Hydrogen Peroxide Generated During Cellular Insulin Stimulation Is Integral to Activation of the Distal Insulin Signaling Cascade in 3T3-L1 Adipocytes

In a variety of cell types, insulin stimulation elicits the rapid production of H 2 O 2 which causes the oxidative inhibition of protein-tyrosine phosphatases (PTPases) and enhances the tyrosine phosphorylation of proteins in the early insulin action cascade (Mahadev et al., J.Biol.Chem. 276 :21938, 2001). In the present work, we explored the potential role of insulin-induced H 2 O 2 generation on downstream insulin signaling using diphenyleneiodonium (DPI), an inhibitor of cellular NADPH oxidase that blocks insulin-stimulated cellular H 2 O 2 production. DPI completely inhibited the activation of phosphatidylinositol (PI) 3’-kinase activity by insulin and reduced the insulin-induced activation of the serine kinase Akt by up to 49%; these activities were restored when H 2 O 2 was added back to cells that had been pretreated with DPI. Interestingly, the H 2 O 2 -induced activation of Akt was entirely mediated by upstream stimulation of PI 3’-kinase activity, since treatment of 3T3-L1 adipocytes with the PI 3’-kinase inhibitors wortmannin or LY294002 completely blocked the subsequent activation of Akt by exogenous H 2 O 2 . Preventing oxidant generation with DPI also blocked insulin-stimulated glucose uptake and GLUT4 translocation to the plasma membrane, providing further evidence for an oxidant signal in the regulation of the distal insulin signaling cascade. Finally, in contrast to the cellular mechanism of H 2 O 2 generation by other growth factors, such as platelet-derived growth factor, we also found that insulin-stimulated cellular production of H 2 O 2 may occur through a unique pathway, independent of cellular PI 3’-kinase activity. Overall, these data provide insight into the physiological role of insulin-dependent H 2 O 2 generation which is not only involved in the regulation of tyrosine phosphorylation events in the early insulin signaling cascade, but also has important effects on the regulation of downstream insulin signaling, involving the activation of PI 3’-kinase, Akt and ultimately cellular glucose transport in response to insulin.


INTRODUCTION
Major advances in our understanding of the regulation of the insulin action pathway have focused on the key role of tyrosine phosphorylation of the insulin receptor and its cellular substrate proteins (1). Insulin binding leads to autophosphorylation of specific residues of the transmembrane insulin receptor and activation of the intrinsic tyrosine kinase activity of its intracellular domains (2). The insulin signal is then transmitted further into the cell through the tyrosine phosphorylation of specific sites on cellular substrate proteins (e.g., IRS and Shc), which act as docking sites for the binding and activation of a variety of src-homology 2 (SH2) domain-containing signaling proteins (3). Much of insulin's downstream signaling to metabolic events involves the activation of phosphatidylinositol (PI) 3'-kinase activity by the docking of its p85 subunit to tyrosine phosphorylated IRS-1 and IRS-2 (4-6), which is linked to a number of distal responses in adipocytes including the activation of the protein kinase Akt and subsequent vesicle translocation and glucose transport activation (7).
This reversible protein-tyrosine phosphorylation of components in the insulin signaling pathway has been shown to be regulated in a variety of ways. A major regulatory influence is exerted by specific cellular protein-tyrosine phosphatases (PTPases) which are involved in balancing the steady-state tyrosine phosphorylation of the insulin receptor and its substrate proteins (8). In turn, the enzymes in the PTPase superfamily are themselves regulated by oxidation/reduction reactions in vivo, since they require a reduced form of the thiol side chain of the catalytic cysteine residue for phosphotyrosine hydrolysis (9)(10)(11). In a recent study, we also showed that a burst of intracellular H 2 O 2 resulting from insulin stimulation results in the reversible oxidative inhibition of cellular PTPase activity (12). The rapid inhibition of PTPases that negatively regulate insulin signaling was associated with enhanced insulin-stimulated tyrosine phosphorylation of the insulin receptor and high M r IRS proteins, and was found to play an important role in the propagation of the early insulin signal.
In addition to the oxidative inactivation of thiol-dependent PTPases, it has been intriguing to speculate that reactive oxygen species, especially the H 2 O 2 burst generated shortly after insulin binding, may be involved in the regulation of more distal cellular insulin signaling.
