lable at ScienceDirect

Nutrition 26 (2010) 1146–1150

Contents lists avai

Nutrition

journal homepage: www.nutr i t ionjrnl .com

Basic nutritional investigation

Hyperglycemia-induced endoplasmic reticulum stress in endothelial cells

Mae Sheikh-Ali M.D. *, Senan Sultan M.D., Abdul-Razzak Alamir M.D., Michael J. Haas Ph.D.,Arshag D. Mooradian M.D.Division of Endocrinology, Diabetes and Metabolism, Department of Medicine, University of Florida, Jacksonville, Florida, USA

a r t i c l e i n f o

Article history:Received 8 June 2009Accepted 15 August 2009

Keywords:Endoplasmic reticulumOxidative stressHyperglycemiaMitochondria

* Corresponding author. Tel.: þ904-244-3702; fax:E-mail address: mae.sheikh-ali@jax.ufl.edu (M. Sh

0899-9007/$ – see front matter Published by Elseviedoi:10.1016/j.nut.2009.08.019

a b s t r a c t

Objective: Hyperglycemia-induced endothelial cell dysfunction in vascular disease can occur due toincreased oxidative stress and a concomitant increase in endoplasmic reticulum (ER) stress. Toinvestigate whether these cellular stresses are independent or causally linked, we determinedwhether or not specific glycolytic intermediates that induce oxidative stress also induce ER stress.Methods: Human umbilical vein endothelial cells were treated with dextrose, partially metabo-lizable (e.g., fructose and galactose) and non-metabolizable sugars (e.g., 3-O-methyglucose), andvarious intermediates of the glycolytic and tricarboxylic acid pathways. Activation of the unfoldedprotein response and subsequent generation of ER stress was measured by the ER stress-responsivealkaline phosphatase method, and superoxide (SO) generation was measured using the hydro-ethidene–fluorescence method. The mitochondrial origin of the SO and the generation of ER stressby dextrose and the intermediate metabolites were confirmed with experiments using allopurinoland diphenyleneiodonium chloride to block SO generation by xanthine oxidase and nicotinamideadenosine dinucleotide phosphate oxidase, respectively.Results: Although ER stress could be induced by glycolytic intermediates up to and includingpyruvate, the SO generation occurred in the presence of glycolytic and mitochondrial metabolites.Conclusion: Although the mitochondria are the site of signals generated by dextrose to initiateoxidative stress, the dextrose-induced ER stress, unlike SO generation, does not require pyruvateoxidation in the mitochondria.

Published by Elsevier Inc.

Introduction

The endoplasmic reticulum (ER) is an organelle that can senseand respond to various stresses related to inhibition of proteinsynthesis and can transmit apoptotic signals [1]. The ER isa cellular calcium store and is responsible for post-translationalmodification, folding, and assembly of newly synthesizedsecretory and membrane-bound proteins [2]. ER dysfunction dueto activation of the unfolded protein response is commonlyreferred to as ‘‘ER stress’’ [3–5] and is implicated in the endo-thelial cell dysfunction observed in type 2 diabetes and obesity[6]. Although ER relies on an oxidizing environment to achievestructural–functional changes of the protein machinery involvedin the redox folding process, in diabetes a reducing environmentis present in the ER that presumably alters protein syntheticcapacity [7,8]. In addition, ascorbate/dihydro-ascorbate redox

þ904-244-5650.eikh-Ali).

r Inc.

coupling is implicated in oxidative folding of proteins, whereasvarious antioxidants (e.g., tocopherol, ubiquinone, and vitamin K)may regulate the electron transfer chain responsible for oxida-tive protein folding in the ER [9].

Hyperglycemia-induced endothelial cell dysfunction canoccur secondary to increased oxidative stress and a concomi-tant increase in ER stress. These processes have been shown tobe linked because oxidative stress can induce ER stress [8,9].However, it is not clear if the two stress responses are causallylinked or are two independent effects of hyperglycemia. Todetermine if the hyperglycemia-related ER stress and oxidativestress signals originate in the cytoplasm or within the mito-chondria, activation of the unfolded protein response andsuperoxide (SO) generation were measured in human umbil-ical vein endothelial cells (HUVECs) treated with dextrose,partially metabolizable (e.g., fructose and galactose) and nonmetabolizable sugars (e.g., 3-O-methyglucose) and variousmetabolites of the glycolytic pathway and the tricarboxylicacid cycle.

