Portal de Periódicos da CAPES

High Glucose Inhibits Glucose-6-phosphate Dehydrogenase via cAMP in Aortic Endothelial Cells

2000; Elsevier BV; Volume: 275; Issue: 51 Linguagem: Inglês

10.1074/jbc.m007505200

1083-351X

Zhiquan Zhang, Kira Apse, Jiongdong Pang, Robert C. Stanton,

Hyperglycemia and glycemic control in critically ill and hospitalized patients

Recent studies have shown that hyperglycemia is a principal cause of cellular damage in patients with diabetes mellitus. A major consequence of hyperglycemia is increased oxidative stress. Glucose-6-phosphate dehydrogenase (G6PD) plays an essential role in the regulation of oxidative stress by primarily regulating NADPH, the main intracellular reductant. In this paper we show that increased glucose (10–25 mm) caused inhibition of G6PD resulting in decreased NADPH levels in bovine aortic endothelial cells (BAEC). Inhibition was seen within 15 min. High glucose-induced inhibition of G6PD predisposed cells to cell death. High glucose via increased activity of adenylate cyclase also stimulated an increase in cAMP levels in BAEC. Agents that increased cAMP caused a decrease in G6PD activity. Inhibition of cAMP-dependent protein kinase A ameliorated the high glucose-induced inhibition of G6PD. Finally, high glucose stimulated phosphorylation of G6PD. These results suggest that, in BAEC, high glucose stimulated increased cAMP, which led to increased protein kinase A activity, phosphorylation of G6PD, and inhibition of G6PD activity. We conclude that these changes in G6PD activity play an important role in high glucose-induced cell damage/death. Recent studies have shown that hyperglycemia is a principal cause of cellular damage in patients with diabetes mellitus. A major consequence of hyperglycemia is increased oxidative stress. Glucose-6-phosphate dehydrogenase (G6PD) plays an essential role in the regulation of oxidative stress by primarily regulating NADPH, the main intracellular reductant. In this paper we show that increased glucose (10–25 mm) caused inhibition of G6PD resulting in decreased NADPH levels in bovine aortic endothelial cells (BAEC). Inhibition was seen within 15 min. High glucose-induced inhibition of G6PD predisposed cells to cell death. High glucose via increased activity of adenylate cyclase also stimulated an increase in cAMP levels in BAEC. Agents that increased cAMP caused a decrease in G6PD activity. Inhibition of cAMP-dependent protein kinase A ameliorated the high glucose-induced inhibition of G6PD. Finally, high glucose stimulated phosphorylation of G6PD. These results suggest that, in BAEC, high glucose stimulated increased cAMP, which led to increased protein kinase A activity, phosphorylation of G6PD, and inhibition of G6PD activity. We conclude that these changes in G6PD activity play an important role in high glucose-induced cell damage/death. High glucose inhibits glucose-6-phosphate dehydrogenase via cAMP in aortic endothelial cells.Journal of Biological ChemistryVol. 276Issue 7PreviewPage 40046, legend to Fig. 5 A:The first sentence should read "the PKA inhibitor H89 causes a stimulation of G6PD" (not "an inhibition of G6PD"). The figure and its correct legend are shown below. Full-Text PDF Open Access glucose-6-phosphate dehydrogenase bovine aortic endothelial cells Dulbecco's modified Eagle's medium reactive oxygen species protein kinase A dehydroepiandrosterone Diabetes mellitus can lead to complications in the eye, heart, kidney, and nerves. Hyperglycemia has been shown to play a principal role in the pathogenesis of diabetic complications in both type 1 and type 2 diabetes mellitus (1UK Prospective Diabetes GroupLancet. 1998; 352: 837-853Abstract Full Text Full Text PDF PubMed Scopus (19145) Google Scholar, 2The Diabetes Control and Complications Trial Research GroupN. Engl. J. Med. 1993; 329: 977-986Crossref PubMed Scopus (22937) Google Scholar). Thus it is critical to understand how hyperglycemia leads to cellular damage. Hyperglycemia-induced increases in oxidative stress have been suggested to be of central pathogenic importance (3Low P.A. Nickander K.K. Tritschler H.J. Diabetes. 1997; 46 (suppl.): 38-42Crossref Google Scholar, 4Giugliano D. Ceriello A. Paolisso G. Diabetes Care. 1996; 19: 257-267Crossref PubMed Scopus (1705) Google Scholar). An increase in oxidative stress may be due to an increase in processes that produce oxidants or due to a decrease in antioxidant defenses. Glucose-6-phosphate dehydrogenase (G6PD),1 the rate-limiting enzyme of the pentose phosphate pathway, is required for the antioxidant defense system because it produces NADPH, the cells' principal reductant (5Tian W.-N. Braunstein L.D. Pang J. Stuhlmeier K. Xi Q.