdx1). Pdx1 was transiently downregulated after 6 h of palmitate exposure, while insulin 1 (Ins1) gene expression was unaffected (Supplementary Fig. 1H and I), suggesting that GLT-induced transient differences in transcriptional activity are not responsible for the observed differences in GSIS.
Iron-mediated beta-cell apoptosis is associated with an increase in cytosolic thiol oxidation
To investigate the causal role of GLT-induced increases in intracellular beta-cell LIP for apoptosis, we used the pharmacological iron chelator DFO to decrease the LIP in islets. DFO reduced the beta-cell LIP in a concentration-dependent manner after 48 h of treatment (Fig. 2A), without impacting absolute levels of GSIS, islet insulin content, accumulated insulin secretion or HIF1α nuclear translocation (Supplementary Fig. 2A, B and C and data not shown). The stimulatory index of the GSIS from the iron-chelated islets was decreased as expected (Supplementary Fig. 2A). The impact of the beta-cell LIP on GLT-induced apoptosis was then investigated in isolated islets exposed to vehicle or GLT for 48 h in the presence or absence of DFO. GLT significantly increased islet apoptosis two-fold, which was completely prevented by co-treatment with DFO (Fig. 2B).
Glucolipotoxicity (GLT) induces iron-dependent cytosolic ROS production, apoptosis and secretory dysfunction. (A) Quantification of the labile iron pool (LIP) from isolated islets treated with different concentrations of DFO using calcein assay. Means + s.e.m. of 4–6 independent experiments. (B) Apoptosis measurement using DNA–histone complex detection by ELISA in isolated islets exposed to GLT (400 µM palmitate and 30 mM glucose) or vehicle (BSA and ethanol) with or without DFO (0.1 mM) for 48 h. Means + s.e.m. of four independent experiments, *P < 0.05 Control GLT vs DFO GLT, #P < 0.05 Control Vehicle vs Control GLT, two-way ANOVA. Cytosolic (C) and mitochondrial (D) thiol oxidation in islets exposed to medium containing 11 mM glucose (G11), G11 supplemented with 400 µM palmitate (G11_PA), 30 mM glucose (G30) or G30 with 400 µM palmitate (G30_PA) with or without DFO (0.1 mM) for 48 h using cyto-roGFP1 and mito-roGFP1 probes, respectively. Means + s.e.m. of 3–5 independent experiments, **P < 0.01 G30_PA vs G30_PA DFO, ###P<0.001 G11 vs G30_PA, Student’s t-test. (E) GSIS from isolated islets exposed to FFA (200 µM palmitate and 200 µM oleate) or vehicle (BSA and ethanol) with or without DFO (0.1 mM) for 48 h. Means + s.e.m. of four independent experiments, **P < 0.01, Student’s t-test.
Citation: Journal of Molecular Endocrinology 61, 2; 10.1530/JME-17-0262
Glucolipotoxicity (GLT) induces iron-dependent cytosolic ROS production, apoptosis and secretory dysfunction. (A) Quantification of the labile iron pool (LIP) from isolated islets treated with different concentrations of DFO using calcein assay. Means + s.e.m. of 4–6 independent experiments. (B) Apoptosis measurement using DNA–histone complex detection by ELISA in isolated islets exposed to GLT (400 µM palmitate and 30 mM glucose) or vehicle (BSA and ethanol) with or without DFO (0.1 mM) for 48 h. Means + s.e.m. of four independent experiments, *P < 0.05 Control GLT vs DFO GLT, #P < 0.05 Control Vehicle vs Control GLT, two-way ANOVA. Cytosolic (C) and mitochondrial (D) thiol oxidation in islets exposed to medium containing 11 mM glucose (G11), G11 supplemented with 400 µM palmitate (G11_PA), 30 mM glucose (G30) or G30 with 400 µM palmitate (G30_PA) with or without DFO (0.1 mM) for 48 h using cyto-roGFP1 and mito-roGFP1 probes, respectively. Means + s.e.m. of 3–5 independent experiments, **P < 0.01 G30_PA vs G30_PA DFO, ###P<0.001 G11 vs G30_PA, Student’s t-test. (E) GSIS from isolated islets exposed to FFA (200 µM palmitate and 200 µM oleate) or vehicle (BSA and ethanol) with or without DFO (0.1 mM) for 48 h. Means + s.e.m. of four independent experiments, **P < 0.01, Student’s t-test.