M. Sheikh-Ali et al. / Nutrition 26 (2010) 1146–1150 1147

Materials and methods

Cell cultures

Human umbilical endothelial cells (Cell Applications, Inc., San Diego, CA,USA) were maintained in T-75 flasks coated with endothelial cell attachmentfactor (Cell Applications, Inc.) in complete endothelial cell growth medium (CellApplications, Inc.). Cells were maintained in a humidified incubator at 37 �C and5% CO2.

For comparison, some studies were also carried out in human hepatoma-derived HepG2 cells. These cells were maintained in Dulbecco’s Modified Eagle’sMedium containing 5% fetal bovine serum, penicillin (100 U/mL), and strepto-mycin (100 mg/mL) and were housed in a humidified environment at 37 �C and 5%CO2.

Superoxide detection assay

Superoxide generation was measured using a hydro-ethidene (HE) (Invi-trogen, Carlsbad, CA) fluorescence method [10,11]. HE is freely permeable to cellsand is subject to oxidation by intracellular SO to form a fluorescent endproduct.For each experiment, HUVECs were released with trypsin and transferred to 96-well black, clear-bottom fluorescence assay microplates (Corning, Inc., Corning,NY, USA) at a density of 50 000 cells per well. Twenty-four hours after plating, thecells were washed with sterile Hank’s Balanced Salt Solution containing 1.26 mMCaCl2, 5.37 mM KCl, 0.44 mM KH2PO4, 0.49 mM Mg Cl2 $ 6 H2O, 0.41 mM MgSO4 $

7 H2O, 136.7 mM NaCl, 4.2 mM NaHCO3, 0.34 mM Na2HPO4, and 5.5 mM D-glucose. Hydro-ethidene dissolved in Hank’s Balanced Salt Solution was added toa final concentration of 10 mmol/L. Fluorescence was measured every 10 min fora total of 60 min using a microplate fluorescence reader with excitation at488 nm and emission at 610 nm. The rate of SO generation was calculated fromthe rate of appearance of the oxidized HE within the first hour of adding theindicator HE. Each experiment was repeated six times.

To determine if xanthine oxidase and nicotinamide adenosine dinucleotidephosphate (NADPH) oxidase contribute to SO generation in HUVECs exposed todextrose and the glycolytic and mitochondrial metabolites, HUVECs were treatedwith dextrose (5.5 and 27.5 mM) or 20 mM pyruvate, 20 mM dihydroxyacetone,7 mM citrate, or 5 mM a-ketoisocaproate and SO generation was measured(Fig. 1) in the presence or absence of 5 mM allopurinol (AP; to inhibit xanthineoxidase) and 10 mM di-phenylene-iodonium chloride (DPI; to inhibit NADPHoxidase). AP and DPI were obtained from Sigma (St. Louis, MO, USA). Theconcentrations of AP and DPI used were selected based on a previous study usingendothelial cells [12].

ER stress measurements

The ER stress was measured using the ER stress-responsive placental alkalinephosphatase assay [13] and the Great EscAPe kit from Clontech Laboratories, Inc.(Mountain View, CA, USA). Measurement of ER stress using this assay has beenshown to correlate closely with changes of markers of the unfolded proteinresponse, including glucose-regulated proteins 78 and 94, in renal proximaltubule cells, hepatocytes, and alveolar macrophage cells in response to treatmentwith tunicamycin and thapsigargin, and in ER stress-responsive placental alka-line phosphatase transgenic mice exposed to lipopolysaccharide [13]. Briefly,HUVECs at 70–80% confluence in six-well plates were transfected with theplasmid pSEAP2-Control, containing a truncated placental alkaline phosphatasegene adjacent to the simian virus 40 early gene promoter [14], and 24 h latertreated with select carbohydrates (5.5 and 27.5 mM glucose, 27.5 mM 3-O-methylglucose, 27.5 mM galactose, and 27.5 mM fructose) and metabolicintermediates (20 mM acetate, 20 mM pyruvate, 20 mM glycerol, 20 mMdi-hydroxy-acetone, 5 mM a-ketoisocaproate, and 7 mM citrate). After 24 h,secreted alkaline phosphatase (SAP) activity was measured by mixing 25 mL ofconditioned medium with 25 mL of dilution buffer and incubating at 65 �C for30 min in a tightly covered black-walled, clear-bottom 96-well microtiter plate.After cooling on ice for 3 min, 97 mL of assay buffer and 3 mL of 1 mM 4-meth-ylumbelliferyl phosphate were added to each well. The samples were incubatedat room temperature for 60 min and SAP activity was measured with a fluores-cence spectrophotometer (360-nm excitation and 449-nm emission). As a nega-tive control, SAP activity was measured in cells transfected with pSEAP2-Basic,which contains the placental SAP gene but lacks a eukaryotic promoter to driveexpression.