There are an increasing number of examples in the literature indicating that at least some of the effects of various hormones, growth factors and cytokines on specific signaling pathways can 4 involve superoxide and H 2 O 2 (13)(14)(15). For example, in the case of epidermal growth factor (EGF) and platelet-derived growth factor (PDGF), evidence has been accumulating to suggest that ligand binding is integrally associated with the generation of a cellular oxidant signal and also that redox species can mimic ligand-mediated signaling by the growth factors themselves (13;16). One report suggested that H 2 O 2 may be involved in "redox priming" of the insulin receptor in muscle cells, which may render receptor kinase activation more efficient (17).
However, little is known about how reactive oxygen species may affect downstream components of the insulin action pathway and a fuller characterization of these potential effects in various types of insulin-sensitive cells is needed.
To explore how the oxidant signal arising from insulin stimulation might be involved in the transmission of distal post-receptor insulin signaling, we used diphenyleneiodonium chloride (DPI), an inhibitor of cellular NADPH-oxidases (18), which we found to fully block the insulinstimulated generation of H 2 O 2 in 3T3-L1 adipocytes. This inhibitory effect was used to explore the role of the insulin-induced oxidant burst on distal insulin signaling, including glucose transport. We found that insulin-stimulated H 2 O 2 was essential for the activation of Akt via stimulation of PI 3'-kinase activity, which also subsequently enhanced glucose transport activation in 3T3-L1 adipocytes. Furthermore, the mechanism by which insulin elicits cellular H 2 O 2 production was shown to be independent of cellular PI 3'-kinase activity, unlike PDGF, which has been reported to generate cellular H 2 O 2 by a PI 3'-kinase-mediated pathway (16).
Thus, insulin-dependent H 2 O 2 generation appears to occur by a unique metabolic pathway and has a cellular role not only in enhancing the early activation of insulin receptor autophosphorylation and kinase activity, but is also integrally involved in the regulation of events in the distal insulin signaling cascade. Monoclonal anti-phosphotyrosine (4G10) and polyclonal antibodies to the insulin receptor βsubunit, IRS-1 and the p85 subunit of PI 3'-kinase were from Upstate Biotechnology (Lake Placid, NY). Antibodies to phosphorylated Akt (Ser473) and Akt protein (not isoform specific) and the Akt kinase activity kit were purchased from New England Biolabs (Beverly, MA). Cell culture -3T3-L1 preadipocytes were cultured in Dulbecco's modified Eagle's medium containing 25 mM glucose (DMEM) plus 10% fetal calf serum in an 5% CO 2 atmosphere and were differentiated to adipocytes as previously described (19). Briefly, confluent cells were placed in differentiation medium (DMEM containing 10% fetal bovine serum, 100 nM insulin, 0.25 µM dexamethasone and 500 µM isobutylmethylxanthine) for 2 days. The medium was then changed to DMEM containing 10% fetal bovine serum and 100 nM insulin. After an additional 6 days, cells were starved overnight in DMEM containing 0.5% (w/v) bovine serum albumin (BSA) and used for the experiments.

Para-nitrophenylphosphate
Assay of intracellular H 2 O 2 in 3T3-L1 adipocytes -Intracellular generation of H 2 O 2 was visualized as described (12;20). At various times indicated (0, 1, 5, 10 minutes) after stimulation with 100 nM insulin with and without prior treatment with 10 µM DPI for 30 minutes, dishes of differentiated 3T3-L1 cells were washed with MEM medium (lacking phenol red) and then incubated in the dark for 10 minutes with 5,6-chloromethyl-2',7'-dichlorodihydrofluorescein diacetate (CM-DCF). The fluorescence of CM-DCF was measured by Bio-Rad confocal microscope at an excitation wavelength of 488 nm and emission at 515-540 nm. To avoid photooxidation of the indicator dye, the fluorescence image was collected by a single rapid scan with identical parameters for all samples. Where indicated, the fluorescence intensity was quantitated from sampled images using Scion Image software (Scion Corporation, Fredrick, MD). PTPase enzyme activity using 32 P-RCM-lysozyme as substrate -Recombinant human insulin receptors from transfected CHO cells (23) were partially purified on wheat germ lectinagarose (Vector Laboratories, Burlingame, CA) as described (24). RCM-lysozyme was radioactively labeled on tyrosine by phosphorylation with the insulin receptor preparation and [γ- 32 7 4 o C. The pellet was washed 3 times with 20% TCA and dialyzed overnight against 50 mM imidazole-HCl, pH 7.2. PTPase activity was assayed using the indicated amount of cell fraction protein in 50 mM HEPES, pH 7.0 and 2 mM EDTA, without DTT, as indicated. The reaction was initiated by the addition of 20 µl of [ 32 P]-phosphotyrosyl RCM-lysozyme (~20,000 dpm) and terminated by the addition of 0.9 ml of acidic charcoal mixture, consisting of 0.9 M NaCl, 90 mM sodium pyrophosphate, 2 mM NaH 2 PO 4 and 4% (w/v) Norit A activated charcoal (26).