-C. Tian X. Stanton R.C. J. Biol. Chem. 1998; 273: 10609-10617Abstract Full Text Full Text PDF PubMed Scopus (402) Google Scholar, 6Tian W.-N. Braunstein L.D. Apse K. Pang J. Rose M. Tian X. Stanton R.C. Am. J. Physiol. 1999; 276: C1121-C1131Crossref PubMed Google Scholar). Previous research from our laboratory has shown that G6PD, an enzyme traditionally thought to be under little post-translational control, is regulated both in its activity and intracellular location by specific signal transduction molecules (7Tian W.-N. Pignatare J.N. Stanton R.C. J. Biol. Chem. 1994; 269: 14798-14805Abstract Full Text PDF PubMed Google Scholar, 8Stanton R.C. Seifter J.L. Boxer D.C. Zimmerman E. Cantley L.C. J. Biol. Chem. 1991; 266: 12442-12448Abstract Full Text PDF PubMed Google Scholar, 9Stanton R.C. Seifter J.L. Am. J. Physiol. 1988; 254: C267-C271Crossref PubMed Google Scholar). Importantly, our laboratory has also shown that modest changes in G6PD activity itself have significant effects on cell growth and cell death in a variety of cell types (5Tian W.-N. Braunstein L.D. Pang J. Stuhlmeier K. Xi Q.-C. Tian X. Stanton R.C. J. Biol. Chem. 1998; 273: 10609-10617Abstract Full Text Full Text PDF PubMed Scopus (402) Google Scholar, 6Tian W.-N. Braunstein L.D. Apse K. Pang J. Rose M. Tian X. Stanton R.C. Am. J. Physiol. 1999; 276: C1121-C1131Crossref PubMed Google Scholar). Thus proper activity of G6PD is required for adequate defense against oxidative stress and prevention of cell damage/death. Previous researchers have shown that the pentose phosphate pathway (as measured by release of radioactively labeled CO2) was impaired after exposure to high glucose. Asahina et al. (10Asahina T. Kashiwagi A. Nishio Y. Ikebuchi M. Harada N. Tanaka Y. Takagi Y. Saeki Y. Kikkawa R. Shigeta Y. Diabetes. 1995; 44: 520-526Crossref PubMed Scopus (105) Google Scholar) showed that when human umbilical endothelial cells in high glucose (33 mm) were exposed to H2O2 the expected increase in pentose phosphate pathway activity was almost completely inhibited as compared with cells incubated in normal glucose conditions (5 mm). The data suggested that high glucose predisposes cells to oxidative damage because of inadequate activation of the pentose phosphate pathway after exposure to an oxidant. G6PD per se was not evaluated in this study, nor was the mechanism of high glucose effect on G6PD reported. Thus we hypothesized that under high glucose conditions impaired activity of G6PD would predispose cells to oxidant damage and cell death. The data reported in this paper support the previous studies (10Asahina T. Kashiwagi A. Nishio Y. Ikebuchi M. Harada N. Tanaka Y. Takagi Y. Saeki Y. Kikkawa R. Shigeta Y. Diabetes. 1995; 44: 520-526Crossref PubMed Scopus (105) Google Scholar) and provide mechanistic information by showing that inhibition of G6PD occurs at least in part by high glucose-stimulated increases in cAMP levels. Chemicals were obtained from Sigma. G6PD antibody was a generous gift from Rolf Kletzien (Upjohn). Bovine aortic endothelial cells (BAEC) were grown in DMEM + 10% calf serum. Cells were used between passages 5 and 10. Cells were harvested by scraping and then lysed in a buffer containing phosphate-buffered saline + 5% Nonidet P-40, and in protease inhibitors. G6PD activity of the lysates was measured as described previously except that substrate concentrations were 500 μm glucose 6-phosphate and 250 μm NADP+ (7Tian W.-N. Pignatare J.N. Stanton R.C. J. Biol. Chem. 1994; 269: 14798-14805Abstract Full Text PDF PubMed Google Scholar). Cell death was measured by trypan blue exclusion as has been done previously (6Tian W.-N. Braunstein L.D. Apse K. Pang J. Rose M. Tian X. Stanton R.C. Am. J. Physiol. 1999; 276: C1121-C1131Crossref PubMed Google Scholar). Apoptosis was measured by 4′,6′-diamidine-2′-phenylindole dihydrochloride as described previously (6Tian W.-N. Braunstein L.D. Apse K. Pang J. Rose M. Tian X. Stanton R.C. Am. J. Physiol. 1999; 276: C1121-C1131Crossref PubMed Google Scholar). NADPH was measured by a new method designed by us (11Zhang Z., Yu, J. Stanton R.C. Anal. Biochem. 2000; 285: 163-167Crossref PubMed Scopus (98) Google Scholar). In brief, the method is based on the fact that only NADH and NADPH (and not NAD+ and NADP+) affect absorbance at 340 nm. Cell extracts are separated into 3 aliquots (A1, A2, and A3). A1 is untreated and the absorbance at 340 nm is measured. A2 is treated with an enzyme that converts all of the NADP+ to NADPH, and then the absorbance at 340 nm is measured. A3 is treated with an enzyme that converts all of the NADPH to NADP+, and then the absorbance at 340 nm is measured. A1 − A3 is the NADPH content, and A2 − A1 is the NADP+ content of the extract. BAEC, which were 80–90% confluent, were incubated with 5.5 and 25 mm glucose for 3 h. GSH was measured using the GSH kit BIOXYTECH GSH-400 from OXIS International, Inc. ROS were measured as described previously using the dye 2′,7′-dichlorofluorescein diacetate (6Tian W.