Statistical analysis

All results are expressed as mean � standard deviation. Analysis of variancefollowed by the Neuman-Keuls procedure for subgroup analysis was carried outusing Statistica for Windows (Statsoft, Inc., Tulsa, OK, USA). Significance wasdefined as a two-tailed P value <0.05.

Results

Oxidative stress and ER stress were measured in HUVECs inthe presence of select carbohydrates and substrates of interme-diary metabolism as indicated in Figure 1. Fructose, galactose,and 3-O-methylglucose were poorly metabolized or poorsubstrates for the glucose transporters. Glycerol and dihy-droxyacetone were glycolytic substrates, whereas pyruvateacetate, citrate, and a-ketoisocaproate were mitochondrialsubstrates (Fig. 1).

Treatment of HUVEC with high concentrations of dextrose andfructose caused significant ER stress (Fig. 2). The meanSAP activity was reduced from 0.94� 0.11 relative light units in5.5-mM dextrose-treated control cells to 0.63� 0.03 in cellstreated with 27.5 mM dextrose (P< 0.004) and 0.656� 0.12 incells treated with 27.5 mM fructose (P< 0.002). However, SAPactivity in cells treated with 27.5 mM 3-O-methyl glucose(1.04� 0.10) or 27.5 mM galactose (0.92� 0.04) was not signifi-cantly different from control cells. The SAP activity in cells treatedwith 20 mM pyruvate (0.46� 0.08), 20 mM glycerol (0.49� 0.08),and 20 mM dihydroxyacetone (0.46� 0.05) was significantlyreduced (P< 0.003, P< 0.003, P< 0.002, respectively). Incontrast, SAP activity in cell cultures treated with 20 mM acetate(0.87� 0.05), 5 mM a-ketoisocaproate (0.79� 0.08), or 7 mMcitrate (0.95� 0.07) was not significantly reduced and was similarto control cells.

The ability of dextrose to induce ER stress is not unique toendothelial cells as similar changes were observed in HepG2cells. In these cells, treatment with 27.5 mM dextrose reducedthe SAP reporter gene activity to 0.572� 0.075 compared withcells treated with 5.5 mM dextrose (0.907� 0.092), whereastreatment with dimethylsulfoxide, an ER stress inhibitor,normalized SAP activity to 0.991�0.051.

The effects of select carbohydrates and substrates of inter-mediary metabolism on SO generation are shown in Figure 3.Generation of SO was increased significantly in cells treated with27.5 mm dextrose compared with control cells treated with5.5 mm dextrose (135.0� 8.7 versus 61.0� 3.6, P< 0.0002).Although SO generation did not increase in cells treated with27.5 mM 3-O-methyl glucose (61.3� 5.7) or 27.5 mM galactose(56.3�7.1), it was significantly elevated in cells treated with27.5 mM fructose-treated cells (83.3� 5.0, P< 0.003). Genera-tion of SO increased significantly in cells treated with all theintermediary metabolites tested including 20 mM acetate(106.7�14.2), 20 mM pyruvate (124.7�14.2), 20 mM glycerol(100.7�10.0), 20 mM dihydroxyacetone (124.0�10.8), 5 mM a-ketoisocaproate (101.7�14.9), and 7 mM citrate (110.3�7.6,P< 0.006, P< 0.002, P< 0.003, P< 0.0007, P< 0.0, andP< 0.0005; respectively).

The effect of xanthine oxidase and NADPH oxidase inhibitionon SO generation is shown in Figure 4A. In the absence of anypharmacologic inhibition, the SO generation increased in cellsexposed to 27.5 mM dextrose (0.163� 0.009) compared withcells exposed to 5.5 mM dextrose (0.099� 0.011, P< 0.002).Addition of AP/DPI had no effect on SO generation in cellsexposed to 27.5 mM dextrose (0.166� 0.011) compared withcells exposed to 5.5 mM dextrose (0.089� 0.007, P< 0.0004).Likewise, in cells treated with pyruvate, dihydroxyacetone,citrate, or a-ketoisocaproate, AP/DPI addition had no effect on SOgeneration. In cells treated with pyruvate, SO generation was0.157� 0.014 compared with 0.162� 0.014 in cells treated withpyruvate and AP/DPI (not significant). In cells treated withdihydroxyacetone, SO generation was 0.162� 0.009 comparedwith 0.170� 0.015 in cells treated with dihydroxyacetone and