After centrifugation in a microfuge, the amount of radioactivity in 0.4 ml of supernatant was measured by Cerenkov counting in a liquid scintillation counter. The initial rate of RCMlysozyme hydrolysis was estimated from the linear portion of the earliest time points of the enzymatic reaction where less than 20% of the RCM-lysozyme was hydrolyzed during the 5 minute reaction period.
Immunoblotting -After the indicated experimental treatments, cells were lysed in buffer containing 50 mM HEPES, pH 7.5, 150 mM NaCl, 1% Triton X-100, 100 mM sodium fluoride, 1 mM EGTA, 1 mM EDTA, 2 mM sodium vanadate, 10 mM sodium pyrophosphate, 1 mM phenylmethylsulfonyl fluoride, and a protease inhibitor cocktail (Sigma). The lysates were briefly sonicated, centrifuged at 13,000 x g for 10 minutes, and 75 µg protein of the cleared supernatant was resolved by SDS-polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride (PVDF) membrane using a semidry western blotting apparatus (AP Biotech). PVDF membranes were subjected to immunoblotting with either monoclonal antibody for phosphotyrosine (4G10) to detect insulin receptor β-subunit and IRS tyrosine phosphorylation, polyclonal antibody to detect phospho-Akt or additional antibodies to detect total protein levels of the insulin receptor β-subunit, IRS-1 and Akt, where indicated. Following incubation with horseradish peroxidase-conjugated secondary antibodies, proteins were visualized by enhanced chemiluminescence, according to the instructions provided by the manufacturer. The immunoblotting signals were quantitated using an ImageStation 440 (Kodak). (v/v) Triton X-100) in the presence or absence of 10 µM DPI for 2 hrs at 4 o C. The reaction was stopped using Laemmli gel sample buffer (27). Samples were boiled at 100 o C for 3 minutes and resolved by SDS-polyacrylamide gel electrophoresis. Protein was transferred to PVDF membrane and the membranes were subjected to immunoblotting with the monoclonal antibody for phosphotyrosine (4G10). Following incubation with horseradish peroxidase-conjugated secondary antibody, insulin receptor tyrosine phosphorylation was visualized by enhanced chemiluminescence.
PI 3'-kinase activity in 3T3-L1 adipocytes -PI 3'-kinase activity was determined as previously described (28). Briefly, 1 mg of 3T3-L1 adipocyte cell lysate was incubated overnight  (29). The reaction was stopped 4 minutes later by washing the cells three times with ice-cold PBS. The cells were then solubilized in 0.05% sodium dodecyl sulfate at 37 o C for 30 minutes, and aliquots were subjected to scintillation counting. Nonspecific uptake (<10% of the total) was determined in the presence of cytochalasin B (50 µM), and was subtracted from the total uptake.
Translocation of GLUT4 to the cell surface was visualized using the plasma membrane sheets assay using polyclonal anti-GLUT4 antibodies as described previously (29;30).
Statistical Analyses: Quantitative data are presented as the mean ± SEM for 3-5 experiments. Statistical analysis was based on Student's t test for comparison of two groups and one-way analysis of variance for multiple group comparisons. A p value less than 0.05 was used to determine statistical significance.