-N. Braunstein L.D. Apse K. Pang J. Rose M. Tian X. Stanton R.C. Am. J. Physiol. 1999; 276: C1121-C1131Crossref PubMed Google Scholar). Dichlorofluorescein fluorescence was measured using a microplate fluorometer (Cambridge Technology). BAEC were incubated with 5.5 and 25 mm glucose for 3 h. Adenylate cyclase activity was measured by a new method designed by us. The method is based on relative differences between two cell lysates rather than absolute activities. Cell lysates either from cells exposed to 5 mm glucose or lysates from cells exposed to 25 mm glucose were prepared. Then each lysate was separated into two aliquots (A and B). To be certain there was no limitation on substrate availability, ATP (the substrate for adenylate cyclase) was added to each aliquot, and then aliquot A was immediately boiled for 10 min. Aliquot B was incubated at 37 °C for 15 min and then boiled for 10 min. Then cAMP was measured using cAMP kit TRK432 from Amersham Pharmacia Biotech. Isobutylmethylxanthine (1 mm) was included in all solutions to inhibit cAMP-dependent phosphodiesterase activity. The amount of cAMP from aliquot A represents the base-line cAMP level. The difference between the cAMP amount in aliquot B and in aliquot A (i.e. B − A) represents the adenylate cyclase-produced cAMP. In the figures, the results are expressed as the increase in cAMP (aliquot B cAMP − aliquot A cAMP level). BAEC that were 80% confluent were exposed to 1 mCi of 32P/P100 culture plate in phosphate-free buffer for 2 h. Then the glucose concentration was increased by adding glucose to the plate. After 3 h cells were harvested and lysed, and G6PD was immunoprecipitated using antibody to G6PD as described previously (5Tian W.-N. Braunstein L.D. Pang J. Stuhlmeier K. Xi Q.-C. Tian X. Stanton R.C. J. Biol. Chem. 1998; 273: 10609-10617Abstract Full Text Full Text PDF PubMed Scopus (402) Google Scholar). Following immunoprecipitation, G6PD protein was resolved by SDS-polyacrylamide gel electrophoresis and then exposed to x-ray film. Antisense experiments were done with the excellent assistance of Drs. Jane Alexander and Joseph Loscalzo of the Boston University Medical Center, who provided both the antisense sequence and suggestions on transfection techniques. Transfection was carried out using the method described by the Sequitur, Inc. kit. In brief, BAEC, which were 80–90% confluent, were washed once with Opti-MEM without serum or antibiotics. Then 13.2 μl of Oligofectin I (stock solution, 2 mg/ml) was added in 1 ml of Opti-MEM in a polystyrene tube. In another tube 4.0 μl of antisense oligomer was added to 1 ml of Opti-MEM. The two solutions were mixed together and allowed to sit for 15 min. The mixture was then added to cells for 5 h. 0.5 ml of mixture was added to each well of a 24-well plate. After 5 h of incubation, the medium was removed, and Opti-MEM with 10% serum was added to the cells for 24 h. Experiments were then started at this time. The antisense sequence we used was 5′-AGUAGUACCCACGUAGCCCACUGGGA-3′. After transfection, BAEC were incubated with 5.5 and 25 mm glucose for 24 h. Then, cell death and G6PD activity were measured as described above. Vascular endothelial cells are known targets of hyperglycemia. Thus we have used BAEC, a commonly used system, for the studies. BAEC were incubated with various concentrations of glucose. Cells were harvested, and cell lysates were obtained as described previously (7Tian W.-N. Pignatare J.N. Stanton R.C. J. Biol. Chem. 1994; 269: 14798-14805Abstract Full Text PDF PubMed Google Scholar, 8Stanton R.C. Seifter J.L. Boxer D.C. Zimmerman E. Cantley L.C. J. Biol. Chem. 1991; 266: 12442-12448Abstract Full Text PDF PubMed Google Scholar). G6PD activity was measured as described previously (7Tian W.-N. Pignatare J.N. Stanton R.C. J. Biol. Chem. 1994; 269: 14798-14805Abstract Full Text PDF PubMed Google Scholar, 8Stanton R.C. Seifter J.L. Boxer D.C. Zimmerman E. Cantley L.C. J. Biol. Chem. 1991; 266: 12442-12448Abstract Full Text PDF PubMed Google Scholar). Fig. 1 A shows that glucose concentrations as low as 10 mm were sufficient to inhibit G6PD activity. As an osmotic control another sugar, raffinose, was added to medium containing 5.5 mm glucose to achieve the same concentrations as seen with the high glucose conditions. Raffinose had no effect on G6PD activity (data not shown). Also note that a raffinose control was done for all the remaining experiments and in all cases did not cause any of the effects seen with increased glucose. High glucose has been shown to increase cell death in BAEC (12Morishita R. Nakamura S. Nakamura Y. Aoki M. Moriguchi A. Kida I. Yo Y. Matsumoto K. Nakamura T. Higaki J. Ogihara T. Diabetes. 