Glucose

Glucose-6-P

Fructose-6-P

Fructose-1,6-bis-P

Triose Phosphates

Phosphoenolpyruvate

Pyruvate

Acetyl CoA

Oxaloacetate Citrate

Succinate Succinyl CoA

Fumarate -Ketoglutarate

Malate Isocitrate

Fructose

3-O-Methylglucose

Galactose

Acetate

Glycerol Dihydroxyacetone

Citrate

-Ketoisocaproate

Cytosol

Mitochondria

Extracellular Space

Pyruvate

Pyruvate

Fig. 1. Select carbohydrates and substrates of intermediary metabolism. Human umbilical vein endothelial cells were treated with the indicated metabolites (bold) that feedinto the glycolytic and tricarboxylic acid cycles as indicated, and oxidative stress and endoplasmic reticulum stress were measured. CoA, coenzyme A.

M. Sheikh-Ali et al. / Nutrition 26 (2010) 1146–11501148

AP/DPI (not significant). In citrate treated cells, SO generationwas 0.171�0.013 compared with 0.165� 0.009 in cells treatedwith citrate and AP/DPI (not significant). In cells treated with a-ketoisocaproate, SO generation was 0.170� 0.014 compared with0.164� 0.008 AU in cells treated with a-ketoisocaproate and AP/DPI (not significant). These results suggest that in dextrose-treated endothelial cells, xanthine oxidase and NADPH oxidasedo not contribute significantly to SO generation.

To determine if xanthine oxidase and NADPH oxidasecontribute to ER stress in HUVECs exposed to dextrose andthe glycolytic and mitochondrial metabolites, HUVECs weretreated with dextrose, pyruvate, dihydroxyacetone, citrate, anda-ketoisocaproate in the presence or absence of AP and DPI asdescribed above, and ER stress was measured (Fig. 4B). Activity ofthe SAP reporter decreased in cells exposed to 27.5 mM dextrose(0.703� 0.057) compared with cells exposed to 5.5 mM dextrose(0.929� 0.077, P< 0.02). Addition of AP/DPI had no effect on SAP

reporter gene activity in cells exposed to 27.5 mM dextrose(0.656� 0.055) compared with cells exposed to 5.5 mM dextrose(0.982� 0.081, P< 0.005). Similar to SO generation, in cellstreated with pyruvate, dihydroxyacetone, citrate, and a-ketoi-socaproate, AP/DPI addition had no effect on ER stress (Fig. 4B).These results suggest that in dextrose-treated endothelial cells,xanthine oxidase and NADPH oxidase do not contribute signifi-cantly to ER stress generation.

Discussion

The results clearly indicate that ER stress is elevated inHUVECs treated with 27.5 mM dextrose and fructose but not incells treated with poorly metabolizable galactose or the non-metabolizable glucose analog 3-O-methylglucose. The latterobservation indicates that the increase in ER stress is indepen-dent of the osmotic effects of 27.5 mM dextrose.

0

0.2

0.4

0.6

0.8

1.0

1.2

SA

P A

ctivity

(A

.U

.)

SAP - + + + + + + + + + + +

5.5 27.5

Dextrose

3-M

e-G

lc

Galacto

se

Fru

cto

se

Acetate

Pyru

va

te

Gly

ce

ro

l

DH

A

-K

eto

-IC

Citrate

* **

**

Fig. 2. Effect of select carbohydrates and intermediary metabolism substrates onendoplasmic reticulum stress. Human umbilical vein endothelial cells were exposedto each substrate and SAP activity was measured after 24 h and expressed inarbitrary units. Hyperglycemia (glucose 500 mg/dL) induced endoplasmicreticulum stress compared with euglycemia (glucose 100 mg/dL). Galactose,3-O-methylglucose, and the mitochondrial substrates acetate, a-ketoisocaproate,and citrate had no effect on endoplasmic reticulum stress, whereas fructose and theglycolytic metabolites pyruvate, glycerol, and dihydroxyacetone increased endo-plasmic reticulum stress (n¼ 6, * P< 0.02). a-Keto-IC, a-ketoisocaproate; AU, arbi-trary units; DHA, dihydroxyacetone; 3-Me-Glc, 3-O-methylglucose; SAP þ, cellstransfected with pSEAP2-Control; SAP �, cells transfected with pSEAP2-Basic(negative control); SAP, secreted alkaline phosphatase.

