Increase in Glucose-6-Phosphate Dehydrogenase in Adipocytes Stimulates Oxidative Stress and Inflammatory Signals

DHEA (dehydroepiandrosterone) and lipopolysaccharide were purchased from Sigma. BAY11-7082, diphenyleneiodonium chloride, and N-acetyl-l-cysteine (NAC) were purchased from Calbiochem. hTNFα was purchased from R&D Systems.

3T3-L1 preadipocytes were grown to confluence in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% bovine calf serum (Gibco BRL). At 2 days after confluence (day 0), differentiation of the 3T3-L1 cells was induced in DMEM containing 10% fetal bovine serum (FBS; Gibco BRL), methylisobutylxanthine (500 μmol/l), dexamethasone (1 μmol/l), and insulin (5 μg/ml) for 48 h. Then, culture medium was changed on alternate days with DMEM containing 10% FBS and insulin (5 μg/ml). Raw264 cells were maintained with DMEM containing 10% FBS, and THP-1 cells were maintained with RPMI containing 10% FBS.

Epididymal adipose tissues were excised from C57BL/6 and db/db mice (12 weeks old), rinsed in PBS, minced, and digested for 45 min at 37°C in Krebs-Ringer phosphate buffer (pH 7.4) with 2% BSA and type I collagenase (1.75 mg/ml). Digested tissue was filtered through a 250-μm nylon mesh to remove undigested tissue and centrifuged at 3,000 rpm for 5 min. The floating adipocyte fraction and stromal-vascular pellet were washed several times. Peritoneal macrophages were isolated from C57BL/6 mice. The mice were injected intraperitoneally with sterile fluid thioglycollate medium. After 4 days, the peritoneal cells were harvested and washed with PBS containing 3 mmol/l EDTA. Primary macrophages were cultured with DMEM containing 10% FBS to allow cell adherence. The nonadherent cells were removed by washing with PBS, and the adherent macrophages were refed with DMEM containing 10% FBS or adipocyte culture supernatants (conditioned media).

G6PD adenovirus was produced by Neurogenex (Seoul, Korea). G6PD cDNA was in-frame fused with Flag epitope tag in its NH₂ terminus, which was cloned into an adenoviral vector, pEntrBHRNX. For adenoviral infection, 3T3-L1 adipocytes (at day 6 after differentiation) were incubated with serum-free DMEM and various titers (10–30 multiplicity of infection) of adenovirus for 16 h at 37°C. Then, culture medium was replaced with fresh medium. Each experiment was performed at 48–72 h after viral infection.

Total RNA was isolated with Trizol (Invitrogen Life Technologies). cDNA was synthesized using a SuperScript first-strand synthesis system (Invitrogen Life Technologies). Quantitative real-time PCR was performed on a iCycler real-time PCR detection system (Bio-Rad), using SYBR Green I (BioWhittaker Molecular Applications), with each primer set shown in online appendix Table 1 (available at http://diabetes.diabetesjournals.org). The relative values of each mRNA were normalized with the levels of cyclophilin or GAPDH mRNA.

Nuclear extracts from the 3T3-L1 adipocytes were isolated as described in a previous report, with minor modifications (24), and used for electrophoretic mobility shift assay. Target DNA sequences of NF-κB were used as probes; one strand is as follows: 5′-AGTTGAGGGGACTTTCCCAGGC-3′. Double-stranded oligonucleotides were end-labeled with [γ-³²P]ATP and T4 polynucleotide kinase. The nuclear extracts were mixed with radio-labeled probes (1 pmol per 30,000 cpm) in reaction buffer: 4 mmol/l Tris (pH 7.9), 23 mmol/l HEPES, 66 mmol/l NaCl, 5 mmol/l MgCl₂, 0.7 mmol/l EDTA, 1 mmol/l dithiothreitol, 14% (vol/vol) glycerol, and 4 μg poly(dI-dC). After incubation for 25 min at room temperature, the samples were loaded onto a native polyacrylamide gel (4%).