1997; 46: 138-142Crossref PubMed Scopus (120) Google Scholar). As seen in Fig. 1 B, increased glucose led to cell death. To further determine an association between G6PD activity and BAEC cell death, two inhibitors of G6PD were used. Both dehydroepiandrosterone (5Tian W.-N. Braunstein L.D. Pang J. Stuhlmeier K. Xi Q.-C. Tian X. Stanton R.C. J. Biol. Chem. 1998; 273: 10609-10617Abstract Full Text Full Text PDF PubMed Scopus (402) Google Scholar, 6Tian W.-N. Braunstein L.D. Apse K. Pang J. Rose M. Tian X. Stanton R.C. Am. J. Physiol. 1999; 276: C1121-C1131Crossref PubMed Google Scholar) and 6-aminonicotinamide (5Tian W.-N. Braunstein L.D. Pang J. Stuhlmeier K. Xi Q.-C. Tian X. Stanton R.C. J. Biol. Chem. 1998; 273: 10609-10617Abstract Full Text Full Text PDF PubMed Scopus (402) Google Scholar, 6Tian W.-N. Braunstein L.D. Apse K. Pang J. Rose M. Tian X. Stanton R.C. Am. J. Physiol. 1999; 276: C1121-C1131Crossref PubMed Google Scholar) worsened cell death at all glucose levels. Importantly there was a close correlation with impaired G6PD activity and cell death (Fig. 1, compare A andB). If high glucose-induced cell death is due, at least in part, to impaired activity of G6PD, then cells exposed to high glucose should be more susceptible to oxidant-induced cell death. Fig. 1 C shows that H2O2-induced cell death was significantly enhanced in cells exposed to high glucose. We have recently published that G6PD activity can affect apoptosis in a variety of cell types including BAEC (6Tian W.-N. Braunstein L.D. Apse K. Pang J. Rose M. Tian X. Stanton R.C. Am. J. Physiol. 1999; 276: C1121-C1131Crossref PubMed Google Scholar). Because high glucose led to an increase in cell death we explored whether there was an increase in apoptosis. Fig. 1 D shows that high glucose increased apoptosis in BAEC. Considering the relative lack of specificity of the G6PD inhibitor, DHEA, we more specifically addressed the role of G6PD by using an antisense oligonucleotide to reduce G6PD activity. Fig. 2 A shows that both in normal and high glucose conditions transfection with an antisense sequence led to a highly significant decrease in G6PD activity. Note that the decrease in G6PD activity was greater in the BAEC exposed to high glucose. Also the transfection efficiency for the Sequitur Oligofectin system is about 30%. As is clear from the results in Fig. 2 A, there is almost a 60% decrease in G6PD activity following transfection. Although we have no specific answer as to why there is such a profound decrease, it is our belief that there is on-going cell death, which leads to release of the oligonucleotide and subsequent transfection of other cells over time. This observation and idea as to the reason for the profound effect on G6PD activity was also seen and suggested by a collaborator. 2J. Loscalzo, personal communication. Fig. 2 B shows that the antisense sequence also led to a significant enhancement of the cell death especially under high glucose conditions. These results strongly support an important role for G6PD in high glucose-induced cell death. We have previously shown that inhibition of G6PD by pharmacologic agents led to a decrease in NADPH levels and an increase in ROS (5Tian W.-N. Braunstein L.D. Pang J. Stuhlmeier K. Xi Q.-C. Tian X. Stanton R.C. J. Biol. Chem. 1998; 273: 10609-10617Abstract Full Text Full Text PDF PubMed Scopus (402) Google Scholar, 6Tian W.-N. Braunstein L.D. Apse K. Pang J. Rose M. Tian X. Stanton R.C. Am. J. Physiol. 1999; 276: C1121-C1131Crossref PubMed Google Scholar). Fig. 3, A and B, shows that a high glucose-induced inhibition of G6PD also led to a decrease in NADPH and an increase in ROS. Because NADPH is the critical substrate required to maintain reduced levels of glutathione, we also measured GSH levels under normal and high glucose conditions. Fig. 3 C shows that high glucose also led to a decrease in GSH levels consistent with our findings on NADPH.Figure 1A, high glucose causes inhibition of G6PD activity. Concentrations as low as 10 mm glucose as early as 15 min after exposure are sufficient to suppress G6PD activity. BAEC were initially grown in DMEM (5.5 mmglucose) + 10% serum that were about 80% confluent and then switched to DMEM with 2% serum plus various concentrations of glucose ranging from 5.5 to 25 mm for 15 min to 24 h. The same experiment done on confluent cells yielded similar results as shown above. All subsequent figures use the same approach with slight changes in either percent confluence or incubation time. The data were normalized by