Citation: Journal of Molecular Endocrinology 61, 2; 10.1530/JME-17-0262
Glucolipotoxicity (GLT) induces iron-dependent cytosolic ROS production, apoptosis and secretory dysfunction. (A) Quantification of the labile iron pool (LIP) from isolated islets treated with different concentrations of DFO using calcein assay. Means + s.e.m. of 4–6 independent experiments. (B) Apoptosis measurement using DNA–histone complex detection by ELISA in isolated islets exposed to GLT (400 µM palmitate and 30 mM glucose) or vehicle (BSA and ethanol) with or without DFO (0.1 mM) for 48 h. Means + s.e.m. of four independent experiments, *P < 0.05 Control GLT vs DFO GLT, #P < 0.05 Control Vehicle vs Control GLT, two-way ANOVA. Cytosolic (C) and mitochondrial (D) thiol oxidation in islets exposed to medium containing 11 mM glucose (G11), G11 supplemented with 400 µM palmitate (G11_PA), 30 mM glucose (G30) or G30 with 400 µM palmitate (G30_PA) with or without DFO (0.1 mM) for 48 h using cyto-roGFP1 and mito-roGFP1 probes, respectively. Means + s.e.m. of 3–5 independent experiments, **P < 0.01 G30_PA vs G30_PA DFO, ###P<0.001 G11 vs G30_PA, Student’s t-test. (E) GSIS from isolated islets exposed to FFA (200 µM palmitate and 200 µM oleate) or vehicle (BSA and ethanol) with or without DFO (0.1 mM) for 48 h. Means + s.e.m. of four independent experiments, **P < 0.01, Student’s t-test.
Citation: Journal of Molecular Endocrinology 61, 2; 10.1530/JME-17-0262
We next asked if an increase in the LIP in islets exposed to a moderate glucose concentration (G11), G11 supplemented with palmitate (G11_PA), high glucose (G30) or G30 supplemented with palmitate (G30_PA) caused an increase of ROS in the cytosol or in the mitochondria (Fig. 2C and D). We first verified that DFO did indeed reduce GLT-induced LIP in a concentration-dependent manner in MIN6 cells (Supplementary Fig. 2D). Palmitate exposure for 48 h increased thiol oxidation in the cytosol independently of iron, since DFO co-treatment did not reduce cytosolic thiol oxidation (Fig. 2C). This was consistent with the inability of palmitate alone to increase the beta-cell LIP (Supplementary Fig. 1A). In contrast, high glucose in combination with palmitate increased islet cytosolic thiol oxidation, which was decreased by iron chelation using DFO (Fig. 2C). Islet mitochondrial thiol oxidation was significantly decreased by DFO at basal, G30 and G30_PA, whereas a 48-h exposure to high glucose or palmitate did not change the mitochondrial thiol oxidation (Fig. 2D).
DFO treatment protected against FFA-mediated dysfunction estimated by GSIS and KCl-induced insulin secretion in islets exposed to a combination of palmitate and oleate for 24 h (Fig. 2E) shown above to increase human islet LIP (Fig. 1C),
Iron is required for normal mitochondrial function
Tamoxifen-inducible beta-cell specific DMT1 KO (Cre+Flox+) islets (Supplementary Fig. 3A) had significantly decreased LIP (Fig. 3A) resulting in decreased ROS formation compared to WT (Cre−Flox+) islets (Fig. 3B).
Iron is important for beta-cell mitochondrial function. (A) Quantification of the labile iron pool (LIP) from isolated islets of inducible beta-cell specific WT (Cre−Flox+) or KO (Cre+Flox+) mice. Means + s.e.m. of n = 5–6 independent experiments, ***P < 0.001, Student’s t-test. (B) ROS production in isolated islets of Cre−Flox+ and Cre+Flox+ islets measured by H2DCF fluorescence. Means + s.e.m. of n = 9, *P < 0.05, Student’s t-test. (C) MMP measurements using Rhod123 in isolated Cre−Flox+ and Cre+Flox+ islets. Left, means + s.e.m. fluorescence traces from islets analyzed, n = 8–9. Right, means + s.e.m. AUC of fluorescence traces normalized to time of n = 5 independent experiments, *P < 0.05, Student’s t-test.
Citation: Journal of Molecular Endocrinology 61, 2; 10.1530/JME-17-0262
Iron is important for beta-cell mitochondrial function. (A) Quantification of the labile iron pool (LIP) from isolated islets of inducible beta-cell specific WT (Cre−Flox+) or KO (Cre+Flox+) mice. Means + s.e.m. of n = 5–6 independent experiments, ***P < 0.001, Student’s t-test. (B) ROS production in isolated islets of Cre−Flox+ and Cre+Flox+ islets measured by H2DCF fluorescence. Means + s.e.m. of n = 9, *P < 0.05, Student’s t-test. (C) MMP measurements using Rhod123 in isolated Cre−Flox+ and Cre+Flox+ islets. Left, means + s.e.m. fluorescence traces from islets analyzed, n = 8–9. Right, means + s.e.m. AUC of fluorescence traces normalized to time of n = 5 independent experiments, *P < 0.05, Student’s t-test.