SAP - + + + + + + + + + + + +

Dex 5.5 5.5 27.5 5.5 27.5 - - - - - - - -

AP/DPI - - - + + - + - + - + - +

0

0.04

0.08

0.12

0.16

0.2S

O G

en

era

tio

n

(A

.U

.)

Pyr DHA Cit KIC

* * * * **** **A

1.0

1.2B

M. Sheikh-Ali et al. / Nutrition 26 (2010) 1146–1150 1149

Pyruvate, glycerol, and dihydroxyacetone induced ER stresssimilar to glucose, whereas compounds that enter themetabolic scheme distal to pyruvate oxidation such as acetate,a-ketoisocaproate, and citrate were without effect.

As reported previously, the high ambient dextrose concen-trations also increase the SO generation (Fig. 3) [11]. In theseexperiments, fructose was also a potent SO generator, whereasgalactose and the non-metabolizable 3-O-methyl glucose werenot. The latter observation indicates that glucose metabolism isa prerequisite for generating intracellular SO, and that thechanges in SO generation are independent of the osmotic effectsof 27.5 mM dextrose. These results indicate that an ER stress-related signal is generated by hyperglycemia in endothelial cellsand, unlike SO generation, it does not require pyruvate oxidationin the mitochondria.

The effect of fructose on SO generation could be secondaryto its partial metabolism, cytotoxicity, or depletion of adenosine

0

20

40

60

80

100

120

140

160

SO

G

en

eratio

n

(A

.U

.)

SAP - + + + + + + + + + + +

5.5 27.5

Dextrose

3-M

e-G

lc

Galacto

se

Fru

cto

se

Acetate

Pyru

va

te

Gly

ce

ro

l

DH

A

-K

eto

-IC

Citrate

*

**

**

** *

Fig. 3. Effect of select carbohydrates and intermediary metabolism substrates onSO generation. Human umbilical vein endothelial cells were exposed to eachsubstrate and SO generation was measured after 1 h and expressed in arbitraryunits. Hyperglycemia (dextrose 27.5 mM) induced SO compared with euglycemia(dextrose 5.5 mM). Galactose and 3-O methylglucose had no effect on SO genera-tion, whereas fructose and the glycolytic (pyruvate, glycerol, and dihydroxyacetone)and mitochondrial (acetate, a-ketoisocaproate and citrate) substrates hadsubstantial SO-generating capacity (n¼ 6, * P< 0.01). a-Keto-IC, a-ketoisocaproate;AU, arbitrary units; DHA, dihydroxyacetone; 3-Me-Glc, 3-O-methylglucose; SO,superoxide; SAP þ, cells transfected with pSEAP2-Control; SAP �, cells transfectedwith pSEAP2-Basic (negative control); SAP, secreted alkaline phosphatase.

triphosphate. It could also be a reflection of its pro-oxidantproperties, as demonstrated previously in cell-free assays wherefructose, unlike galactose, was a potent pro-oxidant [15].

Although the source of glucose-induced SO generation isprobably the mitochondria [12], the source of the SO induced bythe other metabolites of the glycolysis could be mitochondrial orother sites including xanthine oxidase and NADPH oxidase.Therefore, additional experiments were carried out in the pres-ence of pharmacologic inhibitors of these extramitochondrialsources of SO generation. The results showed that dextrose andother substrates of intermediary metabolism induced significantSO generation and ER stress and inhibition of xanthine oxidaseand NADPH oxidase had no effect on this response (Fig. 4).

A potential limitation of these data is that some of the metab-olites used to induce SO generation such as pyruvate and all othera-keto acids have potent antioxidant activity. This is in agreementwith the notion that many compounds including simple carbo-hydrates and ascorbic acid have oxidative and antioxidativeproperties depending on the cellular milieu [15].

0

0.2

0.4

0.6

0.8

SA

P A

ctivity

(A

.U

.)