Nitrite was measured using the Griess reaction (25). Culture media (100 μl each) were collected and incubated with an equal volume of Griess reagent for 10 min at room temperature. The nitrite concentration was determined from the absorbance measured at 550 nm, using sodium nitrite as the standard. Hydrogen peroxide was measured using an Amplex Red Hydrogen peroxide assay kit (Molecular Probes) (8). Culture media (50 μl each) were collected and incubated with Amplex Red reagent for 30 min at room temperature. Hydrogen peroxide was determined by an Envision 2102 multilabel reader at 560 nm (PerkinElmer). Cellular ROS was measured using 2,7-dichlorodihydrofluorescein diacetate (DCF-DA) and luminol (Molecular Probes). 3T3-L1 adipocytes were washed with PBS and incubated in the dark for 30 min with DCF-DA (5 μmol/l). The fluorescence of DCF-DA was monitored by a fluorescence microscope (Olympus) and quantitated by an Envision 2102 multilabel reader at an excitation wavelength of 492 nm and emission wavelength of 530 nm. Luminol (50 μmol/l) was used for the quantitative measurement of the cellular ROS. Chemiluminescence of luminol was determined using a Berthold LB9501 luminometer for 1 min. As a positive control of ROS measurement, differentiated 3T3-L1 adipocytes were treated with H₂O₂ (100 μmol/l) simultaneously with DCF-DA or luminol, and increased DCF-DA fluorescence or chemiluminescence caused by H₂O₂ was determined.

3T3-L1 cells were differentiated on glass coverslips and infected with the indicated adenovirus. After 2 days, cells were cocultured with THP-1 monocytes and nonadherent THP-1 cells removed by washing with PBS. The THP-1 cells adhered to the 3T3-L1 cells were fixed with 3.7% paraformaldehyde, permeabilized with 0.5% PBS with Triton X-100 (PBST), and incubated with 0.1% PBST containing 3% BSA for blocking. Subsequently, the cells were incubated with CD68 monoclonal antibody (DakoCytomation) at room temperature for 1 h and washed with 0.1% PBST. They were then incubated at room temperature with tetramethyl rhodamine isothiocyanate–conjugated secondary antibodies. The coverslips were rinsed and placed on a slide glass with mounting solution containing 4′,6-diamino-2-phenylindole. The cells were visualized using a fluorescence microscope (Olympus), and CD68-positive cells were quantitated using imaging analysis program (LSM510 version 3.5; Carl Zeiss).

Cocultured 3T3-L1 adipocytes and Raw264 macrophage cells were separated following the manufacturer’s protocol, using a CD11b MicroBeads system (MACS; Miltenyi Biotec). Briefly, cocultured cells were collected and washed twice with 10 ml washing buffer (PBS supplemented with 2 mmol/l EDTA and 0.5% BSA). Total cells (<10⁷) were finally suspended in 100 μl washing buffer, mixed with 10 μl CD11b microbeads, and incubated at 12°C for 15 min. Then, cells were washed with 1 ml washing buffer. The cell pellet was suspended with 500 μl washing buffer and loaded in a large cell column. Adipocyte fractions were flow throughed with 500 μl washing buffer in a magnetic field, and macrophage fractions were flow throughed in a nonmagnetic field with 1 ml washing buffer, using the plunger supplied, after column washing three times with 500 μl washing buffer.

Previously, we reported that G6PD expression is increased in the adipose tissue of several obese and diabetic mouse models (23). However, the specific stimuli that regulate G6PD expression in adipocytes under such conditions have not been identified. To investigate this, 3T3-L1 adipocytes were exposed to several conditions associated with metabolic disorders, and the levels of G6PD mRNA were examined by quantitative real-time PCR. As shown in Fig. 1A, G6PD mRNA expression was significantly increased by high-glucose treatment, whereas the other NADPH-producing enzymes, such as malic enzyme and isocitrate dehydrogenase, were only marginally affected. Similarly, high levels of insulin, TNF-α, and lipopolysaccharide remarkably elevated G6PD mRNA in adipocytes (Fig. 1B), implying that in metabolic disorders, G6PD expression is promoted in adipocytes.

To determine the specific cell types in which G6PD is expressed in adipose tissue, we separated primary adipocytes and stromal-vascular cells from epididymal fat tissue and examined their G6PD mRNA levels. Compared with the stromal-vascular fraction, G6PD mRNA was abundantly expressed in the adipocyte fraction and was higher in db/db than C57BL/6 lean mice (Fig. 2A). In accordance with previous reports (3,5,6,26), stromal-vascular cells abundantly express MCP1 and CD68 (Fig. 2B and C), whereas the adipocyte fraction predominantly expresses adiponectin (Fig. 2D). In addition, mRNA levels of MCP1 and CD68 were increased in adipocytes of obese mice, whereas adiponectin was decreased (Fig. 2). These results clearly indicate that G6PD is primarily expressed in adipocytes in fat tissues.