Citation: Journal of Molecular Endocrinology 61, 2; 10.1530/JME-17-0262
Iron is important for beta-cell mitochondrial function. (A) Quantification of the labile iron pool (LIP) from isolated islets of inducible beta-cell specific WT (Cre−Flox+) or KO (Cre+Flox+) mice. Means + s.e.m. of n = 5–6 independent experiments, ***P < 0.001, Student’s t-test. (B) ROS production in isolated islets of Cre−Flox+ and Cre+Flox+ islets measured by H2DCF fluorescence. Means + s.e.m. of n = 9, *P < 0.05, Student’s t-test. (C) MMP measurements using Rhod123 in isolated Cre−Flox+ and Cre+Flox+ islets. Left, means + s.e.m. fluorescence traces from islets analyzed, n = 8–9. Right, means + s.e.m. AUC of fluorescence traces normalized to time of n = 5 independent experiments, *P < 0.05, Student’s t-test.
Citation: Journal of Molecular Endocrinology 61, 2; 10.1530/JME-17-0262
Since cytosolic ROS production promotes mitochondrial superoxide generation that may impair mitochondrial function (Coughlan et al. 2009), we next investigated the change in MMP of isolated Cre−Flox+ and Cre+Flox+ islets upon glucose stimulation (Fig. 3C). At 3 mM glucose, Cre+Flox+ islets exhibited a significant hyperpolarization of the MMP compared to Cre−Flox+ islets, confirming that the decreased GSIS observed in Cre+Flox+ islets, is due to inability to increase the MMP upon high glucose exposure. The MMP of Cre−Flox+ and Cre+Flox+ islets were similar when stimulated with 20 mM glucose, as well as when the mitochondria were completely depolarized using NaN3.
DMT1 KO protects from GLT-induced islet apoptosis associated with decreased ROS formation and preservation of the MMP
GLT-induced islet apoptosis was completely inhibited in Cre+Flox+ islets compared to Cre−Flox+ islets (Fig. 4A) and correlated with decreased GLT-induced ROS formation (Fig. 4B). Additionally, since lipotoxic exposure of islets promotes attraction of pro-inflammatory macrophages to the islets to facilitate islet dysfunction (Eguchi et al. 2012), we investigated islet secretion of IL-6, TNF-α and MCP-1 (Supplementary Fig. 4A, B, C, D, E and F). We did not observe upregulation of the cytokines or the chemoattractant by GLT; however, DFO treatment significantly reduced the production of MCP-1. Iron chelation with DFO, but not DMT1 KO, significantly increased islet production of TNF-α, suggesting that this action is not associated with iron-chelating properties of DFO.
Beta-cell-specific DMT1 KO reduces GLT-mediated apoptosis ROS production and mitochondrial damage. (A) Apoptosis measurement using DNA-histone complex detection by ELISA in isolated Cre−Flox+ and Cre+Flox+ islets exposed to GLT (400 µM palmitate and 30 mM glucose) or vehicle (BSA and ethanol). Means + s.e.m. of four independent experiments, **P < 0.01, Student’s t-test. (B) ROS production in isolated islets of Cre−Flox+ and Cre+Flox+ islets exposed to GLT (400 µM palmitate and 30 mM glucose) or vehicle (BSA and ethanol) measured by H2DCF fluorescence. Means + s.e.m. of three independent experiments, **P < 0.01 and **P < 0.01 Vehicle Cre−Flox+ vs GLT Cre−Flox+, ##P < 0.01 GLT Cre−Flox+ vs GLT Cre+Flox+, Student’s t-test. (C) Insulin secretion measured by perifusion in isolated Cre−Flox+ and Cre+Flox+ islets exposed to GLT (400 µM palmitate and 30 mM glucose) or vehicle (BSA and ethanol). Left panel, perifusion traces. Middle panel, stimulatory index of islet perifusion experiment. Right panel, total islet insulin after perifusion. Graphs show means + s.e.m. of four independent experiments, **P < 0.01, Student’s t-test. (D) MMP measurements using Rhod123 in isolated Cre−Flox+ islets exposed to GLT (400 µM palmitate and 30 mM glucose) or vehicle (BSA and ethanol). Left panel, means + s.e.m. of fluorescence traces from islets analyzed, n = 5–7. Right panel, means + s.e.m. AUC of fluorescence traces normalized to time of n = 4–7 independent experiments, *P < 0.05, **P < 0.01, Student’s t-test. (E) MMP measurements using Rhod123 in isolated Cre+Flox+ islet exposed to GLT (400 µM palmitate and 30 mM glucose) or vehicle (BSA and ethanol). Left panel, means + s.e.m. of fluorescence traces from islets analyzed, n = 4–5. Right, means + s.e.m. AUC of fluorescence traces normalized to time of n = 4–5 independent experiments, *P < 0.05, **P < 0.01, Student’s t-test.