The insulin-induced burst of intracellular oxidant is abolished by DPI -To visualize the
intracellular generation of reactive oxygen species in response to insulin, 3T3-L1 adipocytes were incubated with 5,6-chloromethyl-2',7'-dichlorodihydrofluorescein diacetate (CM-DCF), a sensitive oxidant indicator dye that readily diffuses into cells and is trapped after cleavage by cellular esterases. After oxidation by cellular reactive oxygen species, and H 2 O 2 in particular, CM-DCF becomes highly fluorescent as detected by confocal microscopy (12;20). Following stimulation of cells with 100 nM insulin, a rapid increase in CM-DCF fluorescence was detected within 1 minute that begins to diminish at 5 minutes ( Figure 1). In the presence of 10 µM DPI, an inhibitor of NADPH oxidases (31), the insulin-induced CM-DCF fluorescence is completely abolished, consistent with previous observations that the insulin effect is mediated by an NADPH oxidase system in adipocytes (32;33). In our previous work, this oxidant signal was also completely blocked by catalase, identifying H 2 O 2 as one of the relevant reactive oxygen species that is generated in response to insulin (12).

Insulin-induced inhibition of PTPase activity is reversed by DPI -We recently
demonstrated that one of the major consequences of the insulin-induced generation of H 2 O 2 was to inhibit cellular PTPase activity by a mechanism that involved oxidation of the catalytic PTPase thiol residue. To further assess this effect, we tested how DPI, as an inhibitor of the oxidant generated by cellular insulin stimulation, affected the inhibition of cellular PTPase activity by insulin. Using our recently reported technique involving strict anaerobic conditions for cell lysate preparation as well as the PTPase assays to prevent the oxidation and enzyme inhibition that occurs on exposure to air, we measured changes in the cellular PTPase activity that accompanies insulin stimulation (34). These experiments were performed without reducing agents added during the assay to evaluate the endogenous level of enzyme activity dependent on the oxidation state of the PTPase catalytic thiol residues.
Following treatment of 3T3-L1 adipocytes with 100 nM insulin, using the pNPP assay there was a significant reduction in PTPase enzyme activity of 32 and 62% of the basal level Chinese Hamster Ovary cells overexpressing insulin receptors were exposed to 10 µM DPI and insulin-stimulated tyrosine autophosphorylation of the receptor was evaluated in vitro by western blot analysis with anti-phosphotyrosine antibodies (23). In addition, we assayed PTPase activity in the presence or absence of 10 µM DPI added to 3T3-L1 cell lysates in vitro. DPI had no direct effect in the cell lysates on the rate or extent of insulin receptor autophosphorylation or on PTPase activity (data not shown). These findings provided further support for our conclusion that the mechanism of DPI action on protein tyrosine phosphorylation and cellular PTPase activities involves its inhibition of NADPH oxidase in the intact cells which abrogates the insulin-induced production of oxidant species.

Inhibition of insulin-induced generation of H 2 O 2 by DPI down-regulates PI 3'-kinase
activity -In order to gain insight into how the oxidant signal from insulin stimulation might play a role in the downstream activation of kinases in the insulin action pathway, we tested whether inhibition of insulin-induced generation of H 2 O 2 by DPI affected PI 3'-kinase activation induced by insulin (Figure 4). In control cells, without DPI exposure, 5 minutes of stimulation with 100 nM insulin increased the PI 3'-kinase activity in the 3T3-L1 adipocytes by 1.8 fold over basal (p=0.036; n=3). Cellular treatment with 10 µM DPI for 30 minutes did not affect basal PI 3'kinase activity; however, DPI exposure completely blocked the increase in PI 3'-kinase activity observed with insulin stimulation at 5 minutes (p=0.006; n=3). These results suggested that the oxidant signal resulting from cellular insulin stimulation played an important role in the transmission of the insulin signal to PI 3'-kinase.
The role of H 2 O 2 in PI 3'-kinase activation by insulin was further tested by adding back H 2 O 2 to the cells that had been pretreated with DPI ( Figure 4). Treatment of cells with 1 mM H 2 O 2 had a strong activating effect on the basal level of PI 3'-kinase activity to 3.6-fold over the level in control cells. There was also a trend for insulin treatment of the cells exposed to H 2 O 2 to further stimulate PI 3'-kinase activity, but this was not statistically significant. Pretreatment of the cells with DPI also was inhibitory to the H 2 O 2 -induced increase in PI 3'-kinase activity prior to insulin and also blocked the apparent insulin-stimulated increase in PI 3'-kinase activity in the H 2 O 2 -treated cells.