protein and presented as average ± S.E. from three separate experiments, each run in at least triplicate (*,p < 0.01; **, p < 0.005 compared with control). B, high glucose and G6PD inhibitors increase cell death as measured by trypan blue exclusion. High glucose increased cell death. Inhibition of G6PD enhanced BAEC cell death. BAEC were treated as in Fig. 1 A (except used at 60% confluent), switched to DMEM + 2% serum and glucose for 24 h, and then exposed to an inhibitor of G6PD (either 100 μm DHEA or 5 mm6-aminonicotinamide (ANAD) for 3 h. Data were normalized by cell number and expressed as means ± S.E. of 10 separate experiments, each run in triplicate (***, p < 0.001 compared with control). C,H2O2 enhances cell death in the presence of high glucose. BAEC were prepared as described in B. Then cells were treated with 400 μmH2O2 for 3 h. Cell death was determined by trypan blue exclusion. Data were normalized by cell number and expressed as means ± S.E. of five separate experiments, each run in triplicate (***, p < 0.001; ****, p< 0.0001 compared with control). D, high glucose increases apoptosis. Apoptosis was measured by 4′,6′-diamidine-2′-phenylindole dihydrochloride after 12 h of incubation to glucose ± 100 μm DHEA and 5 mm6-aminonicotinamide (ANAD). Data were normalized by cell number and expressed as means ± S.E. of three separate experiments, each run in triplicate (**, p < 0.005; ****, p < 0.0001 compared with control).View Large Image Figure ViewerDownload Hi-res image Download (PPT)Figure 3A, high glucose leads to a suppression of NADPH levels. BAEC were treated as described in Fig. 1 Aexcept the incubation time was 3 h. NADPH was determined by a newly developed method that is based on a single extract method (11Zhang Z., Yu, J. Stanton R.C. Anal. Biochem. 2000; 285: 163-167Crossref PubMed Scopus (98) Google Scholar). Data were normalized by protein and expressed as means ± S.E. of three separate experiments, each run in triplicate (***,p < 0.001 compared with control). B, high glucose causes an increase in ROS. BAEC were treated as described in Fig. 1 A, and then exposed to glucose for 3 h. Then 2′,7′-dichlorofluorescein diacetate was added and incubated for 10 min at room temperature. ANAD, aminonicotinamide. Data were normalized by protein and expressed as means ± S.E. of five separate experiments, each run in triplicate (*, p < 0.01; **, p < 0.005 compared with control).C, high glucose leads to a suppression of GSH levels. BAEC were treated as in Fig. 1 A except the incubation time was 3 h. GSH was determined as described under "Experimental Procedures." Data were normalized by protein and expressed as means ± S.E. of three separate experiments, each run in triplicate (***, p < 0.001 compared with control).View Large Image Figure ViewerDownload Hi-res image Download (PPT) Published reports led us to hypothesize that high glucose-induced increase in cAMP may mediate the inhibition of G6PD, at least in part. First, increased glucose leads to an increase in cAMP levels (13Kamal K. Du W. Mills I. Sumpio B.E. J. Cell. Biochem. 1998; 71: 491-501Crossref PubMed Scopus (56) Google Scholar). Second, increased cAMP has led to enhancement of cell death in certain cell types (13Kamal K. Du W. Mills I. Sumpio B.E. J. Cell. Biochem. 1998; 71: 491-501Crossref PubMed Scopus (56) Google Scholar). Third, Costa Rosa et al. (14Costa Rosa L.-F.B.P. Curi R. Murphy C. Newsholme P. Biochem. J. 1995; 310: 709-714Crossref PubMed Scopus (68) Google Scholar) showed that compounds that increase cAMP levels led to a decrease in G6PD activity from rat peritoneal macrophages. Fig. 4 A shows that high glucose caused highly significant increases in cAMP. The high glucose-induced increase in cAMP in BAEC could be due to either increased adenylate cyclase activity and/or to decreased phosphodiesterase activity. Fig. 4 B shows that high glucose led to an increase in adenylate cyclase activity. Considering the significant increase in adenylate cyclase activity in Fig. 4 B and the more modest increase in cAMP in Fig. 4 A, it is likely that phosphodiesterase activity is increased as well. If increased cAMP leads to inhibition of G6PD, then pharmacologically induced increases in cAMP should lead to inhibition of G6PD activity. Fig. 4 C shows that isobutylmethylxanthine, a phosphodiesterase inhibitor, led to inhibition of G6PD activity. Fig. 4 D shows that an analog of cAMP, 8-bromo-cAMP, also led to inhibition of G6PD activity. These results confirm that increased cAMP levels cause inhibition of G6PD activity in BAEC. cAMP is believed to cause most of its effects via cAMP-dependent protein kinase A (PKA). H89 is an inhibitor of PKA activity. If PKA is inhibiting G6PD then inhibition of PKA