Citation: Journal of Molecular Endocrinology 61, 2; 10.1530/JME-17-0262
Beta-cell-specific DMT1 KO reduces GLT-mediated apoptosis ROS production and mitochondrial damage. (A) Apoptosis measurement using DNA-histone complex detection by ELISA in isolated Cre−Flox+ and Cre+Flox+ islets exposed to GLT (400 µM palmitate and 30 mM glucose) or vehicle (BSA and ethanol). Means + s.e.m. of four independent experiments, **P < 0.01, Student’s t-test. (B) ROS production in isolated islets of Cre−Flox+ and Cre+Flox+ islets exposed to GLT (400 µM palmitate and 30 mM glucose) or vehicle (BSA and ethanol) measured by H2DCF fluorescence. Means + s.e.m. of three independent experiments, **P < 0.01 and **P < 0.01 Vehicle Cre−Flox+ vs GLT Cre−Flox+, ##P < 0.01 GLT Cre−Flox+ vs GLT Cre+Flox+, Student’s t-test. (C) Insulin secretion measured by perifusion in isolated Cre−Flox+ and Cre+Flox+ islets exposed to GLT (400 µM palmitate and 30 mM glucose) or vehicle (BSA and ethanol). Left panel, perifusion traces. Middle panel, stimulatory index of islet perifusion experiment. Right panel, total islet insulin after perifusion. Graphs show means + s.e.m. of four independent experiments, **P < 0.01, Student’s t-test. (D) MMP measurements using Rhod123 in isolated Cre−Flox+ islets exposed to GLT (400 µM palmitate and 30 mM glucose) or vehicle (BSA and ethanol). Left panel, means + s.e.m. of fluorescence traces from islets analyzed, n = 5–7. Right panel, means + s.e.m. AUC of fluorescence traces normalized to time of n = 4–7 independent experiments, *P < 0.05, **P < 0.01, Student’s t-test. (E) MMP measurements using Rhod123 in isolated Cre+Flox+ islet exposed to GLT (400 µM palmitate and 30 mM glucose) or vehicle (BSA and ethanol). Left panel, means + s.e.m. of fluorescence traces from islets analyzed, n = 4–5. Right, means + s.e.m. AUC of fluorescence traces normalized to time of n = 4–5 independent experiments, *P < 0.05, **P < 0.01, Student’s t-test.
Citation: Journal of Molecular Endocrinology 61, 2; 10.1530/JME-17-0262
Beta-cell-specific DMT1 KO reduces GLT-mediated apoptosis ROS production and mitochondrial damage. (A) Apoptosis measurement using DNA-histone complex detection by ELISA in isolated Cre−Flox+ and Cre+Flox+ islets exposed to GLT (400 µM palmitate and 30 mM glucose) or vehicle (BSA and ethanol). Means + s.e.m. of four independent experiments, **P < 0.01, Student’s t-test. (B) ROS production in isolated islets of Cre−Flox+ and Cre+Flox+ islets exposed to GLT (400 µM palmitate and 30 mM glucose) or vehicle (BSA and ethanol) measured by H2DCF fluorescence. Means + s.e.m. of three independent experiments, **P < 0.01 and **P < 0.01 Vehicle Cre−Flox+ vs GLT Cre−Flox+, ##P < 0.01 GLT Cre−Flox+ vs GLT Cre+Flox+, Student’s t-test. (C) Insulin secretion measured by perifusion in isolated Cre−Flox+ and Cre+Flox+ islets exposed to GLT (400 µM palmitate and 30 mM glucose) or vehicle (BSA and ethanol). Left panel, perifusion traces. Middle panel, stimulatory index of islet perifusion experiment. Right panel, total islet insulin after perifusion. Graphs show means + s.e.m. of four independent experiments, **P < 0.01, Student’s t-test. (D) MMP measurements using Rhod123 in isolated Cre−Flox+ islets exposed to GLT (400 µM palmitate and 30 mM glucose) or vehicle (BSA and ethanol). Left panel, means + s.e.m. of fluorescence traces from islets analyzed, n = 5–7. Right panel, means + s.e.m. AUC of fluorescence traces normalized to time of n = 4–7 independent experiments, *P < 0.05, **P < 0.01, Student’s t-test. (E) MMP measurements using Rhod123 in isolated Cre+Flox+ islet exposed to GLT (400 µM palmitate and 30 mM glucose) or vehicle (BSA and ethanol). Left panel, means + s.e.m. of fluorescence traces from islets analyzed, n = 4–5. Right, means + s.e.m. AUC of fluorescence traces normalized to time of n = 4–5 independent experiments, *P < 0.05, **P < 0.01, Student’s t-test.