In additional control experiments, we demonstrated that DPI or H 2 O 2 had no direct effects on cellular PI 3'-kinase activity. Activated PI 3'-kinase was immunoprecipitated with anti-p85 antibody from lysates of cells treated with insulin or H 2 O 2 . Following exposure to 10 µM DPI in vitro for 30 minutes, there was no change in production of PI3P, assayed as described in the Experimental Procedures (data not shown). In addition, treatment of immunoprecipitated PI 3'-kinase from resting cells with 1 mM H 2 O 2 did not affect enzyme activity. Overall, these results indicate that H 2 O 2 plays an important role in the activation of the PI 3'-kinase pathway by insulin in 3T3-L1 adipocytes, and can act when added exogenously, as well as when generated with the target cells by insulin stimulation. Furthermore, the effects of DPI and H 2 O 2 on PI 3'kinase activity are indirect, mediated through the cellular signaling pathways present in the intact cells.

Abolition of H 2 O 2 generation by DPI inhibits insulin-stimulated Akt activation in 3T3-L1
adipocytes -Downstream of PI 3'-kinase, we next evaluated whether blockade of the insulinstimulated oxidant signal with DPI affected the activation of Akt by insulin ( Figure 5). 3T3-L1 cells were incubated with or without 10 µM DPI for 30 minutes prior to treatment with 100 nM insulin for 1 or 5 minutes. DPI treatment reduced the level of Akt phosphorylation by 38 and 49% at 1 and 5 minutes of insulin stimulation, respectively.
To confirm the involvement of H 2 O 2 in the cellular activation of Akt, we then treated the cells with 1 mM H 2 O 2 for 10 minutes and the phosphorylation of Akt protein was measured as above using immunoblotting with phospho-Akt antibodies. In the presence of H 2 O 2 , there was a dramatic increase in the basal level of Akt phosphorylation, which was not significantly increased further by insulin stimulation. Exposure of adipocytes to DPI prior to treatment with H 2 O 2 also did not alter the level of Akt phosphorylation, indicating that DPI acted at a step prior to Akt activation by H 2 O 2 and had no direct effect of its own on the intracellular phosphorylation of Akt ( Figure 5).

H 2 O 2 -induced phosphorylation of Akt is mediated by activation of PI 3'-kinase-Since
activation of Akt in the insulin signaling pathway is known to be linked to the upstream activation of PI 3'-kinase, we then tested whether inhibition of Akt activation by DPI treatment was mediated by the effect of DPI to reduce the activation of PI 3'-kinase under these conditions. As shown above in Figure 5, H 2 O 2 is a potent activator of Akt in an insulin-independent manner.
Interestingly, treatment of 3T3-L1 adipocytes with the PI 3'-kinase inhibitors wortmannin (100 nM) or LY294002 (50 µM) for 10 minutes completely blocked the subsequent activation of Akt phosphorylation in response to cell stimulation with exogenous H 2 O 2 ( Figure 6). These important results indicated that an oxidant signal in the cell is tightly coupled to the activation of PI 3'kinase, which in turn, triggers Akt phosphorylation and activation.

Inhibition of H 2 O 2 production suppresses insulin-stimulated glucose uptake and GLUT4
plasma membrane translocation-Since the oxidant signal from insulin action was involved in the activation of PI 3'-kinase and subsequently of Akt, we also evaluated how inhibition of H 2 O 2 generation with DPI affected a key biological activity in the differentiated 3T3-L1 adipocyte model, glucose transport. In control cells, without DPI treatment, insulin stimulated a mean of 3.7-fold increase in glucose uptake (Figure 7). In the presence of DPI, basal glucose transport was not significantly affected (88% of control); however, insulin stimulation of glucose transport was completely nullified by prior incubation of the 3T3-L1 adipocytes with DPI, with only a 23% increase above basal levels. For comparison, cells were treated with H 2 O 2 , which caused a stimulation of glucose uptake to the same extent as insulin alone, to 3.5-fold over basal levels.