should lead to an increase in G6PD activity. Fig. 5 A shows that H89 led to an increase in G6PD activity. To determine whether PKA acts directly on G6PD, PKA was incubated with G6PD in cell lysates. Fig. 5 Bshows that incubation of the catalytic subunit of PKA with G6PDin vitro led to inhibition of G6PD. If PKA is acting directly on G6PD, then high glucose should cause phosphorylation of G6PD. Fig. 5 C shows that high glucose led to a highly significant increase in phosphorylation. There is compelling evidence that G6PD is the principal source of NADPH utilized in redox regulation. First, under oxidative stress conditions many studies have shown that G6PD and the pentose phosphate pathway are routinely elevated (15Slekar K.H. Kosman D.J. Culotta V.C. J. Biol. Chem. 1996; 271: 28831-28836Abstract Full Text Full Text PDF PubMed Scopus (179) Google Scholar, 16Ursini M.V. Parella A. Rosa G. Salzano S. Martini G. Biochem. J. 1997; 323: 801-806Crossref PubMed Scopus (134) Google Scholar, 17Cramer T. Cooke S. Ginsberg L.C. Kletzien R.F. Stapleton S.R. Ulrich R.G. J. Biochem. Toxicol. 1995; 10: 293-298Crossref PubMed Scopus (34) Google Scholar). Second, Pandolfi et al. (24Pandolfi P.P. Sonati F. Rivi R. Mason P. Grosveld F. Luzzatto L. EMBO J. 1995; 14: 5209-5215Crossref PubMed Scopus (469) Google Scholar) used homologous recombination in mouse embryonic stem cells to produce a cell line totally deficient in G6PD. The G6PD null cells were exquisitely sensitive to oxidative stress. In addition the null cells had significantly decreased growth rates as compared with cells expressing G6PD and significantly reduced cloning efficiencies. The authors concluded that G6PD was critical for NADPH production and was the principal source of NADPH. Earlier work by Rosenstraus and Chaisin (18Rosenstraus M. Chaisin L. Proc. Natl. Acad. Sci. U. S. A. 1975; 72: 493-497Crossref PubMed Scopus (76) Google Scholar) using a G6PD-deficient Chinese hamster ovary cell line, which was produced through classic mutagenesis techniques, also showed that the G6PD-deficient cells were more susceptible to oxidative stress. Recently Ursini and colleagues (19Salvemini F. Franze A. Iervolino A. Filosa S. Salzano S. Ursini M.V. J. Biol. Chem. 1999; 274: 2750-2757Abstract Full Text Full Text PDF PubMed Scopus (297) Google Scholar) overexpressed G6PD in Hl-60 cells and showed that the glutathione levels in the cell were elevated and the cells were more resistant to oxidative stress. Taken together, these results show that G6PD activity is of central importance to cellular redox regulation. Thus Fig. 3 A shows that the decrease in NADPH caused by high glucose is consistent with the fact that decreased activity of G6PD is the mechanistic reason for this decrease. The redox level in the cell is a balance between reductants and oxidants. Intracellular increases in ROS can be due to exposing cells to external oxidants such as H2O2 and diamide (20Fisher, A. B. (1988) in Upjohn Symposium on Oxygen Radicals, Park City, UT, January 24–30, 1988 (Cerrutti, P., Fridovich, I., and McCord, J. M., ed) pp. 34–38, New York.Google Scholar, 21Fridovich, I. (1988) in Upjohn Symposium on Oxygen Radicals, Park City, UT, January 24–30, 1988 (Cerrutti, P., Fridovich, I., and McCord, J. M., ed.) pp. 1–5, New York.Google Scholar, 22Fridovich I. J. Biol. Chem. 1997; 272: 18515-18517Abstract Full Text Full Text PDF PubMed Scopus (1071) Google Scholar) or due to the intracellular production of ROS in the form of H2O2, O⨪2, ⋅OH, and singlet oxygen (20Fisher, A. B. (1988) in Upjohn Symposium on Oxygen Radicals, Park City, UT, January 24–30, 1988 (Cerrutti, P., Fridovich, I., and McCord, J. M., ed) pp. 34–38, New York.Google Scholar, 21Fridovich, I. (1988) in Upjohn Symposium on Oxygen Radicals, Park City, UT, January 24–30, 1988 (Cerrutti, P., Fridovich, I., and McCord, J. M., ed.) pp. 1–5, New York.Google Scholar, 22Fridovich I. J. Biol. Chem. 1997; 272: 18515-18517Abstract Full Text Full Text PDF PubMed Scopus (1071) Google Scholar). ROS can be formed by normal biologic reactions via the actions of a variety of intracellular oxidases such as xanthine oxidase, monoamine oxidase, NADPH oxidase, and urate oxidase (20Fisher, A. B. (1988) in Upjohn Symposium on Oxygen Radicals, Park City, UT, January 24–30, 1988 (Cerrutti, P., Fridovich, I., and McCord, J. M., ed) pp. 34–38, New York.Google Scholar, 21Fridovich, I. (1988) in Upjohn Symposium on Oxygen Radicals, Park City, UT, January 24–30, 1988 (Cerrutti, P., Fridovich, I., and McCord, J. M., ed.) pp. 1–5, New York.Google Scholar). Nonenzymatic production of ROS can also occur. Superoxide may be formed from the auto-oxidation by molecular oxygen (21Fridovich, I. (1988) in Upjohn Symposium on Oxygen