Citation: Journal of Molecular Endocrinology 61, 2; 10.1530/JME-17-0262
Since Cre+Flox+ islets were completely protected from apoptosis induced by 48 h of GLT exposure, we wanted to investigate if Cre+Flox+ islets had preserved secretory function. The stimulatory index of the GLT-exposed DMT1 KO islets was significantly increased compared to GLT-exposed WT islets (Fig. 4C), showing that DMT1 KO in addition to protecting from GLT-induced apoptosis improves the ability of the islets to respond to glucose.
We next wanted to investigate if reduced ROS formation in Cre+Flox+ islets protected mitochondrial function by measuring islet MMP after 48-h exposure to GLT (Fig. 4D and E). GLT did not affect the MMP at 3 mM glucose in Cre−Flox+ islets; notably however, at 20 mM glucose, Cre−Flox+ islets exposed to GLT had significantly decreased MMP compared to vehicle-exposed Cre−Flox+ islets as measured by the area under the curve. In Cre+Flox+ islets, 48-h exposure to GLT neither affected the MMP at 3 mM glucose nor the MMP at 20 mM glucose compared to vehicle-treated Cre+Flox+ islet, suggesting that protection from cytosolic ROS formation prevented mitochondrial damage and subsequent impairment in MMP.
Discussion
In the present study, we investigated the mechanisms of GLT-induced beta-cell ROS production and cellular dysfunction. We found that GLT specifically alters the iron metabolism in beta-cells both in vivo and in vitro resulting in apoptosis. GLT induced Dmt1 and Trfr mRNA upregulation in beta-cells, and this was associated with increased iron uptake, increased intracellular LIP, cytosolic ROS production, mitochondrial dysfunction and beta-cell apoptosis. Importantly, we observed an increase in islet LIP from islets isolated from mice fed HFD for 8 weeks, where severe beta-cell damage is yet to occur but dysfunction is progressing. This suggests that increases in LIP and ROS formation might precede apoptosis in vivo.
Our work is in line with previously published findings by ourselves and others, where decreased dietary iron or decreased iron uptake protects beta-cells from HFD-induced dysfunction (Cooksey et al. 2010, Hansen et al. 2012). However, this observation was recently challenged in a study claiming that palmitate-mediated beta-cell apoptosis is mediated through a reduction in LIP and that DFO promotes beta-cell apoptosis (Jung et al. 2015). Because DFO is a potent iron chelator and iron is an absolute growth requirement for insulin-producing cells (Nielsen 1989), it is not surprising to us that DFO induces beta-cell apoptosis, which we have in fact demonstrated in MIN6 cells at concentrations of ≥0.1 mM (JB Hansen & T Mandrup-Poulsen, unpublished data). It is therefore important to carefully titrate iron chelators for such studies.
Dmt1 and Trfr are likely upregulated by JNK activation and ER stress, since (1) they contain predicted binding sites for the transcription factors AP1 and XBP1 respectively; (2) AP1 activity is increased by glucotoxicity in a beta-cell line and (3) Xbp1 mRNA is increased in rat islets by palmitate exposure and by palmitate and oleate exposure in human islets, respectively (Cunha et al. 2008, Maedler et al. 2008). These observations also provide a potential mechanism for the synergistic effect of glucose and FFAs in inducing the LIP in mouse beta-cells.
From studies in hepatocytes, another highly metabolically active cell, it has been proposed that GLT-induced ROS formation is caused by physical interaction between the mitochondria and ER with release of calcium from the ER to the mitochondria, resulting in decreased mitochondrial function, increased mitochondrial ROS formation and JNK activation (Arruda et al. 2014). Importantly, our data suggest that GLT induces cytosolic ROS production, rather than increasing mitochondrial ROS production. This, however, does not exclude GLT-induced beta-cell ER–mitochondrial interactions via calcium signaling and mitochondrial ROS production post or prior to iron accumulation, since DFO and DMT1 have limited or no calcium affinity (Munoz et al. 2011, Ehrnstorfer et al. 2014).