Although oxidative stress is a major causative factor in the etiology of metabolic disorders, defined mechanisms by which obesity induced oxidative stress have not been clearly understood. Consistent with previous reports showing that differentiated adipocytes preserve high amounts of ROS (8), we observed that the activity of G6PD, a major adipogenic NADPH-producing enzyme, was positively correlated with the degree of differentiation and the levels of ROS during adipogenesis (online appendix Fig. 1). Nevertheless, the effect of G6PD overexpression in adipocytes on oxidative stress control has not been thoroughly examined. To address this issue, we investigated whether the overexpression of G6PD in adipocytes would affect pro- or antioxidative signaling. In the G6PD-overexpressing adipocytes infected with G6PD adenovirus (AdG6PD), the mRNA levels of pro-oxidative enzymes, including iNOS and NADPH oxidase components (gp91^phox, p22^phox, p67^phox, p47^phox, and p40^phox), were elevated, whereas those of the antioxidative enzymes, including superoxide dismutase (Cu,Zn-SOD) and GPx-1 (glutathione peroxidase), were slightly increased or not altered (Fig. 3A). Our observation that adipogenic G6PD overexpression was unlikely to have induced pro-oxidative gene expression resulted from the effect of G6PD on adipocyte differentiation because AdG6PD-infected adipocytes exhibited little difference of adipogenic gene expression compared with mock adenovirus (AdMock)-infected adipocytes (online appendix Fig. 2). These results suggest that adipogenic G6PD overexpression promotes the pro-oxidative pathways rather than the antioxidative pathways, which might be linked to the increase of ROS production in the adipose tissue of obese subjects. To verify these observations, we directly determined the levels of cellular ROS and NO in G6PD-overexpressing adipocytes. Compared with AdMock controls, the AdG6PD-infected adipocytes exhibited increased cellular ROS (Fig. 3B–D) and NO levels (Fig. 3E). In parallel, we used TNF-α and hydrogen peroxide as positive controls for pro-oxidant signaling molecules (online appendix Fig. 3). TNF-α–treated adipocytes significantly induced expression of proinflammatory genes, such as IL-6 and MCP1, as well as pro-oxidative enzymes, such as iNOS and NADPH oxidase (gp91^phox). In contrast, hydrogen peroxide–treated 3T3-L1 cells did not significantly increase a certain set of genes, suggesting that hydrogen peroxide–induced oxidative stress might regulate a gene set slightly different from that of G6PD-dependent oxidative stress in adipocytes (online appendix Fig. 3). In addition, increased ROS level by G6PD overexpression was attenuated by an antioxidant, NAC, or an NADPH oxidase inhibitor, diphenyleneiodonium chloride (Fig. 3C and D). These data indicate that G6PD overexpression in adipocytes primarily contributes to the production of oxidative stress, which might induce proinflammatory response and insulin resistance.

It is well known that oxidative stress activates NF-κB signaling, which plays an essential role in insulin resistance and chronic inflammation in adipocytes (27–29). Because G6PD overexpression increased oxidative stress in adipocytes (Fig. 3), we examined whether G6PD overexpression in adipocytes would also alter NF-κB signaling and its target gene expression. To test this idea, 3T3-L1 adipocytes were infected with AdMock or AdG6PD and subjected to Western blot analyses with antibodies against p65, p50, inhibitory κB (IκB) kinase-β (IKK-β), and IκB-α. Compared with AdMock, the AdG6PD-infected adipocytes revealed increased expression of the p65 subunits of NF-κB in the nuclear fraction (Fig. 4A, lane 3 vs. lane 4). TNF-α treatment was used as a positive control for the activation of NF-κB signaling (Fig. 4A, lane 2). In addition, the level of IκB-α, which is an inhibitor of NF-κB, was reduced, and that of IKK-β was 2.5-fold increased by G6PD overexpression compared with the AdMock controls (Fig. 4B, lane 1 vs. lane 3). These effects were augmented by TNF-α treatment (Fig. 4B, lane 4). To confirm these results, the nuclear extracts of the 3T3-L1 adipocytes infected with either AdMock or AdG6PD were examined for the DNA-binding activity of NF-κB. In accordance with the above results, the DNA-binding activity of NF-κB was 1.8-fold increased in the G6PD-overexpressing adipocytes (Fig. 4C, lane 1 vs. lane 3).