Cell treatment with DPI prior to H 2 O 2 partially inhibited both the basal stimulation of glucose uptake by H 2 O 2 as well as the uptake stimulated by the combination of insulin and H 2 O 2 (by 54% and 50%, respectively). Although differing somewhat in magnitude, the general pattern of response to DPI and H 2 O 2 in glucose transport is similar to the profile observed in the PI 3'kinase assay (Figure 4).
The glucose uptake data was also validated by the plasma membrane sheets assay to assess GLUT4 translocation ( Figure 8). The visualization of GLUT4 appearance at the cell  In addition to the effects of oxidants when added from outside of the cell, attention has recently been paid to the important role in cellular signaling played by endogenous H 2 O 2 and other reactive oxygen species that are rapidly generated directly within target cells following treatment with various growth factors and hormones (13;16). Elaboration of H 2 O 2 in response to insulin signal transduction has been known to occur in adipose tissue for many years, although the significance of reactive oxygen species for the insulin action cascade has not been identified (32;42;43). We recently showed that treatment of 3T3-L1 adipocytes and HepG2 hepatoma cells with insulin rapidly generates a burst of H 2 O 2 , which has a major impact on the early transmission of the insulin receptor signal by modulation of the steady-state tyrosine phosphorylation of the insulin receptor and its cellular substrate proteins (12). In these studies, we showed that the effect of insulin-stimulated cellular H 2 O 2 was mediated, at least in part, by inhibition of the catalytic activity of cellular PTPases. We demonstrated that this effect also specifically involved the single-domain intracellular PTPase, PTP1B, which has been strongly implicated in the negative regulation of the insulin action pathway (44).

Role of cellular PI 3'-kinase activity in insulin-stimulated H 2 O 2 production-
In the present work, we capitalized on the observation that DPI, a flavoprotein NADPH oxidase inhibitor, blocked the production of intracellular H 2 O 2 arising from stimulation with insulin. This inhibitory agent was then used to determine which of a variety of steps in the postreceptor insulin signaling cascade might be affected by the loss of insulin-generated H 2 O 2 in the 3T3-L1 adipocyte model. Initially, we confirmed the findings of our previous study to show that eradicating the oxidant signal from insulin diminished the insulin-induced inactivation of cellular PTPase activity, and that this effect was associated with enhanced tyrosine phosphorylation of the insulin receptor and its high molecular mass substrates, IRS-1 and 2 (Figures 2 and 3) (12).
Thus, H 2 O 2 coupled to the insulin-receptor interaction appears to be essential for the initiation of the early insulin signal, by reducing the rapid and potent dephosphorylating activity of cellular PTPases proximate to the receptor (45).
The novel finding reported in the present study is that the oxidant signal from insulin functions not only to enhance early insulin action, but also serves an essential role in the downstream insulin action cascade. In particular, the endogenous oxidant enhances the activation of PI 3'-kinase by insulin, since this process is inhibited by DPI. Furthermore, we have demonstrated that the oxidant effect on PI 3'-kinase activation is integral to the downstream activation of the Akt kinase, since the PI 3'-kinase inhibitors wortmannin and LY-294002 completely blocked the activation of Akt mediated by exogenous H 2 O 2 . Signaling subsequent to Akt activation was also blocked, notably involving the stimulation of glucose uptake and GLUT4 transporter translocation by insulin, indicating that there are multiple potential cellular signaling effects that arise from the insulin-stimulated generation of H 2 O 2 . Interestingly, the activation of Akt by H 2 O 2 in fibroblasts, embryonic kidney cells and vascular smooth muscle cells was also abrogated by inhibition of PI 3'-kinase, suggesting that cellular PI 3'-kinase is a critical upstream mediator of Akt activation by oxidative stress in a variety of cell types and involving various signaling pathways, including insulin action (46;47). An additional consideration is that the serine protein phosphatase PP2A, which has been implicated in the negative regulation of Akt by dephosphorylation of Ser473, has a redox-sensitive cysteine residue that is potentially susceptible to inhibition by H 2 O 2 (48). One of the possible effects of DPI in this regard could thus be to accentuate the activity of a phosphatase acting directly on Akt which would facilitate the inactivation of Akt. Thus, multiple influences, affecting upstream as well as more distal mechanisms may be influenced by the oxidant signal stemming from cellular insulin stimulation.