Recently, iron uptake in metabolically unstressed human beta-cells was proposed to be carried out primarily via non-transferrin-dependent iron import by the plasma membrane protein ZIP14 (solute carrier family 39 member 14, SLC39A14), independently of DMT1 and TRFR (Coffey & Knutson 2017), suggesting the existence of two distinct pathways for iron uptake. We hypothesize a physiological role for DMT1 to maintain insulin secretion by controlling the LIP and thereby the ROS levels required to fine-tune insulin secretion driven by the stimulus–secretion coupling. Of note, we found that the iron chelator DFO decreased the islet mitochondrial but not the cytosolic thiol oxidation in the absence of metabolic stress. This suggests that under normal conditions, the cytosolic LIP and ROS production is minimal and their contribution to GSIS likely to be limited. In addition, it confirms previous findings that mitochondrial function and ROS production primarily contribute to GSIS (Leloup et al. 2009) and shows the importance of iron in maintaining mitochondrial ROS production.
A caveat for the in vivo translation of the findings in this study is the diverse reports regarding the cellular effects and mechanisms of glucotoxicity, lipotoxicity and GLT and their dependence on experimental contexts (type of cell line, species, animal model for ex vivo studies, duration or concentrations of exposures). It should be noted, however, that we observed protection against HFD-induced glucose intolerance associated with increased circulating insulin levels in mice with inducible beta-cell-specific deletion of DMT-1 (Hansen et al. 2012), providing in vivo evidence for the validity of the in vitro observations in the present study.
We conclude that iron transport via DMT-1 is a novel central mechanism relaying GLT with generation of hydroxyl radicals from hydrogen peroxide through the Fenton reaction and ROS formation in the cytosolic compartment associated with dissipation of the MMP and beta-cell apoptosis. Taken together with the accumulating epidemiological evidence suggesting that iron overload leads to diabetes (Ellervik et al. 2011, Bonfils et al. 2015) and that patients with diabetes have increased tissue iron accumulation (Rajpathak et al. 2009, Gabrielsen et al. 2012), our findings highlight control of iron levels as an attractive novel treatment option for the prevention of diabetes. Clinical trials are warranted to validate this translation of our preclinical findings.
Supplementary data
This is linked to the online version of the paper at https://doi.org/10.1530/JME-17-0262.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding
This work was supported by a postdoc grant from Danish Council of Independent Research for J B H.
Acknowledgements
The authors would like to thank Dr Nancy C. Andrews for providing the floxed Slc11a2 mice and Drs Doug Melton and Palle Serup for the Cre-ERT mice.
References
Arruda AP, Pers BM, Parlakgul G, Guney E, Inouye K & Hotamisligil GS 2014 Chronic enrichment of hepatic endoplasmic reticulum-mitochondria contact leads to mitochondrial dysfunction in obesity. Nature Medicine 20 1427–1435. (https://doi.org/10.1038/nm.3735)
Bedard K & Krause KH 2007 The NOX family of ROS-generating NADPH oxidases: physiology and pathophysiology. Physiological Reviews 87 245–313. (https://doi.org/10.1152/physrev.00044.2005)
Bonfils L, Ellervik C, Friedrich N, Linneberg A, Sandholt CH, Jorgensen ME, Jorgensen T, Hansen T, Pedersen O & Allin KH 2015 Fasting serum levels of ferritin are associated with impaired pancreatic beta cell function and decreased insulin sensitivity: a population-based study. Diabetologia 58 523–533. (https://doi.org/10.1007/s00125-014-3469-4)
Chan CB, De Leo D, Joseph JW, McQuaid TS, Ha XF, Xu F, Tsushima RG, Pennefather PS, Salapatek AM & Wheeler MB 2001 Increased uncoupling protein-2 levels in beta-cells are associated with impaired glucose-stimulated insulin secretion: mechanism of action. Diabetes 50 1302–1310. (https://doi.org/10.2337/diabetes.50.6.1302)