Radicals, Park City, UT, January 24–30, 1988 (Cerrutti, P., Fridovich, I., and McCord, J. M., ed.) pp. 1–5, New York.Google Scholar, 22Fridovich I. J. Biol. Chem. 1997; 272: 18515-18517Abstract Full Text Full Text PDF PubMed Scopus (1071) Google Scholar) of such chemical groups as thiols, quinones, catechols, etc. (20Fisher, A. B. (1988) in Upjohn Symposium on Oxygen Radicals, Park City, UT, January 24–30, 1988 (Cerrutti, P., Fridovich, I., and McCord, J. M., ed) pp. 34–38, New York.Google Scholar, 21Fridovich, I. (1988) in Upjohn Symposium on Oxygen Radicals, Park City, UT, January 24–30, 1988 (Cerrutti, P., Fridovich, I., and McCord, J. M., ed.) pp. 1–5, New York.Google Scholar). ROS can also be produced in cellular compartments such as mitochondria, peroxisomes, and microsomes (20Fisher, A. B. (1988) in Upjohn Symposium on Oxygen Radicals, Park City, UT, January 24–30, 1988 (Cerrutti, P., Fridovich, I., and McCord, J. M., ed) pp. 34–38, New York.Google Scholar, 21Fridovich, I. (1988) in Upjohn Symposium on Oxygen Radicals, Park City, UT, January 24–30, 1988 (Cerrutti, P., Fridovich, I., and McCord, J. M., ed.) pp. 1–5, New York.Google Scholar). The consequences of increased oxidative stress include oxidation of lipids, proteins, carbohydrates, and nucleic acids that lead to defects in metabolism, cell growth, and other cell physiologic processes (20Fisher, A. B. (1988) in Upjohn Symposium on Oxygen Radicals, Park City, UT, January 24–30, 1988 (Cerrutti, P., Fridovich, I., and McCord, J. M., ed) pp. 34–38, New York.Google Scholar, 21Fridovich, I. (1988) in Upjohn Symposium on Oxygen Radicals, Park City, UT, January 24–30, 1988 (Cerrutti, P., Fridovich, I., and McCord, J. M., ed.) pp. 1–5, New York.Google Scholar). Proteins involved in redox regulation are glutathione, catalase, and superoxide dismutase. All of these enzymes require a reductant for their operation. The main intracellular reductant is NADPH, which is principally produced by G6PD (23Simonian N.A. Coyle J.T. Annu. Rev. Pharmacol. Toxicol. 1996; 36: 83-106Crossref PubMed Scopus (992) Google Scholar, 24Pandolfi P.P. Sonati F. Rivi R. Mason P. Grosveld F. Luzzatto L. EMBO J. 1995; 14: 5209-5215Crossref PubMed Scopus (469) Google Scholar, 25Horecker B.L. J. Chem. Educ. 1965; 42: 244-253Crossref PubMed Scopus (34) Google Scholar, 26Kletzien R.F. Harris P.K.W. Foellmi L.A. FASEB J. 1994; 8: 174-181Crossref PubMed Scopus (461) Google Scholar). The next enzyme in the pentose phosphate pathway is also a dehydrogenase, 6-phosphogluconate dehydrogenase, which also produces NADPH. However, the activity of this enzyme is completely dependent on G6PD as G6PD is the sole source of the substrate for this enzyme (6-phosphogluconate). There is one other enzyme that produces NADPH, NADP+-dependent malate dehydrogenase. In most cells this enzyme cannot replace G6PD as the principal source of NADPH. However, in liver, adipose tissue, pancreatic β-cells, and macrophages NADP+-dependent malate dehydrogenase may play a significant role in NADPH production. Thus in most cells, proper activity of G6PD is required for the cell to properly defend against oxidative stress. Thus the results in Fig. 3 B showing increased ROS following exposure to high glucose are likely due, at least in part, to decreased G6PD activity. The evidence that oxidative stress plays a role in the complications of diabetes mellitus is as follows (3Low P.A. Nickander K.K. Tritschler H.J. Diabetes. 1997; 46 (suppl.): 38-42Crossref Google Scholar, 4Giugliano D. Ceriello A. Paolisso G. Diabetes Care. 1996; 19: 257-267Crossref PubMed Scopus (1705) Google Scholar, 10Asahina T. Kashiwagi A. Nishio Y. Ikebuchi M. Harada N. Tanaka Y. Takagi Y. Saeki Y. Kikkawa R. Shigeta Y. Diabetes. 1995; 44: 520-526Crossref PubMed Scopus (105) Google Scholar, 27Ha H. Yoon S.J. Kim K.H. Kidney Int. 1994; 46: 1620-1626Abstract Full Text PDF PubMed Scopus (47) Google Scholar). 1) Cells exposed to high glucose have increased levels of ROS. 2) Cellular antioxidant mechanisms are altered in cells exposed to high glucose and in cells from patients with diabetes. 3) Studies have shown increased levels of lipid peroxidation and increased levels of reactive oxygen species (4Giugliano D. Ceriello A. Paolisso G. Diabetes Care. 1996; 19: 257-267Crossref PubMed Scopus (1705) Google Scholar, 27Ha H. Yoon S.J. Kim K.H. Kidney Int. 1994; 46: 1620-1626Abstract Full Text PDF PubMed Scopus (47) Google Scholar). 4) In cell culture systems protection against various effects of high glucose can be obtained by adding a variety of intracellular antioxidants such as glutathione, superoxide dismutase, or catalase to the cells (4Giugliano D. Ceriello A. Paolisso G. Diabetes Care. 1996; 19: 257-267Crossref PubMed Scopus (1705) Google Scholar). 