SAP - + + + + + + + + + + + +

Dex 5.5 5.5 27.5 5.5 27.5 - - - - - - - -

AP/DPI - - - + + - + - + - + - +

Pyr DHA Cit KIC

* * *** *

Fig. 4. Effect of xanthine oxidase and nicotinamide adenosine dinucleotide phos-phate oxidase inhibition on SO generation and endoplasmic reticulum stress. (A) SOgeneration was measured in human umbilical vein endothelial cells treated withDex and metabolites Pyr, DHA, Cit, and aKIC with or without AP/DPI. Levels of SOare expressed in arbitrary units (n¼ 6, * P< 0.05 versus cells treated with 5.5 mMDex). (B) Reporter gene (SAP) activity was measured in human umbilical veinendothelial cells treated with Dex and the metabolites Pyr, DHA, Cit, and aKIC, withor without AP/DPI. Levels of SAP reporter gene activity are expressed in arbitraryunits (n¼ 6, * P< 0.05 versus cells treated with 5.5 mM Dex). aKIC, a-ketoisocap-roate; AP, allopurinol; AU, arbitrary units; Cit, citrate; Dex, dextrose; DHA, dihy-droxyacetone; DPI, diphenyleneiodonium chloride; Pyr, pyruvate; SAP, secretedalkaline phosphatase; SO, superoxide.

M. Sheikh-Ali et al. / Nutrition 26 (2010) 1146–11501150

Conclusion

Overall these experiments indicate that hyperglycemia-inducedoxidative and ER stresses in endothelial cells are initiated throughdifferent stress-inducing signaling metabolites of glucose metab-olism. Identification of the precise underlying signaling cascadethat induces ER stress may help in the development of futuretherapeutic modalities that prevent diabetes and its complications.

References

[1] Araki E, Oyadomari S, Mori M. Impact of endoplasmic reticulum stresspathway on pancreatic b-cells and diabetes mellitus. Exp Biol Med2003;228:1213–7.

[2] Kopito RR. Aggresomes, inclusion bodies and protein aggregation. TrendsCell Biol 2000;10:524–30.

[3] Mori K. Tripartite management of unfolded proteins in the endoplasmicreticulum. Cell 2000;101:451–4.

[4] Kaufman RJ, Scheuner D, Schroder M, Shen X, Lee K, Liu CY, Arnold SM. Theunfolded protein response in nutrient sensing and differentiation. Nat RevMol Cell Biol 2002;3:411–21.

[5] Oyadomari S, Araki E, Mori M. Endoplasmic reticulum stress-mediatedapoptosis in pancreatic b-cells. Apoptosis 2002;7:335–45.

[6] Tsiotra PC, Tsigos C. Stress, the endoplasmic reticulum, and insulin resis-tance. Ann N Y Acad Sci 2006;1083:63–76.

[7] Nardai G, Korcsmaros T, Papp E, Csermely P. Reduction of the endoplasmicreticulum accompanies the oxidative damage of diabetes mellitus.Biofactors 2003;17:259–67.

[8] Tu BP, Weissman JS. Oxidative protein folding in eukaryotes: mechanismsand consequences. J Cell Biol 2004;164:341–6.

[9] Banhegyi G, Csala M, Szarka A, Varsanvi M, Benedetti A, Mandl J.Role of ascorbate in oxidative protein folding. Biofactors 2003;17:37–46.

[10] Carter WO, Narayanan PK, Robinson JP. Intracellular hydrogen peroxideanion detection in endothelial cells. J Leukoc Biol 1994;55:253–8.

[11] Horani MH, Haas MJ, Mooradian AD. Rapid adaptive down regulation ofoxidative burst induced by high dextrose in human umbilical vein endo-thelial cells. Diabetes Res Clin Pract 2004;66:7–12.

[12] Nishikawa T, Edelstein D, Du XL, Yamagishi S, Matsumura T, Kaneda Y, et al.Normalizing mitochondrial superoxide production blocks three pathwaysof hyperglycaemic damage. Nature 2000;404:787–90.

[13] Hirmatsu N, Kasai A, Hayakawa K, Yao J, Kitamura M. Real-time detectionand continuous monitoring of ER stress in vitro and in vivo by ES-TRAP:evidence for systemic, transient ER stress during endotoxemia. Nucl AcidsRes 2006;34:e93.

[14] Berger J, Hauber J, Hauber R, Geiger R, Cullen BR. Secreted placental alka-line phosphatase: a powerful new quantitative indicator of gene expressionin eukaryotic cells. Gene 1988;66:1–10.

[15] Wehmeier KR, Mooradian AD. Autoxidative and antioxidative potential ofsimple carbohydrates. Free Radic Biol Med 1994;17:83–6.