Cheng K, Ho K, Stokes R, Scott C, Lau SM, Hawthorne WJ, O'Connell PJ, Loudovaris T, Kay TW, Kulkarni RN, et al. 2010 Hypoxia-inducible factor-1alpha regulates beta cell function in mouse and human islets. Journal of Clinical Investigation 120 2171–2183. (https://doi.org/10.1172/JCI35846)
Coffey R & Knutson MD 2017 The plasma membrane metal-ion transporter ZIP14 contributes to nontransferrin-bound iron uptake by human beta-cells. American Journal of Physiology-Cell Physiology 312 C169–C175. (https://doi.org/10.1152/ajpcell.00116.2016)
Cooksey RC, Jouihan HA, Ajioka RS, Hazel MW, Jones DL, Kushner JP & McClain DA 2004 Oxidative stress, beta-cell apoptosis, and decreased insulin secretory capacity in mouse models of hemochromatosis. Endocrinology 145 5305–5312. (https://doi.org/10.1210/en.2004-0392)
Cooksey RC, Jones D, Gabrielsen S, Huang J, Simcox JA, Luo B, Soesanto Y, Rienhoff H, Abel ED & McClain DA 2010 Dietary iron restriction or iron chelation protects from diabetes and loss of beta-cell function in the obese (ob/ob lep-/-) mouse. American Journal of Physiology-Endocrinology and Metabolism 298 E1236–E1243. (https://doi.org/10.1152/ajpendo.00022.2010)
Coughlan MT, Thorburn DR, Penfold SA, Laskowski A, Harcourt BE, Sourris KC, Tan AL, Fukami K, Thallas-Bonke V, Nawroth PP, et al. 2009 RAGE-induced cytosolic ROS promote mitochondrial superoxide generation in diabetes. Journal of the American Society of Nephrology 20 742–752. (https://doi.org/10.1681/ASN.2008050514)
Cunha DA, Hekerman P, Ladriere L, Bazarra-Castro A, Ortis F, Wakeham MC, Moore F, Rasschaert J, Cardozo AK, Bellomo E, et al. 2008 Initiation and execution of lipotoxic ER stress in pancreatic beta-cells. Journal of Cell Science 121 2308–2318. (https://doi.org/10.1242/jcs.026062)
Ehrnstorfer IA, Geertsma ER, Pardon E, Steyaert J & Dutzler R 2014 Crystal structure of a SLC11 (NRAMP) transporter reveals the basis for transition-metal ion transport. Nature Structural & Molecular Biology 21 990–996. (https://doi.org/10.1038/nsmb.2904)
Ellervik C, Mandrup-Poulsen T, Nordestgaard BG, Larsen LE, Appleyard M, Frandsen M, Petersen P, Schlichting P, Saermark T, Tybjaerg-Hansen A, et al. 2001 Prevalence of hereditary haemochromatosis in late-onset type 1 diabetes mellitus: a retrospective study. Lancet 358 1405–1409. (https://doi.org/10.1016/S0140-6736(01)06526-6)
Ellervik C, Mandrup-Poulsen T, Andersen HU, Tybjaerg-Hansen A, Frandsen M, Birgens H & Nordestgaard BG 2011 Elevated transferrin saturation and risk of diabetes: three population-based studies. Diabetes Care 34 2256–2258. (https://doi.org/10.2337/dc11-0416)
Ford ES & Cogswell ME 1999 Diabetes and serum ferritin concentration among U.S. adults. Diabetes Care 22 1978–1983. (https://doi.org/10.2337/diacare.22.12.1978)
Friberg J, Tonnesen MF, Heller S, Pociot F, Bodvarsdottir TB & Karlsen AE 2010 Inhibition of the nuclear factor-kappaB pathway prevents beta cell failure and diet induced diabetes in Psammomys obesus. PLoS ONE 5 e13341. (https://doi.org/10.1371/journal.pone.0013341)
Gabrielsen JS, Gao Y, Simcox JA, Huang J, Thorup D, Jones D, Cooksey RC, Gabrielsen D, Adams TD, Hunt SC, et al. 2012 Adipocyte iron regulates adiponectin and insulin sensitivity. Journal of Clinical Investigation 122 3529–3540. (https://doi.org/10.1172/JCI44421)
Hansen JB, Tonnesen MF, Madsen AN, Hagedorn PH, Friberg J, Grunnet LG, Heller RS, Nielsen AO, Storling J, Baeyens L, et al. 2012 Divalent metal transporter 1 regulates iron-mediated ROS and pancreatic beta cell fate in response to cytokines. Cell Metabolism 16 449–461. (https://doi.org/10.1016/j.cmet.2012.09.001)
Jehn M, Clark JM & Guallar E 2004 Serum ferritin and risk of the metabolic syndrome in U.S. adults. Diabetes Care 27 2422–2428. (https://doi.org/10.2337/diacare.27.10.2422)