Because NF-κB signaling was stimulated in the G6PD-overexpressing adipocytes, we investigated the expression levels of NF-κB targets. As shown in Fig. 4D, iNOS was significantly elevated in the G6PD-overexpressing adipocytes, whereas adiponectin and CAAAT/enhancer-binding protein-α were substantially decreased. Similar results were obtained on treatment with TNF-α (Fig. 4D, lanes 2 and 4). Taken together, these results imply that the oxidative stress produced by G6PD overexpression in adipocytes would, at least in part, contribute to insulin resistance and/or inflammatory signals by activating NF-κB.

Growing evidence suggests that oxidative stress in adipocytes is tightly correlated with inflammatory signals (8,9,30–32). The observation that G6PD overexpression in adipocytes enhanced oxidative stress prompted us to investigate the mRNA expression profiles of several proinflammatory genes, including TNF-α, IL-6, MCP1, resistin, and chemokine (C-C motif) receptor 2 (CCR2), that are associated with insulin resistance. In G6PD-overexpressing adipocytes, the expression of these proinflammatory genes was evidently increased (Fig. 5A). In addition, pretreatment with NF-κB inhibitors, such as BAY11-7082 and rosiglitazone, and the antioxidant NAC, as well as the G6PD inhibitor dehydroepiandrosterone (33), substantially reduced the effects of G6PD on proinflammatory gene expression (Fig. 5A). The effects of these drugs, including the NF-κB inhibitors, the antioxidant, and the G6PD inhibitor, also downregulated the expression of iNOS and NADPH oxidase in the G6PD-overexpressing adipocytes (Fig. 5B). Under the same conditions, mRNA levels of adiponectin and leptin were decreased by G6PD-overexpressing adipocytes (Fig. 5C). Collectively, these data suggest that increase of inflammatory signals in the G6PD-overexpressing adipocytes is mediated by oxidative stress and NF-κB activation.

Adipocytes secrete various adipocytokines that regulate whole-body energy homeostasis by acting on other tissues, including brain, liver, muscles, and macrophages (34). Recently, interactions between adipose tissue and peripheral macrophages have been implicated in the insulin resistance and inflammatory signals observed in obesity (3,6). However, the identity of the molecular linkers of this interaction were largely unknown. The observation that the increase in G6PD expression in adipocytes promoted oxidative stress and proinflammatory signals led to the examination of the effects of adipogenic G6PD overexpression on macrophage biology. To examine whether adipogenic G6PD overexpression alters humoral factors, we determined the levels of proinflammatory cytokines, such as MCP1, anti-inflammatory cytokines, such as adiponectin, and amounts of hydrogen peroxide from AdMock- or AdG6PD-infected adipocytes. In supernatants (conditioned media) from AdG6PD-infected adipocytes, the levels of secreted MCP1 and hydrogen peroxide were elevated, whereas that of adiponectin was decreased (Fig. 6A and B). Next, we investigated the effects of supernatants from G6PD-overexpressing adipocytes on primary cultured peritoneal macrophages. Peritoneal macrophages were incubated with supernatants from adipocytes infected with either AdMock or AdG6PD in the presence or absence of NAC, an antioxidant, or rosiglitazone, an antioxidant and anti-inflammatory drug (3,35,36) (Fig. 6C). Total RNA was isolated from each treated macrophage and analyzed for gene expression. Interestingly, the mRNA expression of macrophage genes, including iNOS, TNF-α, IL-6, MCP1, resistin, and CCR2, that are responsible for the inflammatory signals and macrophage differentiation was enhanced by the supernatants from G6PD-overexpressing adipocytes (Fig. 6D). These effects were attenuated by NAC or rosiglitazone (Fig. 6D), indicating that certain molecules secreted from the G6PD-overexpressing adipocytes modulate proinflammatory signals, as well as macrophage-specific gene expression.