In the present study, we also found that unlike H 2 O 2 generated by cell stimulation with PDGF (16), insulin-stimulated production of H 2 O 2 is not mediated by PI 3'-kinase, since the CM-DCF fluorescence induced by insulin binding was not blocked by the PI 3'-kinase inhibitors wortmannin or LY294002 (Figure 8). Thus, the generation of H 2 O 2 in response to insulin in adipocytes may be different from that found in other cell types or with other growth factors. This is consistent with work by Krieger-Brauer and colleagues showing that the elaboration of cellular H 2 O 2 during the process of physiological insulin signal transduction is generated by a novel plasma membrane-bound Mn 2+ -dependent NADPH oxidase that is coupled to Gα i2 (32;33). Other recent work has provided further evidence in several in vivo models that Gi α2 is closely linked to insulin action. For example, Gi α2 deficiency in a transgenic mouse model produces insulin resistance and impaired glucose tolerance (49); conversely, conditional expression of a constitutively active mutant of Gi α2 in insulin-sensitive tissues mimics insulin action in vivo with enhanced glucose tolerance and activation of adipocyte GLUT 4 recruitment, hexose transport and glycogen synthase (50). These data are consistent with the hypothesis that Gi α2 plays a permissive role for insulin signaling, possibly involving a reactive oxygen speciescoupled mechanism, although this has not yet been directly evaluated.
Our data also provides a contrast between the oxidant signal arising from insulin stimulation, inhibited in this work by DPI, and cellular treatment with exogenous H 2 O 2 , which has been used to potentiate or mimic insulin action in previously published studies noted above.
We show that insulin-stimulated activation of PI 3'-kinase and glucose transport (Figures 4 and   7) are completely blocked by prior treatment with DPI, demonstrating an important regulatory role for the endogenous oxidant species in the transmission of the insulin signal. In contrast, in the same experiments, the stimulation of PI 3'-kinase activity or glucose transport by H 2 O 2 in the presence or absence of insulin, is only partially inhibited by DPI. This may be due to several possible mechanisms, including the balance of individual reactive oxygen species, e.g., the proportion of cellular superoxide (reduced by DPI inhibition of NADPH oxidase) and H 2 O 2 provided exogenously, which may differentially regulate various steps in the distal signaling cascade (11). In addition, cellular reactive oxygen species, especially when derived from receptor activation, appear to oxidize a restricted subset of susceptible proteins in the cell, perhaps by colocalization in cellular microdomains (14;51). Although technically challenging, sensitive methods to decipher the subcellular localization and the identity of the various reactive oxygen species in signal transduction are needed to help clarify these issues.
In summary, these data provide new insight into the physiological role of the oxidant signal that arises from cellular insulin stimulation, which affects the initial tyrosine phosphorylation of proteins in the early insulin signaling cascade by oxidative inhibition of PTPases. In addition, the insulin-induced generation of H 2 O 2 is shown here to have important effects on the regulation of downstream insulin signaling, affecting the activation of Akt by a PI 3'-kinase mediated pathway, as well as subsequent steps leading to the activation of glucose uptake and GLUT4 transporter translocation in response to insulin. Defective cellular action of insulin in its target tissues is a fundamental pathophysiologic features of type 2 diabetes mellitus, a prevalent disorder with devastating health and economic consequences in developed societies (52). Thus, a full characterization of the regulation of insulin signal transduction is essential in order to decipher the underlying causes of this disease. Further studies on the formation and cellular role of reactive oxygen species at various steps in the insulin signaling cascade may also provide new targets for therapeutic intervention in diabetes and other insulin-resistant states. Confluent, differentiated 3T3-L1 adipocytes were starved overnight in serum-free medium prior to stimulation with 100 nM insulin for the indicated amount of time. Intracellular H 2 O 2 production was detected by fluorescence of 5,6-chloromethyl-2',7'-dichlorodihydrofluorescein diacetate (CM-DCF), which generates a fluorescent signal in situ visualized by confocal microscopy using fluorescein parameters for excitation and emission as we have described previously (12). Where indicated, cells were pre-incubated for 30 minutes with 10 µM DPI prior to insulin stimulation.