Jiang R, Manson JE, Meigs JB, Ma J, Rifai N & Hu FB 2004 Body iron stores in relation to risk of type 2 diabetes in apparently healthy women. JAMA 291 711–717. (https://doi.org/10.1001/jama.291.6.711)
Jung IR, Choi SE, Jung JG, Lee SA, Han SJ, Kim HJ, Kim DJ, Lee KW & Kang Y 2015 Involvement of iron depletion in palmitate-induced lipotoxicity of beta cells. Molecular and Cellular Endocrinology 407 74–84. (https://doi.org/10.1016/j.mce.2015.03.007)
Leloup C, Tourrel-Cuzin C, Magnan C, Karaca M, Castel J, Carneiro L, Colombani AL, Ktorza A, Casteilla L & Penicaud L 2009 Mitochondrial reactive oxygen species are obligatory signals for glucose-induced insulin secretion. Diabetes 58 673–681. (https://doi.org/10.2337/db07-1056)
Lenzen S 2017 Chemistry and biology of reactive species with special reference to the antioxidative defence status in pancreatic beta-cells. Biochimica et Biophysica Acta 1861 1929–1942. (https://doi.org/10.1016/j.bbagen.2017.05.013)
Lenzen S, Drinkgern J & Tiedge M 1996 Low antioxidant enzyme gene expression in pancreatic islets compared with various other mouse tissues. Free Radical Biology and Medicine 20 463–466. (https://doi.org/10.1016/0891-5849(96)02051-5)
Lu JP & Hayashi K 1994 Selective iron deposition in pancreatic islet B cells of transfusional iron-overloaded autopsy cases. Pathology International 44 194–199. (https://doi.org/10.1111/j.1440-1827.1994.tb02592.x)
Lu JP, Hayashi K, Okada S & Awai M 1991 Transferrin receptors and selective iron deposition in pancreatic B cells of iron-overloaded rats. Acta Pathologica Japonica 41 647–652.
Maedler K, Schulthess FT, Bielman C, Berney T, Bonny C, Prentki M, Donath MY & Roduit R 2008 Glucose and leptin induce apoptosis in human beta-cells and impair glucose-stimulated insulin secretion through activation of c-Jun N-terminal kinases. FASEB Journal 22 1905–1913. (https://doi.org/10.1096/fj.07-101824)
Montonen J, Boeing H, Steffen A, Lehmann R, Fritsche A, Joost HG, Schulze MB & Pischon T 2012 Body iron stores and risk of type 2 diabetes: results from the European Prospective Investigation into Cancer and Nutrition (EPIC)-Potsdam study. Diabetologia 55 2613–2621. (https://doi.org/10.1007/s00125-012-2633-y)
Munoz P, Humeres A, Elgueta C, Kirkwood A, Hidalgo C & Nunez MT 2011 Iron mediates N-methyl-D-aspartate receptor-dependent stimulation of calcium-induced pathways and hippocampal synaptic plasticity. The Journal of Biological Chemistry 286 13382–13392. (https://doi.org/10.1074/jbc.M110.213785)
Muoio DM & Newgard CB 2008 Mechanisms of disease:Molecular and metabolic mechanisms of insulin resistance and beta-cell failure in type 2 diabetes. Nature Reviews Molecular Cell Biology 9 193–205. (https://doi.org/10.1038/nrm2327)
Nielsen JH 1989 Mechanisms of pancreatic beta-cell growth and regeneration: studies on rat insulinoma cells. Experimental and Clinical Endocrinology 93 277–285. (https://doi.org/10.1055/s-0029-1210869)
Prentice KJ, Luu L, Allister EM, Liu Y, Jun LS, Sloop KW, Hardy AB, Wei L, Jia W, Fantus IG, et al. 2014 The furan fatty acid metabolite CMPF is elevated in diabetes and induces beta cell dysfunction. Cell Metabolism 19 653–666. (https://doi.org/10.1016/j.cmet.2014.03.008)
Rajpathak SN, Crandall JP, Wylie-Rosett J, Kabat GC, Rohan TE & Hu FB 2009 The role of iron in type 2 diabetes in humans. Biochimica et Biophysica Acta 1790 671–681. (https://doi.org/10.1016/j.bbagen.2008.04.005)
Roma LP, Pascal SM, Duprez J & Jonas JC 2012 Mitochondrial oxidative stress contributes differently to rat pancreatic islet cell apoptosis and insulin secretory defects after prolonged culture in a low non-stimulating glucose concentration. Diabetologia 55 2226–2237. (https://doi.org/10.1007/s00125-012-2581-6)
Thomas C, Mackey MM, Diaz AA & Cox DP 2009 Hydroxyl radical is produced via the Fenton reaction in submitochondrial particles under oxidative stress: implications for diseases associated with iron accumulation. Redox Report 14 102–108. (https://doi.org/10.1179/135100009X392566)