Recently, Suganami et al. (37) reported that adipocyte-derived mediators influence inflammatory gene expressions in macrophages, using a coculture system with adipocytes and macrophages. To examine the effects of adipogenic G6PD overexpression on the interaction between adipocytes and macrophages, macrophage Raw264 cells were cocultured with AdMock- or AdG6PD-infected 3T3-L1 adipocytes, and the inflammatory gene expressions were analyzed (Fig. 7A). As shown in Fig. 7B, expression of proinflammatory or macrophage-specific genes, including iNOS, TNF-α, IL-6, MCP1, CD11b, CD68, and SR-A, was remarkably increased when adipocytes were cocultured with macrophages, and these effects were augmented by G6PD overexpression in adipocytes. In addition, expression of cell adhesion molecules, such as intercellular adhesion molecule-1 and vascular cell adhesion molecule-1, was enhanced by G6PD overexpression, suggesting that adipogenic G6PD increase would promote cross talk or physical interaction between adipocytes and macrophages, accompanied by modification of gene expression. Next, because inflammatory signaling molecules, such as MCP1, TNF-α, IL-6, and CCR2, are expressed in both adipocytes and macrophages, we decided to investigate the relative contribution of each cell type on inflammatory gene expression. Cocultured adipocytes and Raw264 cells were separated by CD11b microbeads (as detailed in research design and methods), and mRNAs from each cell type were isolated and analyzed (Fig. 7C). Interestingly, mRNA expression of MCP1 and IL-6 in both 3T3-L1 adipocytes and Raw264 macrophages was profoundly increased by coculture of adipocytes and macrophages and even further augmented by G6PD overexpression in adipocytes (Fig. 7D, left panel). In the case of iNOS and TNF-α, the relative induction folds of these genes were elevated in both cell types by coculture (Fig. 7D, middle panel). Although Raw264 cells predominantly expressed iNOS and TNF-α mRNA, induction of these genes was more remarkably increased in G6PD-overexpressing adipocytes by coculture (Fig. 7D, middle panel). In contrast, macrophage-specific genes, including CD68, CCR2, and CD11b, were elevated in Raw264 cells by coculturing with adipocytes, and they were slightly more increased by adipogenic G6PD overexpression (Fig. 7D, right panel). Additionally, the expression of intercellular adhesion molecule-1 in Raw264 cells and vascular cell adhesion molecule-1 in both Raw264 and 3T3-L1 cells was further enhanced by G6PD overexpression, suggesting that adipogenic G6PD might stimulate the interaction between macrophages and adipocytes. These results suggest that cross talk between adipocytes and macrophages promotes inflammatory signals and that the increase of inflammatory gene expression might be differently induced in a cell type–specific manner. Therefore, it is likely that adipogenic G6PD overexpression would activate the cross talk between adipocytes and macrophages and mediate enhanced chronic inflammatory signals in fat tissue in obesity.

Next, we investigated whether G6PD overexpression in adipocytes promotes the direct recruitment of monocytes onto adipose tissue. To address this issue, AdMock- and AdG6PD-infected 3T3-L1 cells were incubated with human monocyte THP-1 cells. Consistent with the results above, we observed that more THP-1 cells were recruited to the G6PD-overexpressing adipocytes compared with the control adipocytes, and antioxidant NAC pretreatment partially attenuated monocyte recruitment (Fig. 8). These results suggest that G6PD-overexpressing adipocytes provide a certain milieu that might promote the recruitment of circulating monocytes onto adipocytes and induce macrophage infiltration into the adipose tissue of obese subjects.

In obese subjects, adipose tissue exhibits increased oxidative stress, which is a major contributor to metabolic disorders such as insulin resistance, type 2 diabetes, cardiovascular disease, and atherosclerosis (8,31,38). It has been reported that antioxidant drugs, including α-lipoic acid and mimetics of SOD or catalase, improve diabetes complications by reducing oxidative stress (31,39,40). In addition, several drugs for cardiovascular diseases or type 2 diabetes, such as statins and thiazolidinediones, also decrease intracellular ROS (36,41,42). Thus, it is very likely that obesity-related oxidative stress is closely associated with metabolic disorders. Despite this, the question of how obesity produces oxidative stress remains to be elucidated.

Insulin resistance, a crucial factor for the development of metabolic disorders, is induced by several molecules, including insulin, free fatty acids, endotoxins, and certain cytokines such as TNF-α, IL-6, and MCP1, as well as by high levels of glucose. We found that the expression of G6PD was elevated in response to these types of stimuli; all of which are elevated in obesity and type 2 diabetes (Fig. 1). Moreover, among several NADPH-producing enzymes, G6PD in adipocytes was notably upregulated by these signals. Because these signals are also known to activate NF-κB signaling, resulting in the induction of inflammation and insulin resistance, we examined whether the upregulation of G6PD under these conditions was mediated by the NF-κB pathway. Of note, we observed that the mouse G6PD promoter was transactivated by NF-κB (data not shown). Thus, it appears that the G6PD in adipocytes might be actively involved in the pathogenesis of metabolic disorders such as obesity and insulin resistance, and the increase of G6PD in adipocytes might play a causative role in the development of metabolic disorders.