5) Many of the proposed mechanisms,e.g. increased protein kinase C (28Chakraborti T. Ghosh S.K. Michael J.R. Batabyal S.K. Chakraborti S. Mol. Cell. Biochem. 1998; 187: 1-10Crossref PubMed Scopus (78) Google Scholar), increased advanced glycation end products (4Giugliano D. Ceriello A. Paolisso G. Diabetes Care. 1996; 19: 257-267Crossref PubMed Scopus (1705) Google Scholar, 29Baynes J.W. Thorpe S.R. Diabetes. 1999; 48: 1-9Crossref PubMed Scopus (2158) Google Scholar), and increased aldose reductase, are all associated with increased oxidative stress. Increased oxidative stress may be a unifying factor for all of these mechanisms. Thus there is strong evidence that increased oxidative stress is of major importance in the etiology of diabetic complications. Recent work by our laboratory has shown that G6PD activity per se is important for proper cell growth (5Tian W.-N. Braunstein L.D. Pang J. Stuhlmeier K. Xi Q.-C. Tian X. Stanton R.C. J. Biol. Chem. 1998; 273: 10609-10617Abstract Full Text Full Text PDF PubMed Scopus (402) Google Scholar) and preventing cell death under various conditions (6Tian W.-N. Braunstein L.D. Apse K. Pang J. Rose M. Tian X. Stanton R.C. Am. J. Physiol. 1999; 276: C1121-C1131Crossref PubMed Google Scholar). In particular cell death caused by oxidative stress was dependent on G6PD levels. Thus the results seen in this paper and our previously published studies are consistent with the hypothesis that altered G6PD activity (caused by high glucose) would predispose cells to cell damage/death. In particular, the studies using the G6PD inhibitors and the antisense oligonucleotide to G6PD strongly support an important role for G6PD in high glucose cell death. In 1975, Zawalich et al. (30Zawalich W.S. Karl R.C. Ferrendelli J.A. Matschinsky F.M. Diabetologia. 1975; 11: 231-235Crossref PubMed Scopus (59) Google Scholar) showed that increased glucose led to an increase in cAMP in pancreatic islets. In 1978, Jackowski et al. (31Jackowski M.M. Tepperman H.M. Tepperman J. J. Cyclic Nucleotide Res. 1978; 4: 323-333PubMed Google Scholar) showed that rats exposed to high glucose had increased cAMP levels in their adipose tissue. More recently, Kamal et al. (13Kamal K. Du W. Mills I. Sumpio B.E. J. Cell. Biochem. 1998; 71: 491-501Crossref PubMed Scopus (56) Google Scholar) showed that increased cAMP likely played an antiproliferative role in human dermal microvascular endothelial cell line. Thus there is published evidence for high glucose causing an increase in cAMP levels. The results shown in Figs. 3 and 4 are consistent with these previous observations. Fig. 4 Bsuggests that a mechanism underlying this effect is at least in part due to increased activation of adenylate cyclase. Interestingly, the effect may differ depending on cell type. For example human corneal cells exposed to increased glucose showed a decrease in cAMP levels (32Wang H.Z. Wu K.Y. Lin C. Fong J.C. Hong S.J. Kaohsiung J. Med. Sci. 1997; 13: 566-571PubMed Google Scholar). The different responses of cAMP levels may truly reflect cell-specific responses or they may reflect different experimental conditions (e.g. incubation time may play a role). Further studies are required to clarify this. As previously noted, Costa Rosa et al. (14Costa Rosa L.-F.B.P. Curi R. Murphy C. Newsholme P. Biochem. J. 1995; 310: 709-714Crossref PubMed Scopus (68) Google Scholar) showed that cAMP could inhibit macrophage G6PD. However, they suggested that this was true for only macrophage G6PD. Costa Rosa et al. (14Costa Rosa L.-F.B.P. Curi R. Murphy C. Newsholme P. Biochem. J. 1995; 310: 709-714Crossref PubMed Scopus (68) Google Scholar) compared their findings on macrophage G6PD to only one other source of G6PD. Thus they did not do an extensive study of various sources of G6PD. The research reported here suggests that cAMP inhibition of G6PD is not limited to macrophages. Thus it is possible that cAMP might regulate G6PD in a variety of cell types. Taken together, our results strongly suggest that in BAEC, high glucose led to an increase in cAMP that, via cAMP-dependent PKA, caused phosphorylation and inhibition of G6PD activity. This led to decreased NADPH, increased ROS, and then cell damage/death. This impaired activity of G6PD likely plays a critical role in the pathogenesis of endothelial cell damage observed in patients with diabetes. We are very grateful to Chun-Rui Zhao for excellent technical assistance in the phosphorylation experiments and to Jia Yu for excellent technical assistance and critical suggestions. We are also very grateful to Drs. Joseph Loscalzo and Jane Leopold of the Boston University Medical Center for their assistance in the antisense experiments.