To further explore the molecular mechanism underlying G6PD-induced insulin resistance (23), we focused on oxidative stress and inflammatory signals because G6PD is a crucial enzyme to maintain cellular redox potential. In adipocytes, G6PD overexpression greatly enhances pro-oxidative enzymes, including iNOS and NADPH oxidase, thus inducing oxidative stress (Fig. 3). Moreover, we believe that the G6PD-dependent increase of pro-oxidative enzymes is limited to adipocytes because G6PD overexpression in primary hepatocytes did not show any effects (online appendix Fig. 4). The elevated oxidative stress in the adipose tissue of obese subjects with high levels of NADPH oxidase activity (8) is closely correlated with the increase in G6PD expression (23). We also observed that the level of cellular NADPH in adipocytes was increased by G6PD overexpression, high glucose, high insulin, or TNF-α, which have been implicated in metabolic disorders (online appendix Fig. 5). In addition, it has been reported that the enzymatic activities of NADPH oxidase and iNOS, the major enzymes involved in oxidative stress, are elevated in obesity and metabolic disorders (8,20,43–46). Therefore, these findings strongly suggest that the abnormal increase of G6PD in the adipocytes of obese subjects leads to oxidative stress by upregulation of NADPH oxidase and iNOS. As mentioned above, mitochondria also play a crucial role in the generation of cellular ROS, and mitochondrial dysregulation has been implicated in metabolic disorders, including insulin resistance and diabetes (17,47–49). In this regard, the effects of adipogenic G6PD overexpression on mitochondrial function need to be further examined to get a better picture of metabolic disorders.

It has been demonstrated that oxidative stress–induced activation of NF-κB signaling appears to be associated with the pathogenesis of insulin resistance and type 2 diabetes (28,29,50,51). Many studies have also reported that antioxidant drugs and IKK-β inhibitors protect against insulin resistance in obesity and type 2 diabetes (27,31,51,52). Thus, it is of interest that G6PD-overexpressing adipocytes enhanced NF-κB expression and its DNA-binding activity (Fig. 4), accompanied by regulation of NF-κB target genes, including inflammatory and chemokine genes (Fig. 5). Consequently, the G6PD-dependent increase in inflammatory gene expression was attenuated by pretreatment with NF-κB inhibitors and antioxidant drugs (Fig. 5), suggesting that the effects of adipogenic G6PD overexpression are likely to be mediated through the oxidative stress and NF-κB signaling pathways.

Recent studies have revealed that the altered adipocyte biology in obesity enhances adipocyte-macrophage interactions, resulting in chronic inflammatory signals and insulin resistance (3,5). Because G6PD-overexpressing adipocytes change the expression of several adipocytokines, we hypothesized that the secretory factors of G6PD-overexpressing adipocytes might affect macrophage gene expression. In support of this idea, we observed that the expression of many macrophage-specific genes and the recruitment of THP-1 monocytes onto adipose tissue was remarkably increased in G6PD-overexpressing adipocytes (Figs. 6–8), implying that G6PD overexpression in adipocytes produces primary signals for inflamed adipocytes. Furthermore, our results provide an important clue to understand how macrophages and adipocytes communicate with each other and how, in obese subjects, peripheral macrophages respond to the signals from adipocytes with promoted infiltration into the adipose tissue.

In summary, we demonstrated a novel role of G6PD in obesity-induced metabolic disorders (Fig. 9). Variable environmental factors, including hyperglycemia, hyperinsulinemia, and the presence of cytokines, induce G6PD expression in adipose tissue. Subsequent elevations in G6PD in adipocytes increase oxidative stress; the activation of NF-κB signaling, which eventually stimulates inflammatory responses and insulin resistance; and dysregulation of adipocytokines. Increase of G6PD in adipocytes promotes macrophage recruitment and affects macrophage-specific gene expression, contributing to the chronic inflammation in obesity. Therefore, the identification of regulatory tools for G6PD activity may prove promising for treating metabolic disorders such as obesity-induced insulin resistance and chronic inflammation.

This work was supported in part by grants from the Molecular and Cellular Biodiscovery Research Program, Stem Cell Research Center of the 21st Century Frontier Research Program (SC13150), Research Center for Functional Cellulomics of Science Research Center Program, and the National Research Laboratory Program of Korea Science and Engineering Foundation. A.H.C., K.H.K., and J.B.K. were supported by BK21 Research Fellowships from the Ministry of Education and Human Resources Development.

We thank N. Houstis and E.D. Rosen for providing materials. We also thank Y.S. Lee, J.J. Chung, and K.C. Hsu for critically reading the manuscript.